rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/src/backend/utils/adt/jsonfuncs.c

4133 lines
102 KiB
C
Raw Normal View History

Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*-------------------------------------------------------------------------
*
* jsonfuncs.c
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* Functions to process JSON data types.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*
Update copyright via script for 2017
2017-01-03 19:48:53 +01:00
* Portions Copyright (c) 1996-2017, PostgreSQL Global Development Group
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* src/backend/utils/adt/jsonfuncs.c
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include <limits.h>
#include "access/htup_details.h"
#include "catalog/pg_type.h"
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
#include "fmgr.h"
#include "funcapi.h"
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
#include "lib/stringinfo.h"
#include "mb/pg_wchar.h"
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
#include "miscadmin.h"
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/hsearch.h"
#include "utils/json.h"
#include "utils/jsonapi.h"
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
#include "utils/jsonb.h"
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
#include "utils/lsyscache.h"
#include "utils/memutils.h"
#include "utils/typcache.h"
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
/* Operations available for setPath */
#define JB_PATH_CREATE 0x0001
#define JB_PATH_DELETE 0x0002
Fix broken jsonb_set() logic for replacing array elements. Commit 0b62fd036 did a fairly sloppy job of refactoring setPath() to support jsonb_insert() along with jsonb_set(). In its defense, though, there was no regression test case exercising the case of replacing an existing element in a jsonb array. Per bug #14366 from Peng Sun. Back-patch to 9.6 where bug was introduced. Report: <20161012065349.1412.47858@wrigleys.postgresql.org>
2016-10-13 06:25:28 +02:00
#define JB_PATH_REPLACE 0x0004
#define JB_PATH_INSERT_BEFORE 0x0008
#define JB_PATH_INSERT_AFTER 0x0010
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
#define JB_PATH_CREATE_OR_INSERT \
(JB_PATH_INSERT_BEFORE | JB_PATH_INSERT_AFTER | JB_PATH_CREATE)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for json_object_keys */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct OkeysState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
char **result;
int result_size;
int result_count;
int sent_count;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} OkeysState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for json_get* functions */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct GetState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
text *tresult;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
char *result_start;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool normalize_results;
bool next_scalar;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
int npath; /* length of each path-related array */
char **path_names; /* field name(s) being sought */
int *path_indexes; /* array index(es) being sought */
bool *pathok; /* is path matched to current depth? */
int *array_cur_index; /* current element index at each path level */
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} GetState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for json_array_length */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct AlenState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
int count;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} AlenState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for json_each */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct EachState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
Tuplestorestate *tuple_store;
TupleDesc ret_tdesc;
MemoryContext tmp_cxt;
char *result_start;
bool normalize_results;
bool next_scalar;
char *normalized_scalar;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} EachState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for json_array_elements */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct ElementsState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
const char *function_name;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Tuplestorestate *tuple_store;
TupleDesc ret_tdesc;
MemoryContext tmp_cxt;
char *result_start;
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
bool normalize_results;
bool next_scalar;
char *normalized_scalar;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} ElementsState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* state for get_json_object_as_hash */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct JhashState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
const char *function_name;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
HTAB *hash;
char *saved_scalar;
char *save_json_start;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} JHashState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
/* hashtable element */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct JsonHashEntry
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
char fname[NAMEDATALEN]; /* hash key (MUST BE FIRST) */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
char *val;
char *json;
bool isnull;
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} JsonHashEntry;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* these two are stolen from hstore / record_out, used in populate_record* */
typedef struct ColumnIOData
{
Oid column_type;
Oid typiofunc;
Oid typioparam;
FmgrInfo proc;
} ColumnIOData;
typedef struct RecordIOData
{
Oid record_type;
int32 record_typmod;
int ncolumns;
Use FLEXIBLE_ARRAY_MEMBER in struct RecordIOData. I (tgl) fixed this last night in rowtypes.c, but I missed that the code had been copied into a couple of other places. Michael Paquier
2015-02-20 23:03:12 +01:00
ColumnIOData columns[FLEXIBLE_ARRAY_MEMBER];
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
} RecordIOData;
/* state for populate_recordset */
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
typedef struct PopulateRecordsetState
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
JsonLexContext *lex;
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
const char *function_name;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
HTAB *json_hash;
char *saved_scalar;
char *save_json_start;
Tuplestorestate *tuple_store;
TupleDesc ret_tdesc;
HeapTupleHeader rec;
RecordIOData *my_extra;
MemoryContext fn_mcxt; /* used to stash IO funcs */
Reindent json.c and jsonfuncs.c. This will help in preparation of clean patches for upcoming json work.
2014-01-22 14:46:51 +01:00
} PopulateRecordsetState;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
/* state for json_strip_nulls */
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
typedef struct StripnullState
{
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
JsonLexContext *lex;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
StringInfo strval;
bool skip_next_null;
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
} StripnullState;
Relocate static function declarations to be after typedefs in jsonfuncs.c. Project style is to put things in this order, for the good and sufficient reason that you often need the typedefs in the function declarations. There already was one function declaration that needed a typedef, which was randomly placed away from all the other static function declarations in consequence. And the submitted patch for better json_populate_record functionality jumped through even more hoops in order to preserve this bad idea. This patch only moves lines from point A to point B, no other changes.
2017-01-22 20:08:26 +01:00
/* semantic action functions for json_object_keys */
static void okeys_object_field_start(void *state, char *fname, bool isnull);
static void okeys_array_start(void *state);
static void okeys_scalar(void *state, char *token, JsonTokenType tokentype);
/* semantic action functions for json_get* functions */
static void get_object_start(void *state);
static void get_object_end(void *state);
static void get_object_field_start(void *state, char *fname, bool isnull);
static void get_object_field_end(void *state, char *fname, bool isnull);
static void get_array_start(void *state);
static void get_array_end(void *state);
static void get_array_element_start(void *state, bool isnull);
static void get_array_element_end(void *state, bool isnull);
static void get_scalar(void *state, char *token, JsonTokenType tokentype);
/* common worker function for json getter functions */
static Datum get_path_all(FunctionCallInfo fcinfo, bool as_text);
static text *get_worker(text *json, char **tpath, int *ipath, int npath,
bool normalize_results);
static Datum get_jsonb_path_all(FunctionCallInfo fcinfo, bool as_text);
/* semantic action functions for json_array_length */
static void alen_object_start(void *state);
static void alen_scalar(void *state, char *token, JsonTokenType tokentype);
static void alen_array_element_start(void *state, bool isnull);
/* common workers for json{b}_each* functions */
static Datum each_worker(FunctionCallInfo fcinfo, bool as_text);
static Datum each_worker_jsonb(FunctionCallInfo fcinfo, const char *funcname,
bool as_text);
/* semantic action functions for json_each */
static void each_object_field_start(void *state, char *fname, bool isnull);
static void each_object_field_end(void *state, char *fname, bool isnull);
static void each_array_start(void *state);
static void each_scalar(void *state, char *token, JsonTokenType tokentype);
/* common workers for json{b}_array_elements_* functions */
static Datum elements_worker(FunctionCallInfo fcinfo, const char *funcname,
bool as_text);
static Datum elements_worker_jsonb(FunctionCallInfo fcinfo, const char *funcname,
bool as_text);
/* semantic action functions for json_array_elements */
static void elements_object_start(void *state);
static void elements_array_element_start(void *state, bool isnull);
static void elements_array_element_end(void *state, bool isnull);
static void elements_scalar(void *state, char *token, JsonTokenType tokentype);
/* turn a json object into a hash table */
static HTAB *get_json_object_as_hash(text *json, const char *funcname);
/* common worker for populate_record and to_record */
static Datum populate_record_worker(FunctionCallInfo fcinfo, const char *funcname,
bool have_record_arg);
/* semantic action functions for get_json_object_as_hash */
static void hash_object_field_start(void *state, char *fname, bool isnull);
static void hash_object_field_end(void *state, char *fname, bool isnull);
static void hash_array_start(void *state);
static void hash_scalar(void *state, char *token, JsonTokenType tokentype);
/* semantic action functions for populate_recordset */
static void populate_recordset_object_field_start(void *state, char *fname, bool isnull);
static void populate_recordset_object_field_end(void *state, char *fname, bool isnull);
static void populate_recordset_scalar(void *state, char *token, JsonTokenType tokentype);
static void populate_recordset_object_start(void *state);
static void populate_recordset_object_end(void *state);
static void populate_recordset_array_start(void *state);
static void populate_recordset_array_element_start(void *state, bool isnull);
/* semantic action functions for json_strip_nulls */
static void sn_object_start(void *state);
static void sn_object_end(void *state);
static void sn_array_start(void *state);
static void sn_array_end(void *state);
static void sn_object_field_start(void *state, char *fname, bool isnull);
static void sn_array_element_start(void *state, bool isnull);
static void sn_scalar(void *state, char *token, JsonTokenType tokentype);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* Turn a jsonb object into a record */
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
static void make_row_from_rec_and_jsonb(Jsonb *element,
PopulateRecordsetState *state);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Relocate static function declarations to be after typedefs in jsonfuncs.c. Project style is to put things in this order, for the good and sufficient reason that you often need the typedefs in the function declarations. There already was one function declaration that needed a typedef, which was randomly placed away from all the other static function declarations in consequence. And the submitted patch for better json_populate_record functionality jumped through even more hoops in order to preserve this bad idea. This patch only moves lines from point A to point B, no other changes.
2017-01-22 20:08:26 +01:00
/* worker function for populate_recordset and to_recordset */
static Datum populate_recordset_worker(FunctionCallInfo fcinfo, const char *funcname,
bool have_record_arg);
/* Worker that takes care of common setup for us */
static JsonbValue *findJsonbValueFromContainerLen(JsonbContainer *container,
uint32 flags,
char *key,
uint32 keylen);
/* functions supporting jsonb_delete, jsonb_set and jsonb_concat */
static JsonbValue *IteratorConcat(JsonbIterator **it1, JsonbIterator **it2,
JsonbParseState **state);
static JsonbValue *setPath(JsonbIterator **it, Datum *path_elems,
bool *path_nulls, int path_len,
JsonbParseState **st, int level, Jsonb *newval,
int op_type);
static void setPathObject(JsonbIterator **it, Datum *path_elems,
bool *path_nulls, int path_len, JsonbParseState **st,
int level,
Jsonb *newval, uint32 npairs, int op_type);
static void setPathArray(JsonbIterator **it, Datum *path_elems,
bool *path_nulls, int path_len, JsonbParseState **st,
int level, Jsonb *newval, uint32 nelems, int op_type);
static void addJsonbToParseState(JsonbParseState **jbps, Jsonb *jb);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* SQL function json_object_keys
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*
* Returns the set of keys for the object argument.
*
* This SRF operates in value-per-call mode. It processes the
* object during the first call, and the keys are simply stashed
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* in an array, whose size is expanded as necessary. This is probably
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
* safe enough for a list of keys of a single object, since they are
* limited in size to NAMEDATALEN and the number of keys is unlikely to
* be so huge that it has major memory implications.
*/
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_object_keys(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
OkeysState *state;
int i;
if (SRF_IS_FIRSTCALL())
{
MemoryContext oldcontext;
Jsonb *jb = PG_GETARG_JSONB(0);
bool skipNested = false;
JsonbIterator *it;
JsonbValue v;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (JB_ROOT_IS_SCALAR(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a scalar",
"jsonb_object_keys")));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (JB_ROOT_IS_ARRAY(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on an array",
"jsonb_object_keys")));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
state = palloc(sizeof(OkeysState));
state->result_size = JB_ROOT_COUNT(jb);
state->result_count = 0;
state->sent_count = 0;
state->result = palloc(state->result_size * sizeof(char *));
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
it = JsonbIteratorInit(&jb->root);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != WJB_DONE)
{
skipNested = true;
if (r == WJB_KEY)
{
char *cstr;
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
cstr = palloc(v.val.string.len + 1 * sizeof(char));
memcpy(cstr, v.val.string.val, v.val.string.len);
cstr[v.val.string.len] = '\0';
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
state->result[state->result_count++] = cstr;
}
}
MemoryContextSwitchTo(oldcontext);
funcctx->user_fctx = (void *) state;
}
funcctx = SRF_PERCALL_SETUP();
state = (OkeysState *) funcctx->user_fctx;
if (state->sent_count < state->result_count)
{
char *nxt = state->result[state->sent_count++];
SRF_RETURN_NEXT(funcctx, CStringGetTextDatum(nxt));
}
/* cleanup to reduce or eliminate memory leaks */
for (i = 0; i < state->result_count; i++)
pfree(state->result[i]);
pfree(state->result);
pfree(state);
SRF_RETURN_DONE(funcctx);
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_object_keys(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
OkeysState *state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
int i;
if (SRF_IS_FIRSTCALL())
{
text *json = PG_GETARG_TEXT_P(0);
JsonLexContext *lex = makeJsonLexContext(json, true);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonSemAction *sem;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
MemoryContext oldcontext;
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = palloc(sizeof(OkeysState));
sem = palloc0(sizeof(JsonSemAction));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->lex = lex;
state->result_size = 256;
state->result_count = 0;
state->sent_count = 0;
state->result = palloc(256 * sizeof(char *));
sem->semstate = (void *) state;
sem->array_start = okeys_array_start;
sem->scalar = okeys_scalar;
sem->object_field_start = okeys_object_field_start;
/* remainder are all NULL, courtesy of palloc0 above */
pg_parse_json(lex, sem);
/* keys are now in state->result */
pfree(lex->strval->data);
pfree(lex->strval);
pfree(lex);
pfree(sem);
MemoryContextSwitchTo(oldcontext);
funcctx->user_fctx = (void *) state;
}
funcctx = SRF_PERCALL_SETUP();
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = (OkeysState *) funcctx->user_fctx;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (state->sent_count < state->result_count)
{
char *nxt = state->result[state->sent_count++];
SRF_RETURN_NEXT(funcctx, CStringGetTextDatum(nxt));
}
/* cleanup to reduce or eliminate memory leaks */
for (i = 0; i < state->result_count; i++)
pfree(state->result[i]);
pfree(state->result);
pfree(state);
SRF_RETURN_DONE(funcctx);
}
static void
okeys_object_field_start(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
OkeysState *_state = (OkeysState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* only collecting keys for the top level object */
if (_state->lex->lex_level != 1)
return;
/* enlarge result array if necessary */
if (_state->result_count >= _state->result_size)
{
_state->result_size *= 2;
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
_state->result = (char **)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
repalloc(_state->result, sizeof(char *) * _state->result_size);
}
/* save a copy of the field name */
_state->result[_state->result_count++] = pstrdup(fname);
}
static void
okeys_array_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
OkeysState *_state = (OkeysState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* top level must be a json object */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on an array",
"json_object_keys")));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
okeys_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
OkeysState *_state = (OkeysState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* top level must be a json object */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a scalar",
"json_object_keys")));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
/*
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* json and jsonb getter functions
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
* these implement the -> ->> #> and #>> operators
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* and the json{b?}_extract_path*(json, text, ...) functions
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Datum
json_object_field(PG_FUNCTION_ARGS)
{
text *json = PG_GETARG_TEXT_P(0);
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
text *fname = PG_GETARG_TEXT_PP(1);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
char *fnamestr = text_to_cstring(fname);
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *result;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
result = get_worker(json, &fnamestr, NULL, 1, false);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (result != NULL)
PG_RETURN_TEXT_P(result);
else
PG_RETURN_NULL();
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_object_field(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
text *key = PG_GETARG_TEXT_PP(1);
JsonbValue *v;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (!JB_ROOT_IS_OBJECT(jb))
PG_RETURN_NULL();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
v = findJsonbValueFromContainerLen(&jb->root, JB_FOBJECT,
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
VARDATA_ANY(key),
VARSIZE_ANY_EXHDR(key));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
if (v != NULL)
PG_RETURN_JSONB(JsonbValueToJsonb(v));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_NULL();
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_object_field_text(PG_FUNCTION_ARGS)
{
text *json = PG_GETARG_TEXT_P(0);
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
text *fname = PG_GETARG_TEXT_PP(1);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
char *fnamestr = text_to_cstring(fname);
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *result;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
result = get_worker(json, &fnamestr, NULL, 1, true);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (result != NULL)
PG_RETURN_TEXT_P(result);
else
PG_RETURN_NULL();
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_object_field_text(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
text *key = PG_GETARG_TEXT_PP(1);
JsonbValue *v;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (!JB_ROOT_IS_OBJECT(jb))
PG_RETURN_NULL();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
v = findJsonbValueFromContainerLen(&jb->root, JB_FOBJECT,
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
VARDATA_ANY(key),
VARSIZE_ANY_EXHDR(key));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
if (v != NULL)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
text *result = NULL;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
switch (v->type)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
case jbvNull:
break;
case jbvBool:
result = cstring_to_text(v->val.boolean ? "true" : "false");
break;
case jbvString:
result = cstring_to_text_with_len(v->val.string.val, v->val.string.len);
break;
case jbvNumeric:
result = cstring_to_text(DatumGetCString(DirectFunctionCall1(numeric_out,
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
PointerGetDatum(v->val.numeric))));
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
break;
case jbvBinary:
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
StringInfo jtext = makeStringInfo();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
(void) JsonbToCString(jtext, v->val.binary.data, -1);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
result = cstring_to_text_with_len(jtext->data, jtext->len);
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
break;
default:
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elog(ERROR, "unrecognized jsonb type: %d", (int) v->type);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
if (result)
PG_RETURN_TEXT_P(result);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
PG_RETURN_NULL();
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_array_element(PG_FUNCTION_ARGS)
{
text *json = PG_GETARG_TEXT_P(0);
int element = PG_GETARG_INT32(1);
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *result;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
result = get_worker(json, NULL, &element, 1, false);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (result != NULL)
PG_RETURN_TEXT_P(result);
else
PG_RETURN_NULL();
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_array_element(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
int element = PG_GETARG_INT32(1);
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
JsonbValue *v;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (!JB_ROOT_IS_ARRAY(jb))
PG_RETURN_NULL();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
/* Handle negative subscript */
if (element < 0)
{
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
uint32 nelements = JB_ROOT_COUNT(jb);
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (-element > nelements)
PG_RETURN_NULL();
else
element += nelements;
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
v = getIthJsonbValueFromContainer(&jb->root, element);
if (v != NULL)
PG_RETURN_JSONB(JsonbValueToJsonb(v));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_NULL();
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_array_element_text(PG_FUNCTION_ARGS)
{
text *json = PG_GETARG_TEXT_P(0);
int element = PG_GETARG_INT32(1);
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *result;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
result = get_worker(json, NULL, &element, 1, true);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (result != NULL)
PG_RETURN_TEXT_P(result);
else
PG_RETURN_NULL();
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_array_element_text(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
int element = PG_GETARG_INT32(1);
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
JsonbValue *v;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (!JB_ROOT_IS_ARRAY(jb))
PG_RETURN_NULL();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
/* Handle negative subscript */
if (element < 0)
{
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
uint32 nelements = JB_ROOT_COUNT(jb);
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (-element > nelements)
PG_RETURN_NULL();
else
element += nelements;
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
v = getIthJsonbValueFromContainer(&jb->root, element);
if (v != NULL)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
text *result = NULL;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
switch (v->type)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
case jbvNull:
break;
case jbvBool:
result = cstring_to_text(v->val.boolean ? "true" : "false");
break;
case jbvString:
result = cstring_to_text_with_len(v->val.string.val, v->val.string.len);
break;
case jbvNumeric:
result = cstring_to_text(DatumGetCString(DirectFunctionCall1(numeric_out,
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
PointerGetDatum(v->val.numeric))));
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
break;
case jbvBinary:
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
StringInfo jtext = makeStringInfo();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
(void) JsonbToCString(jtext, v->val.binary.data, -1);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
result = cstring_to_text_with_len(jtext->data, jtext->len);
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
break;
default:
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elog(ERROR, "unrecognized jsonb type: %d", (int) v->type);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Improve the efficiency of certain jsonb get operations. Instead of iterating over jsonb structures, use the inbuilt functions findJsonbValueFromContainerLen() and getIthJsonbValueFromContainer() to extract values directly. These functions use algorithms that are O(n log n) and O(1) respectively, whereas iterating is O(n), so we should see considerable speedup here. Teodor Sigaev.
2014-06-02 01:04:02 +02:00
if (result)
PG_RETURN_TEXT_P(result);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
PG_RETURN_NULL();
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_extract_path(PG_FUNCTION_ARGS)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
return get_path_all(fcinfo, false);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Datum
json_extract_path_text(PG_FUNCTION_ARGS)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
return get_path_all(fcinfo, true);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
/*
* common routine for extract_path functions
*/
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
get_path_all(FunctionCallInfo fcinfo, bool as_text)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *json = PG_GETARG_TEXT_P(0);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ArrayType *path = PG_GETARG_ARRAYTYPE_P(1);
text *result;
Datum *pathtext;
bool *pathnulls;
int npath;
char **tpath;
int *ipath;
int i;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/*
* If the array contains any null elements, return NULL, on the grounds
* that you'd have gotten NULL if any RHS value were NULL in a nested
* series of applications of the -> operator. (Note: because we also
* return NULL for error cases such as no-such-field, this is true
* regardless of the contents of the rest of the array.)
*/
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (array_contains_nulls(path))
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
PG_RETURN_NULL();
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
deconstruct_array(path, TEXTOID, -1, false, 'i',
&pathtext, &pathnulls, &npath);
tpath = palloc(npath * sizeof(char *));
ipath = palloc(npath * sizeof(int));
for (i = 0; i < npath; i++)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
Assert(!pathnulls[i]);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
tpath[i] = TextDatumGetCString(pathtext[i]);
/*
* we have no idea at this stage what structure the document is so
* just convert anything in the path that we can to an integer and set
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
* all the other integers to INT_MIN which will never match.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (*tpath[i] != '\0')
{
long ind;
char *endptr;
errno = 0;
ind = strtol(tpath[i], &endptr, 10);
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (*endptr == '\0' && errno == 0 && ind <= INT_MAX && ind >= INT_MIN)
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
ipath[i] = (int) ind;
else
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
ipath[i] = INT_MIN;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
else
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
ipath[i] = INT_MIN;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
result = get_worker(json, tpath, ipath, npath, as_text);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (result != NULL)
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
PG_RETURN_TEXT_P(result);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
else
PG_RETURN_NULL();
}
/*
* get_worker
*
* common worker for all the json getter functions
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
*
* json: JSON object (in text form)
* tpath[]: field name(s) to extract
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
* ipath[]: array index(es) (zero-based) to extract, accepts negatives
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
* npath: length of tpath[] and/or ipath[]
* normalize_results: true to de-escape string and null scalars
*
* tpath can be NULL, or any one tpath[] entry can be NULL, if an object
* field is not to be matched at that nesting level. Similarly, ipath can
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
* be NULL, or any one ipath[] entry can be INT_MIN if an array element is
* not to be matched at that nesting level (a json datum should never be
* large enough to have -INT_MIN elements due to MaxAllocSize restriction).
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static text *
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
get_worker(text *json,
char **tpath,
int *ipath,
int npath,
bool normalize_results)
{
JsonLexContext *lex = makeJsonLexContext(json, true);
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
JsonSemAction *sem = palloc0(sizeof(JsonSemAction));
GetState *state = palloc0(sizeof(GetState));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
Assert(npath >= 0);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->lex = lex;
/* is it "_as_text" variant? */
state->normalize_results = normalize_results;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
state->npath = npath;
state->path_names = tpath;
state->path_indexes = ipath;
state->pathok = palloc0(sizeof(bool) * npath);
state->array_cur_index = palloc(sizeof(int) * npath);
if (npath > 0)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->pathok[0] = true;
sem->semstate = (void *) state;
/*
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
* Not all variants need all the semantic routines. Only set the ones that
* are actually needed for maximum efficiency.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
sem->scalar = get_scalar;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (npath == 0)
{
sem->object_start = get_object_start;
sem->object_end = get_object_end;
sem->array_start = get_array_start;
sem->array_end = get_array_end;
}
if (tpath != NULL)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
sem->object_field_start = get_object_field_start;
sem->object_field_end = get_object_field_end;
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (ipath != NULL)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
sem->array_start = get_array_start;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
sem->array_element_start = get_array_element_start;
sem->array_element_end = get_array_element_end;
}
pg_parse_json(lex, sem);
return state->tresult;
}
static void
get_object_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
int lex_level = _state->lex->lex_level;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (lex_level == 0 && _state->npath == 0)
{
/*
* Special case: we should match the entire object. We only need this
* at outermost level because at nested levels the match will have
* been started by the outer field or array element callback.
*/
_state->result_start = _state->lex->token_start;
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
get_object_end(void *state)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
int lex_level = _state->lex->lex_level;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (lex_level == 0 && _state->npath == 0)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* Special case: return the entire object */
char *start = _state->result_start;
int len = _state->lex->prev_token_terminator - start;
_state->tresult = cstring_to_text_with_len(start, len);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
static void
get_object_field_start(void *state, char *fname, bool isnull)
{
GetState *_state = (GetState *) state;
bool get_next = false;
int lex_level = _state->lex->lex_level;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (lex_level <= _state->npath &&
_state->pathok[lex_level - 1] &&
_state->path_names != NULL &&
_state->path_names[lex_level - 1] != NULL &&
strcmp(fname, _state->path_names[lex_level - 1]) == 0)
{
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (lex_level < _state->npath)
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
{
/* if not at end of path just mark path ok */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
_state->pathok[lex_level] = true;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
else
{
/* end of path, so we want this value */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
get_next = true;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
if (get_next)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* this object overrides any previous matching object */
_state->tresult = NULL;
_state->result_start = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->normalize_results &&
_state->lex->token_type == JSON_TOKEN_STRING)
{
/* for as_text variants, tell get_scalar to set it for us */
_state->next_scalar = true;
}
else
{
/* for non-as_text variants, just note the json starting point */
_state->result_start = _state->lex->token_start;
}
}
}
static void
get_object_field_end(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool get_last = false;
int lex_level = _state->lex->lex_level;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* same tests as in get_object_field_start */
if (lex_level <= _state->npath &&
_state->pathok[lex_level - 1] &&
_state->path_names != NULL &&
_state->path_names[lex_level - 1] != NULL &&
strcmp(fname, _state->path_names[lex_level - 1]) == 0)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
if (lex_level < _state->npath)
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
{
/* done with this field so reset pathok */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
_state->pathok[lex_level] = false;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
else
{
/* end of path, so we want this value */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
get_last = true;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* for as_text scalar case, our work is already done */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (get_last && _state->result_start != NULL)
{
/*
* make a text object from the string from the prevously noted json
* start up to the end of the previous token (the lexer is by now
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* ahead of us on whatever came after what we're interested in).
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
if (isnull && _state->normalize_results)
_state->tresult = (text *) NULL;
else
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
{
char *start = _state->result_start;
int len = _state->lex->prev_token_terminator - start;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
_state->tresult = cstring_to_text_with_len(start, len);
}
/* this should be unnecessary but let's do it for cleanliness: */
_state->result_start = NULL;
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
get_array_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
int lex_level = _state->lex->lex_level;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (lex_level < _state->npath)
{
/* Initialize counting of elements in this array */
_state->array_cur_index[lex_level] = -1;
Only adjust negative indexes in json_get up to the length of the path. The previous code resulted in memory access beyond the path bounds. The cure is to move it into a code branch that checks the value of lex_level is within the correct bounds. Bug reported and diagnosed by Piotr Stefaniak.
2015-07-28 23:54:13 +02:00
/* INT_MIN value is reserved to represent invalid subscript */
if (_state->path_indexes[lex_level] < 0 &&
_state->path_indexes[lex_level] != INT_MIN)
{
/* Negative subscript -- convert to positive-wise subscript */
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
int nelements = json_count_array_elements(_state->lex);
Only adjust negative indexes in json_get up to the length of the path. The previous code resulted in memory access beyond the path bounds. The cure is to move it into a code branch that checks the value of lex_level is within the correct bounds. Bug reported and diagnosed by Piotr Stefaniak.
2015-07-28 23:54:13 +02:00
if (-_state->path_indexes[lex_level] <= nelements)
_state->path_indexes[lex_level] += nelements;
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
else if (lex_level == 0 && _state->npath == 0)
{
/*
* Special case: we should match the entire array. We only need this
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
* at the outermost level because at nested levels the match will have
* been started by the outer field or array element callback.
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
*/
_state->result_start = _state->lex->token_start;
}
}
pgindent run for release 9.3 This is the first run of the Perl-based pgindent script. Also update pgindent instructions.
2013-05-29 22:58:43 +02:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
static void
get_array_end(void *state)
{
GetState *_state = (GetState *) state;
int lex_level = _state->lex->lex_level;
if (lex_level == 0 && _state->npath == 0)
{
/* Special case: return the entire array */
char *start = _state->result_start;
int len = _state->lex->prev_token_terminator - start;
_state->tresult = cstring_to_text_with_len(start, len);
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
get_array_element_start(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool get_next = false;
int lex_level = _state->lex->lex_level;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* Update array element counter */
if (lex_level <= _state->npath)
_state->array_cur_index[lex_level - 1]++;
if (lex_level <= _state->npath &&
_state->pathok[lex_level - 1] &&
_state->path_indexes != NULL &&
_state->array_cur_index[lex_level - 1] == _state->path_indexes[lex_level - 1])
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (lex_level < _state->npath)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* if not at end of path just mark path ok */
_state->pathok[lex_level] = true;
}
else
{
/* end of path, so we want this value */
get_next = true;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
}
/* same logic as for objects */
if (get_next)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
_state->tresult = NULL;
_state->result_start = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->normalize_results &&
_state->lex->token_type == JSON_TOKEN_STRING)
{
_state->next_scalar = true;
}
else
{
_state->result_start = _state->lex->token_start;
}
}
}
static void
get_array_element_end(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool get_last = false;
int lex_level = _state->lex->lex_level;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* same tests as in get_array_element_start */
if (lex_level <= _state->npath &&
_state->pathok[lex_level - 1] &&
_state->path_indexes != NULL &&
_state->array_cur_index[lex_level - 1] == _state->path_indexes[lex_level - 1])
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
if (lex_level < _state->npath)
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
{
/* done with this element so reset pathok */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
_state->pathok[lex_level] = false;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
else
{
/* end of path, so we want this value */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
get_last = true;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* same logic as for objects */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (get_last && _state->result_start != NULL)
{
if (isnull && _state->normalize_results)
_state->tresult = (text *) NULL;
else
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
{
char *start = _state->result_start;
int len = _state->lex->prev_token_terminator - start;
_state->tresult = cstring_to_text_with_len(start, len);
}
_state->result_start = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
}
static void
get_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
GetState *_state = (GetState *) state;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
int lex_level = _state->lex->lex_level;
/* Check for whole-object match */
if (lex_level == 0 && _state->npath == 0)
{
if (_state->normalize_results && tokentype == JSON_TOKEN_STRING)
{
/* we want the de-escaped string */
_state->next_scalar = true;
}
else if (_state->normalize_results && tokentype == JSON_TOKEN_NULL)
{
_state->tresult = (text *) NULL;
}
else
{
/*
* This is a bit hokey: we will suppress whitespace after the
* scalar token, but not whitespace before it. Probably not worth
* doing our own space-skipping to avoid that.
*/
char *start = _state->lex->input;
int len = _state->lex->prev_token_terminator - start;
_state->tresult = cstring_to_text_with_len(start, len);
}
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->next_scalar)
{
/* a de-escaped text value is wanted, so supply it */
_state->tresult = cstring_to_text(token);
/* make sure the next call to get_scalar doesn't overwrite it */
_state->next_scalar = false;
}
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_extract_path(PG_FUNCTION_ARGS)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
return get_jsonb_path_all(fcinfo, false);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Datum
jsonb_extract_path_text(PG_FUNCTION_ARGS)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
return get_jsonb_path_all(fcinfo, true);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
get_jsonb_path_all(FunctionCallInfo fcinfo, bool as_text)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Jsonb *jb = PG_GETARG_JSONB(0);
ArrayType *path = PG_GETARG_ARRAYTYPE_P(1);
Fix core dump in jsonb #> operator, and add regression test cases. jsonb's #> operator segfaulted (dereferencing a null pointer) if the RHS was a zero-length array, as reported in bug #11207 from Justin Van Winkle. json's #> operator returns NULL in such cases, so for the moment let's make jsonb act likewise. Also add a bunch of regression test queries memorializing the -> and #> operators' behavior for this and other corner cases. There is a good argument for changing some of these behaviors, as they are not very consistent with each other, and throwing an error isn't necessarily a desirable behavior for operators that are likely to be used in indexes. However, everybody can agree that a core dump is the Wrong Thing, and we need test cases even if we decide to change their expected output later.
2014-08-20 22:48:35 +02:00
Jsonb *res;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum *pathtext;
bool *pathnulls;
int npath;
int i;
bool have_object = false,
have_array = false;
JsonbValue *jbvp = NULL;
JsonbValue tv;
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
JsonbContainer *container;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/*
* If the array contains any null elements, return NULL, on the grounds
* that you'd have gotten NULL if any RHS value were NULL in a nested
* series of applications of the -> operator. (Note: because we also
* return NULL for error cases such as no-such-field, this is true
* regardless of the contents of the rest of the array.)
*/
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (array_contains_nulls(path))
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
PG_RETURN_NULL();
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
deconstruct_array(path, TEXTOID, -1, false, 'i',
&pathtext, &pathnulls, &npath);
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* Identify whether we have object, array, or scalar at top-level */
container = &jb->root;
Fix core dump in jsonb #> operator, and add regression test cases. jsonb's #> operator segfaulted (dereferencing a null pointer) if the RHS was a zero-length array, as reported in bug #11207 from Justin Van Winkle. json's #> operator returns NULL in such cases, so for the moment let's make jsonb act likewise. Also add a bunch of regression test queries memorializing the -> and #> operators' behavior for this and other corner cases. There is a good argument for changing some of these behaviors, as they are not very consistent with each other, and throwing an error isn't necessarily a desirable behavior for operators that are likely to be used in indexes. However, everybody can agree that a core dump is the Wrong Thing, and we need test cases even if we decide to change their expected output later.
2014-08-20 22:48:35 +02:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (JB_ROOT_IS_OBJECT(jb))
have_object = true;
else if (JB_ROOT_IS_ARRAY(jb) && !JB_ROOT_IS_SCALAR(jb))
have_array = true;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
else
{
Assert(JB_ROOT_IS_ARRAY(jb) && JB_ROOT_IS_SCALAR(jb));
/* Extract the scalar value, if it is what we'll return */
if (npath <= 0)
jbvp = getIthJsonbValueFromContainer(container, 0);
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/*
* If the array is empty, return the entire LHS object, on the grounds
* that we should do zero field or element extractions. For the
* non-scalar case we can just hand back the object without much work. For
* the scalar case, fall through and deal with the value below the loop.
* (This inconsistency arises because there's no easy way to generate a
* JsonbValue directly for root-level containers.)
*/
if (npath <= 0 && jbvp == NULL)
{
if (as_text)
{
PG_RETURN_TEXT_P(cstring_to_text(JsonbToCString(NULL,
container,
VARSIZE(jb))));
}
else
{
/* not text mode - just hand back the jsonb */
PG_RETURN_JSONB(jb);
}
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
for (i = 0; i < npath; i++)
{
if (have_object)
{
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
jbvp = findJsonbValueFromContainerLen(container,
JB_FOBJECT,
VARDATA_ANY(pathtext[i]),
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
VARSIZE_ANY_EXHDR(pathtext[i]));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else if (have_array)
{
long lindex;
uint32 index;
char *indextext = TextDatumGetCString(pathtext[i]);
char *endptr;
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
errno = 0;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
lindex = strtol(indextext, &endptr, 10);
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
if (endptr == indextext || *endptr != '\0' || errno != 0 ||
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
lindex > INT_MAX || lindex < INT_MIN)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_NULL();
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (lindex >= 0)
{
index = (uint32) lindex;
}
else
{
/* Handle negative subscript */
uint32 nelements;
/* Container must be array, but make sure */
if ((container->header & JB_FARRAY) == 0)
elog(ERROR, "not a jsonb array");
nelements = container->header & JB_CMASK;
if (-lindex > nelements)
PG_RETURN_NULL();
else
index = nelements + lindex;
}
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
jbvp = getIthJsonbValueFromContainer(container, index);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* scalar, extraction yields a null */
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_NULL();
}
if (jbvp == NULL)
PG_RETURN_NULL();
else if (i == npath - 1)
break;
if (jbvp->type == jbvBinary)
{
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
JsonbIterator *it = JsonbIteratorInit((JsonbContainer *) jbvp->val.binary.data);
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
r = JsonbIteratorNext(&it, &tv, true);
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
container = (JsonbContainer *) jbvp->val.binary.data;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
have_object = r == WJB_BEGIN_OBJECT;
have_array = r == WJB_BEGIN_ARRAY;
}
else
{
have_object = jbvp->type == jbvObject;
have_array = jbvp->type == jbvArray;
}
}
if (as_text)
{
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
/* special-case outputs for string and null values */
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (jbvp->type == jbvString)
Fix corner-case behaviors in JSON/JSONB field extraction operators. Cause the path extraction operators to return their lefthand input, not NULL, if the path array has no elements. This seems more consistent since the case ought to correspond to applying the simple extraction operator (->) zero times. Cause other corner cases in field/element/path extraction to return NULL rather than failing. This behavior is arguably more useful than throwing an error, since it allows an expression index using these operators to be built even when not all values in the column are suitable for the extraction being indexed. Moreover, we already had multiple inconsistencies between the path extraction operators and the simple extraction operators, as well as inconsistencies between the JSON and JSONB code paths. Adopt a uniform rule of returning NULL rather than throwing an error when the JSON input does not have a structure that permits the request to be satisfied. Back-patch to 9.4. Update the release notes to list this as a behavior change since 9.3.
2014-08-22 19:17:58 +02:00
PG_RETURN_TEXT_P(cstring_to_text_with_len(jbvp->val.string.val,
jbvp->val.string.len));
if (jbvp->type == jbvNull)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_NULL();
}
res = JsonbValueToJsonb(jbvp);
if (as_text)
{
PG_RETURN_TEXT_P(cstring_to_text(JsonbToCString(NULL,
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
&res->root,
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
VARSIZE(res))));
}
else
{
/* not text mode - just hand back the jsonb */
PG_RETURN_JSONB(res);
}
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
* SQL function json_array_length(json) -> int
*/
Datum
json_array_length(PG_FUNCTION_ARGS)
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *json = PG_GETARG_TEXT_P(0);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
AlenState *state;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
JsonLexContext *lex;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonSemAction *sem;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
lex = makeJsonLexContext(json, false);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = palloc0(sizeof(AlenState));
sem = palloc0(sizeof(JsonSemAction));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* palloc0 does this for us */
#if 0
state->count = 0;
#endif
state->lex = lex;
sem->semstate = (void *) state;
sem->object_start = alen_object_start;
sem->scalar = alen_scalar;
sem->array_element_start = alen_array_element_start;
pg_parse_json(lex, sem);
PG_RETURN_INT32(state->count);
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_array_length(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
if (JB_ROOT_IS_SCALAR(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot get array length of a scalar")));
else if (!JB_ROOT_IS_ARRAY(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot get array length of a non-array")));
PG_RETURN_INT32(JB_ROOT_COUNT(jb));
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
* These next two checks ensure that the json is an array (since it can't be
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
* a scalar or an object).
*/
static void
alen_object_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
AlenState *_state = (AlenState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot get array length of a non-array")));
}
static void
alen_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
AlenState *_state = (AlenState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot get array length of a scalar")));
}
static void
alen_array_element_start(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
AlenState *_state = (AlenState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* just count up all the level 1 elements */
if (_state->lex->lex_level == 1)
_state->count++;
}
/*
* SQL function json_each and json_each_text
*
* decompose a json object into key value pairs.
*
* Unlike json_object_keys() these SRFs operate in materialize mode,
* stashing results into a Tuplestore object as they go.
* The construction of tuples is done using a temporary memory context
* that is cleared out after each tuple is built.
*/
Datum
json_each(PG_FUNCTION_ARGS)
{
return each_worker(fcinfo, false);
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_each(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return each_worker_jsonb(fcinfo, "jsonb_each", false);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_each_text(PG_FUNCTION_ARGS)
{
return each_worker(fcinfo, true);
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_each_text(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return each_worker_jsonb(fcinfo, "jsonb_each_text", true);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
each_worker_jsonb(FunctionCallInfo fcinfo, const char *funcname, bool as_text)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Jsonb *jb = PG_GETARG_JSONB(0);
ReturnSetInfo *rsi;
Tuplestorestate *tuple_store;
TupleDesc tupdesc;
TupleDesc ret_tdesc;
MemoryContext old_cxt,
tmp_cxt;
bool skipNested = false;
JsonbIterator *it;
JsonbValue v;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (!JB_ROOT_IS_OBJECT(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a non-object",
funcname)));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
rsi->returnMode = SFRM_Materialize;
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("function returning record called in context "
"that cannot accept type record")));
old_cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
ret_tdesc = CreateTupleDescCopy(tupdesc);
BlessTupleDesc(ret_tdesc);
tuple_store =
tuplestore_begin_heap(rsi->allowedModes & SFRM_Materialize_Random,
false, work_mem);
MemoryContextSwitchTo(old_cxt);
tmp_cxt = AllocSetContextCreate(CurrentMemoryContext,
"jsonb_each temporary cxt",
Add macros to make AllocSetContextCreate() calls simpler and safer. I found that half a dozen (nearly 5%) of our AllocSetContextCreate calls had typos in the context-sizing parameters. While none of these led to especially significant problems, they did create minor inefficiencies, and it's now clear that expecting people to copy-and-paste those calls accurately is not a great idea. Let's reduce the risk of future errors by introducing single macros that encapsulate the common use-cases. Three such macros are enough to cover all but two special-purpose contexts; those two calls can be left as-is, I think. While this patch doesn't in itself improve matters for third-party extensions, it doesn't break anything for them either, and they can gradually adopt the simplified notation over time. In passing, change TopMemoryContext to use the default allocation parameters. Formerly it could only be extended 8K at a time. That was probably reasonable when this code was written; but nowadays we create many more contexts than we did then, so that it's not unusual to have a couple hundred K in TopMemoryContext, even without considering various dubious code that sticks other things there. There seems no good reason not to let it use growing blocks like most other contexts. Back-patch to 9.6, mostly because that's still close enough to HEAD that it's easy to do so, and keeping the branches in sync can be expected to avoid some future back-patching pain. The bugs fixed by these changes don't seem to be significant enough to justify fixing them further back. Discussion: <21072.1472321324@sss.pgh.pa.us>
2016-08-27 23:50:38 +02:00
ALLOCSET_DEFAULT_SIZES);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
it = JsonbIteratorInit(&jb->root);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != WJB_DONE)
{
skipNested = true;
if (r == WJB_KEY)
{
text *key;
HeapTuple tuple;
Datum values[2];
bool nulls[2] = {false, false};
/* Use the tmp context so we can clean up after each tuple is done */
old_cxt = MemoryContextSwitchTo(tmp_cxt);
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
key = cstring_to_text_with_len(v.val.string.val, v.val.string.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/*
* The next thing the iterator fetches should be the value, no
* matter what shape it is.
*/
r = JsonbIteratorNext(&it, &v, skipNested);
values[0] = PointerGetDatum(key);
if (as_text)
{
if (v.type == jbvNull)
{
/* a json null is an sql null in text mode */
nulls[1] = true;
values[1] = (Datum) NULL;
}
else
{
text *sv;
if (v.type == jbvString)
{
/* In text mode, scalar strings should be dequoted */
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
sv = cstring_to_text_with_len(v.val.string.val, v.val.string.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else
{
/* Turn anything else into a json string */
StringInfo jtext = makeStringInfo();
Jsonb *jb = JsonbValueToJsonb(&v);
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
(void) JsonbToCString(jtext, &jb->root, 0);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
sv = cstring_to_text_with_len(jtext->data, jtext->len);
}
values[1] = PointerGetDatum(sv);
}
}
else
{
/* Not in text mode, just return the Jsonb */
Jsonb *val = JsonbValueToJsonb(&v);
values[1] = PointerGetDatum(val);
}
tuple = heap_form_tuple(ret_tdesc, values, nulls);
tuplestore_puttuple(tuple_store, tuple);
/* clean up and switch back */
MemoryContextSwitchTo(old_cxt);
MemoryContextReset(tmp_cxt);
}
}
MemoryContextDelete(tmp_cxt);
rsi->setResult = tuple_store;
rsi->setDesc = ret_tdesc;
PG_RETURN_NULL();
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
each_worker(FunctionCallInfo fcinfo, bool as_text)
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *json = PG_GETARG_TEXT_P(0);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
JsonLexContext *lex;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonSemAction *sem;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ReturnSetInfo *rsi;
MemoryContext old_cxt;
TupleDesc tupdesc;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
EachState *state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
lex = makeJsonLexContext(json, true);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = palloc0(sizeof(EachState));
sem = palloc0(sizeof(JsonSemAction));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
rsi->returnMode = SFRM_Materialize;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
(void) get_call_result_type(fcinfo, NULL, &tupdesc);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* make these in a sufficiently long-lived memory context */
old_cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
state->ret_tdesc = CreateTupleDescCopy(tupdesc);
BlessTupleDesc(state->ret_tdesc);
state->tuple_store =
tuplestore_begin_heap(rsi->allowedModes & SFRM_Materialize_Random,
false, work_mem);
MemoryContextSwitchTo(old_cxt);
sem->semstate = (void *) state;
sem->array_start = each_array_start;
sem->scalar = each_scalar;
sem->object_field_start = each_object_field_start;
sem->object_field_end = each_object_field_end;
state->normalize_results = as_text;
state->next_scalar = false;
state->lex = lex;
state->tmp_cxt = AllocSetContextCreate(CurrentMemoryContext,
"json_each temporary cxt",
Add macros to make AllocSetContextCreate() calls simpler and safer. I found that half a dozen (nearly 5%) of our AllocSetContextCreate calls had typos in the context-sizing parameters. While none of these led to especially significant problems, they did create minor inefficiencies, and it's now clear that expecting people to copy-and-paste those calls accurately is not a great idea. Let's reduce the risk of future errors by introducing single macros that encapsulate the common use-cases. Three such macros are enough to cover all but two special-purpose contexts; those two calls can be left as-is, I think. While this patch doesn't in itself improve matters for third-party extensions, it doesn't break anything for them either, and they can gradually adopt the simplified notation over time. In passing, change TopMemoryContext to use the default allocation parameters. Formerly it could only be extended 8K at a time. That was probably reasonable when this code was written; but nowadays we create many more contexts than we did then, so that it's not unusual to have a couple hundred K in TopMemoryContext, even without considering various dubious code that sticks other things there. There seems no good reason not to let it use growing blocks like most other contexts. Back-patch to 9.6, mostly because that's still close enough to HEAD that it's easy to do so, and keeping the branches in sync can be expected to avoid some future back-patching pain. The bugs fixed by these changes don't seem to be significant enough to justify fixing them further back. Discussion: <21072.1472321324@sss.pgh.pa.us>
2016-08-27 23:50:38 +02:00
ALLOCSET_DEFAULT_SIZES);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
pg_parse_json(lex, sem);
Fix whitespace
2014-02-06 05:12:51 +01:00
MemoryContextDelete(state->tmp_cxt);
In json code, clean up temp memory contexts after processing. Craig Ringer.
2014-02-03 16:40:12 +01:00
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi->setResult = state->tuple_store;
rsi->setDesc = state->ret_tdesc;
PG_RETURN_NULL();
}
static void
each_object_field_start(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
EachState *_state = (EachState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* save a pointer to where the value starts */
if (_state->lex->lex_level == 1)
{
/*
* next_scalar will be reset in the object_field_end handler, and
* since we know the value is a scalar there is no danger of it being
* on while recursing down the tree.
*/
if (_state->normalize_results && _state->lex->token_type == JSON_TOKEN_STRING)
_state->next_scalar = true;
else
_state->result_start = _state->lex->token_start;
}
}
static void
each_object_field_end(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
EachState *_state = (EachState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
MemoryContext old_cxt;
int len;
text *val;
HeapTuple tuple;
Datum values[2];
bool nulls[2] = {false, false};
/* skip over nested objects */
if (_state->lex->lex_level != 1)
return;
/* use the tmp context so we can clean up after each tuple is done */
old_cxt = MemoryContextSwitchTo(_state->tmp_cxt);
values[0] = CStringGetTextDatum(fname);
if (isnull && _state->normalize_results)
{
nulls[1] = true;
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
values[1] = (Datum) 0;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
else if (_state->next_scalar)
{
values[1] = CStringGetTextDatum(_state->normalized_scalar);
_state->next_scalar = false;
}
else
{
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
len = _state->lex->prev_token_terminator - _state->result_start;
val = cstring_to_text_with_len(_state->result_start, len);
values[1] = PointerGetDatum(val);
}
tuple = heap_form_tuple(_state->ret_tdesc, values, nulls);
tuplestore_puttuple(_state->tuple_store, tuple);
/* clean up and switch back */
MemoryContextSwitchTo(old_cxt);
MemoryContextReset(_state->tmp_cxt);
}
static void
each_array_start(void *state)
{
EachState *_state = (EachState *) state;
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot deconstruct an array as an object")));
}
static void
each_scalar(void *state, char *token, JsonTokenType tokentype)
{
EachState *_state = (EachState *) state;
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot deconstruct a scalar")));
/* supply de-escaped value if required */
if (_state->next_scalar)
_state->normalized_scalar = token;
}
/*
* SQL functions json_array_elements and json_array_elements_text
*
* get the elements from a json array
*
* a lot of this processing is similar to the json_each* functions
*/
Datum
jsonb_array_elements(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return elements_worker_jsonb(fcinfo, "jsonb_array_elements", false);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Datum
jsonb_array_elements_text(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return elements_worker_jsonb(fcinfo, "jsonb_array_elements_text", true);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elements_worker_jsonb(FunctionCallInfo fcinfo, const char *funcname,
bool as_text)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Jsonb *jb = PG_GETARG_JSONB(0);
ReturnSetInfo *rsi;
Tuplestorestate *tuple_store;
TupleDesc tupdesc;
TupleDesc ret_tdesc;
MemoryContext old_cxt,
tmp_cxt;
bool skipNested = false;
JsonbIterator *it;
JsonbValue v;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (JB_ROOT_IS_SCALAR(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot extract elements from a scalar")));
else if (!JB_ROOT_IS_ARRAY(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot extract elements from an object")));
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
rsi->returnMode = SFRM_Materialize;
/* it's a simple type, so don't use get_call_result_type() */
tupdesc = rsi->expectedDesc;
old_cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
ret_tdesc = CreateTupleDescCopy(tupdesc);
BlessTupleDesc(ret_tdesc);
tuple_store =
tuplestore_begin_heap(rsi->allowedModes & SFRM_Materialize_Random,
false, work_mem);
MemoryContextSwitchTo(old_cxt);
tmp_cxt = AllocSetContextCreate(CurrentMemoryContext,
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
"jsonb_array_elements temporary cxt",
Add macros to make AllocSetContextCreate() calls simpler and safer. I found that half a dozen (nearly 5%) of our AllocSetContextCreate calls had typos in the context-sizing parameters. While none of these led to especially significant problems, they did create minor inefficiencies, and it's now clear that expecting people to copy-and-paste those calls accurately is not a great idea. Let's reduce the risk of future errors by introducing single macros that encapsulate the common use-cases. Three such macros are enough to cover all but two special-purpose contexts; those two calls can be left as-is, I think. While this patch doesn't in itself improve matters for third-party extensions, it doesn't break anything for them either, and they can gradually adopt the simplified notation over time. In passing, change TopMemoryContext to use the default allocation parameters. Formerly it could only be extended 8K at a time. That was probably reasonable when this code was written; but nowadays we create many more contexts than we did then, so that it's not unusual to have a couple hundred K in TopMemoryContext, even without considering various dubious code that sticks other things there. There seems no good reason not to let it use growing blocks like most other contexts. Back-patch to 9.6, mostly because that's still close enough to HEAD that it's easy to do so, and keeping the branches in sync can be expected to avoid some future back-patching pain. The bugs fixed by these changes don't seem to be significant enough to justify fixing them further back. Discussion: <21072.1472321324@sss.pgh.pa.us>
2016-08-27 23:50:38 +02:00
ALLOCSET_DEFAULT_SIZES);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
it = JsonbIteratorInit(&jb->root);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != WJB_DONE)
{
skipNested = true;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (r == WJB_ELEM)
{
HeapTuple tuple;
Datum values[1];
bool nulls[1] = {false};
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* use the tmp context so we can clean up after each tuple is done */
old_cxt = MemoryContextSwitchTo(tmp_cxt);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (!as_text)
{
Jsonb *val = JsonbValueToJsonb(&v);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
values[0] = PointerGetDatum(val);
}
else
{
if (v.type == jbvNull)
{
/* a json null is an sql null in text mode */
nulls[0] = true;
values[0] = (Datum) NULL;
}
else
{
text *sv;
if (v.type == jbvString)
{
/* in text mode scalar strings should be dequoted */
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
sv = cstring_to_text_with_len(v.val.string.val, v.val.string.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else
{
/* turn anything else into a json string */
StringInfo jtext = makeStringInfo();
Jsonb *jb = JsonbValueToJsonb(&v);
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
(void) JsonbToCString(jtext, &jb->root, 0);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
sv = cstring_to_text_with_len(jtext->data, jtext->len);
}
values[0] = PointerGetDatum(sv);
}
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
tuple = heap_form_tuple(ret_tdesc, values, nulls);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
tuplestore_puttuple(tuple_store, tuple);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* clean up and switch back */
MemoryContextSwitchTo(old_cxt);
MemoryContextReset(tmp_cxt);
}
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
MemoryContextDelete(tmp_cxt);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
rsi->setResult = tuple_store;
rsi->setDesc = ret_tdesc;
PG_RETURN_NULL();
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Datum
json_array_elements(PG_FUNCTION_ARGS)
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return elements_worker(fcinfo, "json_array_elements", false);
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
}
Datum
json_array_elements_text(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return elements_worker(fcinfo, "json_array_elements_text", true);
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elements_worker(FunctionCallInfo fcinfo, const char *funcname, bool as_text)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
text *json = PG_GETARG_TEXT_P(0);
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
/* elements only needs escaped strings when as_text */
JsonLexContext *lex = makeJsonLexContext(json, as_text);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonSemAction *sem;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ReturnSetInfo *rsi;
MemoryContext old_cxt;
TupleDesc tupdesc;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ElementsState *state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = palloc0(sizeof(ElementsState));
sem = palloc0(sizeof(JsonSemAction));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
rsi->returnMode = SFRM_Materialize;
/* it's a simple type, so don't use get_call_result_type() */
tupdesc = rsi->expectedDesc;
/* make these in a sufficiently long-lived memory context */
old_cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
state->ret_tdesc = CreateTupleDescCopy(tupdesc);
BlessTupleDesc(state->ret_tdesc);
state->tuple_store =
tuplestore_begin_heap(rsi->allowedModes & SFRM_Materialize_Random,
false, work_mem);
MemoryContextSwitchTo(old_cxt);
sem->semstate = (void *) state;
sem->object_start = elements_object_start;
sem->scalar = elements_scalar;
sem->array_element_start = elements_array_element_start;
sem->array_element_end = elements_array_element_end;
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
state->function_name = funcname;
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
state->normalize_results = as_text;
state->next_scalar = false;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->lex = lex;
state->tmp_cxt = AllocSetContextCreate(CurrentMemoryContext,
"json_array_elements temporary cxt",
Add macros to make AllocSetContextCreate() calls simpler and safer. I found that half a dozen (nearly 5%) of our AllocSetContextCreate calls had typos in the context-sizing parameters. While none of these led to especially significant problems, they did create minor inefficiencies, and it's now clear that expecting people to copy-and-paste those calls accurately is not a great idea. Let's reduce the risk of future errors by introducing single macros that encapsulate the common use-cases. Three such macros are enough to cover all but two special-purpose contexts; those two calls can be left as-is, I think. While this patch doesn't in itself improve matters for third-party extensions, it doesn't break anything for them either, and they can gradually adopt the simplified notation over time. In passing, change TopMemoryContext to use the default allocation parameters. Formerly it could only be extended 8K at a time. That was probably reasonable when this code was written; but nowadays we create many more contexts than we did then, so that it's not unusual to have a couple hundred K in TopMemoryContext, even without considering various dubious code that sticks other things there. There seems no good reason not to let it use growing blocks like most other contexts. Back-patch to 9.6, mostly because that's still close enough to HEAD that it's easy to do so, and keeping the branches in sync can be expected to avoid some future back-patching pain. The bugs fixed by these changes don't seem to be significant enough to justify fixing them further back. Discussion: <21072.1472321324@sss.pgh.pa.us>
2016-08-27 23:50:38 +02:00
ALLOCSET_DEFAULT_SIZES);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
pg_parse_json(lex, sem);
Fix whitespace
2014-02-06 05:12:51 +01:00
MemoryContextDelete(state->tmp_cxt);
In json code, clean up temp memory contexts after processing. Craig Ringer.
2014-02-03 16:40:12 +01:00
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi->setResult = state->tuple_store;
rsi->setDesc = state->ret_tdesc;
PG_RETURN_NULL();
}
static void
elements_array_element_start(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ElementsState *_state = (ElementsState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* save a pointer to where the value starts */
if (_state->lex->lex_level == 1)
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
{
/*
* next_scalar will be reset in the array_element_end handler, and
* since we know the value is a scalar there is no danger of it being
* on while recursing down the tree.
*/
if (_state->normalize_results && _state->lex->token_type == JSON_TOKEN_STRING)
_state->next_scalar = true;
else
_state->result_start = _state->lex->token_start;
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
elements_array_element_end(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ElementsState *_state = (ElementsState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
MemoryContext old_cxt;
int len;
text *val;
HeapTuple tuple;
Datum values[1];
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
bool nulls[1] = {false};
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* skip over nested objects */
if (_state->lex->lex_level != 1)
return;
/* use the tmp context so we can clean up after each tuple is done */
old_cxt = MemoryContextSwitchTo(_state->tmp_cxt);
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
if (isnull && _state->normalize_results)
{
nulls[0] = true;
values[0] = (Datum) NULL;
}
else if (_state->next_scalar)
{
values[0] = CStringGetTextDatum(_state->normalized_scalar);
_state->next_scalar = false;
}
else
{
len = _state->lex->prev_token_terminator - _state->result_start;
val = cstring_to_text_with_len(_state->result_start, len);
values[0] = PointerGetDatum(val);
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
tuple = heap_form_tuple(_state->ret_tdesc, values, nulls);
tuplestore_puttuple(_state->tuple_store, tuple);
/* clean up and switch back */
MemoryContextSwitchTo(old_cxt);
MemoryContextReset(_state->tmp_cxt);
}
static void
elements_object_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ElementsState *_state = (ElementsState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a non-array",
_state->function_name)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
elements_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ElementsState *_state = (ElementsState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* json structure check */
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a scalar",
_state->function_name)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Add json_array_elements_text function. This was a notable omission from the json functions added in 9.3 and there have been numerous complaints about its absence. Laurence Rowe.
2014-01-29 21:39:01 +01:00
/* supply de-escaped value if required */
if (_state->next_scalar)
_state->normalized_scalar = token;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
/*
* SQL function json_populate_record
*
* set fields in a record from the argument json
*
* Code adapted shamelessly from hstore's populate_record
* which is in turn partly adapted from record_out.
*
* The json is decomposed into a hash table, in which each
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* field in the record is then looked up by name. For jsonb
* we fetch the values direct from the object.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_populate_record(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_record_worker(fcinfo, "jsonb_populate_record", true);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
Datum
jsonb_to_record(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_record_worker(fcinfo, "jsonb_to_record", false);
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Datum
json_populate_record(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_record_worker(fcinfo, "json_populate_record", true);
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
}
Datum
json_to_record(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_record_worker(fcinfo, "json_to_record", false);
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
populate_record_worker(FunctionCallInfo fcinfo, const char *funcname,
bool have_record_arg)
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
{
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
int json_arg_num = have_record_arg ? 1 : 0;
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
Oid jtype = get_fn_expr_argtype(fcinfo->flinfo, json_arg_num);
Correct handling of NULL arguments in json funcs. Per gripe from Tom Lane.
2013-04-15 22:20:21 +02:00
text *json;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Jsonb *jb = NULL;
HTAB *json_hash = NULL;
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
HeapTupleHeader rec = NULL;
Silence compiler warnings about possibly unset variables. They are in fact set in every case where they are needed, but the compiler doesn't know that. Per gripe from Tom Lane.
2014-01-30 00:54:14 +01:00
Oid tupType = InvalidOid;
int32 tupTypmod = -1;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
TupleDesc tupdesc;
HeapTupleData tuple;
HeapTuple rettuple;
RecordIOData *my_extra;
int ncolumns;
int i;
Datum *values;
bool *nulls;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Assert(jtype == JSONOID || jtype == JSONBOID);
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (have_record_arg)
{
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
Oid argtype = get_fn_expr_argtype(fcinfo->flinfo, 0);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (!type_is_rowtype(argtype))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("first argument of %s must be a row type",
funcname)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (PG_ARGISNULL(0))
{
if (PG_ARGISNULL(1))
PG_RETURN_NULL();
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
/*
* have no tuple to look at, so the only source of type info is
* the argtype. The lookup_rowtype_tupdesc call below will error
* out if we don't have a known composite type oid here.
*/
tupType = argtype;
tupTypmod = -1;
}
else
{
rec = PG_GETARG_HEAPTUPLEHEADER(0);
if (PG_ARGISNULL(1))
PG_RETURN_POINTER(rec);
/* Extract type info from the tuple itself */
tupType = HeapTupleHeaderGetTypeId(rec);
tupTypmod = HeapTupleHeaderGetTypMod(rec);
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
tupdesc = lookup_rowtype_tupdesc(tupType, tupTypmod);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
else
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
{
/* json{b}_to_record case */
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (PG_ARGISNULL(0))
PG_RETURN_NULL();
Fix crash in json_to_record(). json_to_record() depends on get_call_result_type() for the tuple descriptor of the record that should be returned, but in some cases that cannot be determined. Add a guard to check if the tuple descriptor has been properly resolved, similar to other callers of get_call_result_type(). Also add guard for two other callers of get_call_result_type() in jsonfuncs.c. Although json_to_record() is the only actual bug, it's a good idea to follow convention.
2014-02-26 16:47:09 +01:00
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("function returning record called in context "
"that cannot accept type record"),
errhint("Try calling the function in the FROM clause "
"using a column definition list.")));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (jtype == JSONOID)
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
{
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* just get the text */
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
json = PG_GETARG_TEXT_P(json_arg_num);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Remove use_json_as_text options from json_to_record/json_populate_record. The "false" case was really quite useless since all it did was to throw an error; a definition not helped in the least by making it the default. Instead let's just have the "true" case, which emits nested objects and arrays in JSON syntax. We might later want to provide the ability to emit sub-objects in Postgres record or array syntax, but we'd be best off to drive that off a check of the target field datatype, not a separate argument. For the functions newly added in 9.4, we can just remove the flag arguments outright. We can't do that for json_populate_record[set], which already existed in 9.3, but we can ignore the argument and always behave as if it were "true". It helps that the flag arguments were optional and not documented in any useful fashion anyway.
2014-06-29 19:50:58 +02:00
json_hash = get_json_object_as_hash(json, funcname);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
/*
* if the input json is empty, we can only skip the rest if we were
* passed in a non-null record, since otherwise there may be issues
* with domain nulls.
*/
if (hash_get_num_entries(json_hash) == 0 && rec)
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
{
hash_destroy(json_hash);
Fix two small bugs in json's populate_record_worker The first bug is not releasing a tupdesc when doing an early return out of the function. The second bug is a logic error in choosing when to do an early return if given an empty jsonb object. Bug reports from Pavel Stehule and Tom Lane respectively. Backpatch to 9.4 where these were introduced.
2015-05-04 18:38:58 +02:00
ReleaseTupleDesc(tupdesc);
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
PG_RETURN_POINTER(rec);
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else
{
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
jb = PG_GETARG_JSONB(json_arg_num);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* same logic as for json */
Fix two small bugs in json's populate_record_worker The first bug is not releasing a tupdesc when doing an early return out of the function. The second bug is a logic error in choosing when to do an early return if given an empty jsonb object. Bug reports from Pavel Stehule and Tom Lane respectively. Backpatch to 9.4 where these were introduced.
2015-05-04 18:38:58 +02:00
if (JB_ROOT_COUNT(jb) == 0 && rec)
{
ReleaseTupleDesc(tupdesc);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
PG_RETURN_POINTER(rec);
Fix two small bugs in json's populate_record_worker The first bug is not releasing a tupdesc when doing an early return out of the function. The second bug is a logic error in choosing when to do an early return if given an empty jsonb object. Bug reports from Pavel Stehule and Tom Lane respectively. Backpatch to 9.4 where these were introduced.
2015-05-04 18:38:58 +02:00
}
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ncolumns = tupdesc->natts;
if (rec)
{
/* Build a temporary HeapTuple control structure */
tuple.t_len = HeapTupleHeaderGetDatumLength(rec);
ItemPointerSetInvalid(&(tuple.t_self));
tuple.t_tableOid = InvalidOid;
tuple.t_data = rec;
}
/*
* We arrange to look up the needed I/O info just once per series of
* calls, assuming the record type doesn't change underneath us.
*/
my_extra = (RecordIOData *) fcinfo->flinfo->fn_extra;
if (my_extra == NULL ||
my_extra->ncolumns != ncolumns)
{
fcinfo->flinfo->fn_extra =
MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
Use FLEXIBLE_ARRAY_MEMBER in struct RecordIOData. I (tgl) fixed this last night in rowtypes.c, but I missed that the code had been copied into a couple of other places. Michael Paquier
2015-02-20 23:03:12 +01:00
offsetof(RecordIOData, columns) +
ncolumns * sizeof(ColumnIOData));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
my_extra = (RecordIOData *) fcinfo->flinfo->fn_extra;
my_extra->record_type = InvalidOid;
my_extra->record_typmod = 0;
Fix uninitialized variables in json's populate_record_worker(). Peter Geoghegan.
2014-03-26 23:20:56 +01:00
my_extra->ncolumns = ncolumns;
MemSet(my_extra->columns, 0, sizeof(ColumnIOData) * ncolumns);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (have_record_arg && (my_extra->record_type != tupType ||
my_extra->record_typmod != tupTypmod))
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
MemSet(my_extra, 0,
Use FLEXIBLE_ARRAY_MEMBER in struct RecordIOData. I (tgl) fixed this last night in rowtypes.c, but I missed that the code had been copied into a couple of other places. Michael Paquier
2015-02-20 23:03:12 +01:00
offsetof(RecordIOData, columns) +
ncolumns * sizeof(ColumnIOData));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
my_extra->record_type = tupType;
my_extra->record_typmod = tupTypmod;
my_extra->ncolumns = ncolumns;
}
values = (Datum *) palloc(ncolumns * sizeof(Datum));
nulls = (bool *) palloc(ncolumns * sizeof(bool));
if (rec)
{
/* Break down the tuple into fields */
heap_deform_tuple(&tuple, tupdesc, values, nulls);
}
else
{
for (i = 0; i < ncolumns; ++i)
{
values[i] = (Datum) 0;
nulls[i] = true;
}
}
for (i = 0; i < ncolumns; ++i)
{
ColumnIOData *column_info = &my_extra->columns[i];
Oid column_type = tupdesc->attrs[i]->atttypid;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
JsonbValue *v = NULL;
JsonHashEntry *hashentry = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* Ignore dropped columns in datatype */
if (tupdesc->attrs[i]->attisdropped)
{
nulls[i] = true;
continue;
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (jtype == JSONOID)
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
hashentry = hash_search(json_hash,
NameStr(tupdesc->attrs[i]->attname),
HASH_FIND, NULL);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
else
{
char *key = NameStr(tupdesc->attrs[i]->attname);
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
v = findJsonbValueFromContainerLen(&jb->root, JB_FOBJECT, key,
strlen(key));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
* we can't just skip here if the key wasn't found since we might have
* a domain to deal with. If we were passed in a non-null record
* datum, we assume that the existing values are valid (if they're
* not, then it's not our fault), but if we were passed in a null,
* then every field which we don't populate needs to be run through
* the input function just in case it's a domain type.
*/
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (((jtype == JSONOID && hashentry == NULL) ||
(jtype == JSONBOID && v == NULL)) && rec)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
continue;
/*
* Prepare to convert the column value from text
*/
if (column_info->column_type != column_type)
{
getTypeInputInfo(column_type,
&column_info->typiofunc,
&column_info->typioparam);
fmgr_info_cxt(column_info->typiofunc, &column_info->proc,
fcinfo->flinfo->fn_mcxt);
column_info->column_type = column_type;
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if ((jtype == JSONOID && (hashentry == NULL || hashentry->isnull)) ||
(jtype == JSONBOID && (v == NULL || v->type == jbvNull)))
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
/*
* need InputFunctionCall to happen even for nulls, so that domain
* checks are done
*/
values[i] = InputFunctionCall(&column_info->proc, NULL,
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = true;
}
else
{
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
char *s = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (jtype == JSONOID)
{
/* already done the hard work in the json case */
s = hashentry->val;
}
else
{
if (v->type == jbvString)
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
s = pnstrdup(v->val.string.val, v->val.string.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvBool)
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
s = pnstrdup((v->val.boolean) ? "t" : "f", 1);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvNumeric)
s = DatumGetCString(DirectFunctionCall1(numeric_out,
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
PointerGetDatum(v->val.numeric)));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvBinary)
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
s = JsonbToCString(NULL, (JsonbContainer *) v->val.binary.data, v->val.binary.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elog(ERROR, "unrecognized jsonb type: %d", (int) v->type);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
values[i] = InputFunctionCall(&column_info->proc, s,
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = false;
}
}
rettuple = heap_form_tuple(tupdesc, values, nulls);
ReleaseTupleDesc(tupdesc);
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
if (json_hash)
hash_destroy(json_hash);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
PG_RETURN_DATUM(HeapTupleGetDatum(rettuple));
}
/*
* get_json_object_as_hash
*
* decompose a json object into a hash table.
*/
static HTAB *
Remove use_json_as_text options from json_to_record/json_populate_record. The "false" case was really quite useless since all it did was to throw an error; a definition not helped in the least by making it the default. Instead let's just have the "true" case, which emits nested objects and arrays in JSON syntax. We might later want to provide the ability to emit sub-objects in Postgres record or array syntax, but we'd be best off to drive that off a check of the target field datatype, not a separate argument. For the functions newly added in 9.4, we can just remove the flag arguments outright. We can't do that for json_populate_record[set], which already existed in 9.3, but we can ignore the argument and always behave as if it were "true". It helps that the flag arguments were optional and not documented in any useful fashion anyway.
2014-06-29 19:50:58 +02:00
get_json_object_as_hash(text *json, const char *funcname)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
{
HASHCTL ctl;
HTAB *tab;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JHashState *state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
JsonLexContext *lex = makeJsonLexContext(json, true);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonSemAction *sem;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
memset(&ctl, 0, sizeof(ctl));
ctl.keysize = NAMEDATALEN;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ctl.entrysize = sizeof(JsonHashEntry);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ctl.hcxt = CurrentMemoryContext;
tab = hash_create("json object hashtable",
100,
&ctl,
HASH_ELEM | HASH_CONTEXT);
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
state = palloc0(sizeof(JHashState));
sem = palloc0(sizeof(JsonSemAction));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->function_name = funcname;
state->hash = tab;
state->lex = lex;
sem->semstate = (void *) state;
sem->array_start = hash_array_start;
sem->scalar = hash_scalar;
sem->object_field_start = hash_object_field_start;
sem->object_field_end = hash_object_field_end;
pg_parse_json(lex, sem);
return tab;
}
static void
hash_object_field_start(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JHashState *_state = (JHashState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level > 1)
return;
if (_state->lex->token_type == JSON_TOKEN_ARRAY_START ||
_state->lex->token_type == JSON_TOKEN_OBJECT_START)
{
Remove use_json_as_text options from json_to_record/json_populate_record. The "false" case was really quite useless since all it did was to throw an error; a definition not helped in the least by making it the default. Instead let's just have the "true" case, which emits nested objects and arrays in JSON syntax. We might later want to provide the ability to emit sub-objects in Postgres record or array syntax, but we'd be best off to drive that off a check of the target field datatype, not a separate argument. For the functions newly added in 9.4, we can just remove the flag arguments outright. We can't do that for json_populate_record[set], which already existed in 9.3, but we can ignore the argument and always behave as if it were "true". It helps that the flag arguments were optional and not documented in any useful fashion anyway.
2014-06-29 19:50:58 +02:00
/* remember start position of the whole text of the subobject */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
_state->save_json_start = _state->lex->token_start;
}
else
{
/* must be a scalar */
_state->save_json_start = NULL;
}
}
static void
hash_object_field_end(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JHashState *_state = (JHashState *) state;
JsonHashEntry *hashentry;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool found;
/*
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
* Ignore nested fields.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Fix json_to_record() bug with nested objects. A thinko concerning nesting depth caused json_to_record() to produce bogus output if a field of its input object contained a sub-object with a field name matching one of the requested output column names. Per bug #13996 from Johann Visagie. I added a regression test case based on his example, plus parallel tests for json_to_recordset, jsonb_to_record, jsonb_to_recordset. The latter three do not exhibit the same bug (which suggests that we may be missing some opportunities to share code...) but testing seems like a good idea in any case. Back-patch to 9.4 where these functions were introduced.
2016-03-03 05:31:39 +01:00
if (_state->lex->lex_level > 1)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
return;
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
/*
* Ignore field names >= NAMEDATALEN - they can't match a record field.
* (Note: without this test, the hash code would truncate the string at
* NAMEDATALEN-1, and could then match against a similarly-truncated
* record field name. That would be a reasonable behavior, but this code
* has previously insisted on exact equality, so we keep this behavior.)
*/
if (strlen(fname) >= NAMEDATALEN)
return;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
hashentry = hash_search(_state->hash, fname, HASH_ENTER, &found);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
* found being true indicates a duplicate. We don't do anything about
* that, a later field with the same name overrides the earlier field.
*/
hashentry->isnull = isnull;
if (_state->save_json_start != NULL)
{
int len = _state->lex->prev_token_terminator - _state->save_json_start;
char *val = palloc((len + 1) * sizeof(char));
memcpy(val, _state->save_json_start, len);
val[len] = '\0';
hashentry->val = val;
}
else
{
/* must have had a scalar instead */
hashentry->val = _state->saved_scalar;
}
}
static void
hash_array_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JHashState *_state = (JHashState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot call %s on an array", _state->function_name)));
}
static void
hash_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JHashState *_state = (JHashState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot call %s on a scalar", _state->function_name)));
if (_state->lex->lex_level == 1)
_state->saved_scalar = token;
}
/*
* SQL function json_populate_recordset
*
* set fields in a set of records from the argument json,
* which must be an array of objects.
*
* similar to json_populate_record, but the tuple-building code
* is pushed down into the semantic action handlers so it's done
* per object in the array.
*/
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Datum
jsonb_populate_recordset(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_recordset_worker(fcinfo, "jsonb_populate_recordset", true);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
Datum
jsonb_to_recordset(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_recordset_worker(fcinfo, "jsonb_to_recordset", false);
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
}
Datum
json_populate_recordset(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_recordset_worker(fcinfo, "json_populate_recordset", true);
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
}
Datum
json_to_recordset(PG_FUNCTION_ARGS)
{
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
return populate_recordset_worker(fcinfo, "json_to_recordset", false);
Cleanup around json_to_record/json_to_recordset Set function parameter names and defaults. Add jsonb versions (which the code already provided for so the actual new code is trivial). Add jsonb regression tests and docs. Bump catalog version (which I apparently forgot to do when jsonb was committed).
2014-03-26 15:18:24 +01:00
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
static void
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
make_row_from_rec_and_jsonb(Jsonb *element, PopulateRecordsetState *state)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
Datum *values;
bool *nulls;
int i;
RecordIOData *my_extra = state->my_extra;
int ncolumns = my_extra->ncolumns;
TupleDesc tupdesc = state->ret_tdesc;
HeapTupleHeader rec = state->rec;
HeapTuple rettuple;
values = (Datum *) palloc(ncolumns * sizeof(Datum));
nulls = (bool *) palloc(ncolumns * sizeof(bool));
if (state->rec)
{
HeapTupleData tuple;
/* Build a temporary HeapTuple control structure */
tuple.t_len = HeapTupleHeaderGetDatumLength(state->rec);
ItemPointerSetInvalid(&(tuple.t_self));
tuple.t_tableOid = InvalidOid;
tuple.t_data = state->rec;
/* Break down the tuple into fields */
heap_deform_tuple(&tuple, tupdesc, values, nulls);
}
else
{
for (i = 0; i < ncolumns; ++i)
{
values[i] = (Datum) 0;
nulls[i] = true;
}
}
for (i = 0; i < ncolumns; ++i)
{
ColumnIOData *column_info = &my_extra->columns[i];
Oid column_type = tupdesc->attrs[i]->atttypid;
JsonbValue *v = NULL;
char *key;
/* Ignore dropped columns in datatype */
if (tupdesc->attrs[i]->attisdropped)
{
nulls[i] = true;
continue;
}
key = NameStr(tupdesc->attrs[i]->attname);
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
v = findJsonbValueFromContainerLen(&element->root, JB_FOBJECT,
key, strlen(key));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/*
* We can't just skip here if the key wasn't found since we might have
* a domain to deal with. If we were passed in a non-null record
* datum, we assume that the existing values are valid (if they're
* not, then it's not our fault), but if we were passed in a null,
* then every field which we don't populate needs to be run through
* the input function just in case it's a domain type.
*/
if (v == NULL && rec)
continue;
/*
* Prepare to convert the column value from text
*/
if (column_info->column_type != column_type)
{
getTypeInputInfo(column_type,
&column_info->typiofunc,
&column_info->typioparam);
fmgr_info_cxt(column_info->typiofunc, &column_info->proc,
state->fn_mcxt);
column_info->column_type = column_type;
}
if (v == NULL || v->type == jbvNull)
{
/*
* Need InputFunctionCall to happen even for nulls, so that domain
* checks are done
*/
values[i] = InputFunctionCall(&column_info->proc, NULL,
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = true;
}
else
{
char *s = NULL;
if (v->type == jbvString)
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
s = pnstrdup(v->val.string.val, v->val.string.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvBool)
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
s = pnstrdup((v->val.boolean) ? "t" : "f", 1);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvNumeric)
s = DatumGetCString(DirectFunctionCall1(numeric_out,
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
PointerGetDatum(v->val.numeric)));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else if (v->type == jbvBinary)
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
s = JsonbToCString(NULL, (JsonbContainer *) v->val.binary.data, v->val.binary.len);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
else
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
elog(ERROR, "unrecognized jsonb type: %d", (int) v->type);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
values[i] = InputFunctionCall(&column_info->proc, s,
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = false;
}
}
rettuple = heap_form_tuple(tupdesc, values, nulls);
tuplestore_puttuple(state->tuple_store, rettuple);
}
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
/*
* common worker for json_populate_recordset() and json_to_recordset()
*/
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
static Datum
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
populate_recordset_worker(FunctionCallInfo fcinfo, const char *funcname,
bool have_record_arg)
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
{
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
int json_arg_num = have_record_arg ? 1 : 0;
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
Oid jtype = get_fn_expr_argtype(fcinfo->flinfo, json_arg_num);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ReturnSetInfo *rsi;
MemoryContext old_cxt;
Oid tupType;
int32 tupTypmod;
HeapTupleHeader rec;
TupleDesc tupdesc;
RecordIOData *my_extra;
int ncolumns;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (have_record_arg)
{
pgindent run for 9.4 This includes removing tabs after periods in C comments, which was applied to back branches, so this change should not effect backpatching.
2014-05-06 18:12:18 +02:00
Oid argtype = get_fn_expr_argtype(fcinfo->flinfo, 0);
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
if (!type_is_rowtype(argtype))
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("first argument of %s must be a row type",
funcname)));
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi = (ReturnSetInfo *) fcinfo->resultinfo;
if (!rsi || !IsA(rsi, ReturnSetInfo) ||
(rsi->allowedModes & SFRM_Materialize) == 0 ||
rsi->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that "
"cannot accept a set")));
rsi->returnMode = SFRM_Materialize;
/*
* get the tupdesc from the result set info - it must be a record type
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
* because we already checked that arg1 is a record type, or we're in a
* to_record function which returns a setof record.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Fix crash in json_to_record(). json_to_record() depends on get_call_result_type() for the tuple descriptor of the record that should be returned, but in some cases that cannot be determined. Add a guard to check if the tuple descriptor has been properly resolved, similar to other callers of get_call_result_type(). Also add guard for two other callers of get_call_result_type() in jsonfuncs.c. Although json_to_record() is the only actual bug, it's a good idea to follow convention.
2014-02-26 16:47:09 +01:00
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("function returning record called in context "
"that cannot accept type record")));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/* if the json is null send back an empty set */
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
if (PG_ARGISNULL(json_arg_num))
PG_RETURN_NULL();
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
if (!have_record_arg || PG_ARGISNULL(0))
New json functions. json_build_array() and json_build_object allow for the construction of arbitrarily complex json trees. json_object() turns a one or two dimensional array, or two separate arrays, into a json_object of name/value pairs, similarly to the hstore() function. json_object_agg() aggregates its two arguments into a single json object as name value pairs. Catalog version bumped. Andrew Dunstan, reviewed by Marko Tiikkaja.
2014-01-28 23:48:21 +01:00
rec = NULL;
Tidy up the populate/to_record{set} code for json a bit. In the process fix a small bug.
2014-03-26 02:20:54 +01:00
else
rec = PG_GETARG_HEAPTUPLEHEADER(0);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
tupType = tupdesc->tdtypeid;
tupTypmod = tupdesc->tdtypmod;
ncolumns = tupdesc->natts;
/*
* We arrange to look up the needed I/O info just once per series of
* calls, assuming the record type doesn't change underneath us.
*/
my_extra = (RecordIOData *) fcinfo->flinfo->fn_extra;
if (my_extra == NULL ||
my_extra->ncolumns != ncolumns)
{
fcinfo->flinfo->fn_extra =
MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
Use FLEXIBLE_ARRAY_MEMBER in struct RecordIOData. I (tgl) fixed this last night in rowtypes.c, but I missed that the code had been copied into a couple of other places. Michael Paquier
2015-02-20 23:03:12 +01:00
offsetof(RecordIOData, columns) +
ncolumns * sizeof(ColumnIOData));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
my_extra = (RecordIOData *) fcinfo->flinfo->fn_extra;
my_extra->record_type = InvalidOid;
my_extra->record_typmod = 0;
}
if (my_extra->record_type != tupType ||
my_extra->record_typmod != tupTypmod)
{
MemSet(my_extra, 0,
Use FLEXIBLE_ARRAY_MEMBER in struct RecordIOData. I (tgl) fixed this last night in rowtypes.c, but I missed that the code had been copied into a couple of other places. Michael Paquier
2015-02-20 23:03:12 +01:00
offsetof(RecordIOData, columns) +
ncolumns * sizeof(ColumnIOData));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
my_extra->record_type = tupType;
my_extra->record_typmod = tupTypmod;
my_extra->ncolumns = ncolumns;
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
state = palloc0(sizeof(PopulateRecordsetState));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/* make these in a sufficiently long-lived memory context */
old_cxt = MemoryContextSwitchTo(rsi->econtext->ecxt_per_query_memory);
Remove spurious semicolons. Petr Jelinek
2015-03-31 14:12:27 +02:00
state->ret_tdesc = CreateTupleDescCopy(tupdesc);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
BlessTupleDesc(state->ret_tdesc);
state->tuple_store = tuplestore_begin_heap(rsi->allowedModes &
SFRM_Materialize_Random,
false, work_mem);
MemoryContextSwitchTo(old_cxt);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
state->function_name = funcname;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
state->my_extra = my_extra;
state->rec = rec;
state->fn_mcxt = fcinfo->flinfo->fn_mcxt;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
if (jtype == JSONOID)
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
text *json = PG_GETARG_TEXT_P(json_arg_num);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
JsonLexContext *lex;
JsonSemAction *sem;
sem = palloc0(sizeof(JsonSemAction));
lex = makeJsonLexContext(json, true);
sem->semstate = (void *) state;
sem->array_start = populate_recordset_array_start;
sem->array_element_start = populate_recordset_array_element_start;
sem->scalar = populate_recordset_scalar;
sem->object_field_start = populate_recordset_object_field_start;
sem->object_field_end = populate_recordset_object_field_end;
sem->object_start = populate_recordset_object_start;
sem->object_end = populate_recordset_object_end;
state->lex = lex;
pg_parse_json(lex, sem);
}
else
{
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
Jsonb *jb = PG_GETARG_JSONB(json_arg_num);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
JsonbIterator *it;
JsonbValue v;
bool skipNested = false;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Assert(jtype == JSONBOID);
if (JB_ROOT_IS_SCALAR(jb) || !JB_ROOT_IS_ARRAY(jb))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a non-array",
funcname)));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
it = JsonbIteratorInit(&jb->root);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != WJB_DONE)
{
skipNested = true;
if (r == WJB_ELEM)
{
Jsonb *element = JsonbValueToJsonb(&v);
if (!JB_ROOT_IS_OBJECT(element))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("argument of %s must be an array of objects",
funcname)));
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
make_row_from_rec_and_jsonb(element, state);
}
}
}
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
rsi->setResult = state->tuple_store;
rsi->setDesc = state->ret_tdesc;
PG_RETURN_NULL();
}
static void
populate_recordset_object_start(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
int lex_level = _state->lex->lex_level;
HASHCTL ctl;
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
/* Reject object at top level: we must have an array at level 0 */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on an object",
_state->function_name)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Remove use_json_as_text options from json_to_record/json_populate_record. The "false" case was really quite useless since all it did was to throw an error; a definition not helped in the least by making it the default. Instead let's just have the "true" case, which emits nested objects and arrays in JSON syntax. We might later want to provide the ability to emit sub-objects in Postgres record or array syntax, but we'd be best off to drive that off a check of the target field datatype, not a separate argument. For the functions newly added in 9.4, we can just remove the flag arguments outright. We can't do that for json_populate_record[set], which already existed in 9.3, but we can ignore the argument and always behave as if it were "true". It helps that the flag arguments were optional and not documented in any useful fashion anyway.
2014-06-29 19:50:58 +02:00
/* Nested objects require no special processing */
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
if (lex_level > 1)
return;
/* Object at level 1: set up a new hash table for this object */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
memset(&ctl, 0, sizeof(ctl));
ctl.keysize = NAMEDATALEN;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
ctl.entrysize = sizeof(JsonHashEntry);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
ctl.hcxt = CurrentMemoryContext;
_state->json_hash = hash_create("json object hashtable",
100,
&ctl,
HASH_ELEM | HASH_CONTEXT);
}
static void
populate_recordset_object_end(void *state)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
HTAB *json_hash = _state->json_hash;
Datum *values;
bool *nulls;
int i;
RecordIOData *my_extra = _state->my_extra;
int ncolumns = my_extra->ncolumns;
TupleDesc tupdesc = _state->ret_tdesc;
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
JsonHashEntry *hashentry;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
HeapTupleHeader rec = _state->rec;
HeapTuple rettuple;
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
/* Nested objects require no special processing */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level > 1)
return;
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
/* Otherwise, construct and return a tuple based on this level-1 object */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
values = (Datum *) palloc(ncolumns * sizeof(Datum));
nulls = (bool *) palloc(ncolumns * sizeof(bool));
if (_state->rec)
{
HeapTupleData tuple;
/* Build a temporary HeapTuple control structure */
tuple.t_len = HeapTupleHeaderGetDatumLength(_state->rec);
ItemPointerSetInvalid(&(tuple.t_self));
tuple.t_tableOid = InvalidOid;
tuple.t_data = _state->rec;
/* Break down the tuple into fields */
heap_deform_tuple(&tuple, tupdesc, values, nulls);
}
else
{
for (i = 0; i < ncolumns; ++i)
{
values[i] = (Datum) 0;
nulls[i] = true;
}
}
for (i = 0; i < ncolumns; ++i)
{
ColumnIOData *column_info = &my_extra->columns[i];
Oid column_type = tupdesc->attrs[i]->atttypid;
char *value;
/* Ignore dropped columns in datatype */
if (tupdesc->attrs[i]->attisdropped)
{
nulls[i] = true;
continue;
}
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
hashentry = hash_search(json_hash,
NameStr(tupdesc->attrs[i]->attname),
HASH_FIND, NULL);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
* we can't just skip here if the key wasn't found since we might have
* a domain to deal with. If we were passed in a non-null record
* datum, we assume that the existing values are valid (if they're
* not, then it's not our fault), but if we were passed in a null,
* then every field which we don't populate needs to be run through
* the input function just in case it's a domain type.
*/
if (hashentry == NULL && rec)
continue;
/*
* Prepare to convert the column value from text
*/
if (column_info->column_type != column_type)
{
getTypeInputInfo(column_type,
&column_info->typiofunc,
&column_info->typioparam);
fmgr_info_cxt(column_info->typiofunc, &column_info->proc,
_state->fn_mcxt);
column_info->column_type = column_type;
}
if (hashentry == NULL || hashentry->isnull)
{
/*
* need InputFunctionCall to happen even for nulls, so that domain
* checks are done
*/
values[i] = InputFunctionCall(&column_info->proc, NULL,
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = true;
}
else
{
value = hashentry->val;
values[i] = InputFunctionCall(&column_info->proc, value,
column_info->typioparam,
tupdesc->attrs[i]->atttypmod);
nulls[i] = false;
}
}
rettuple = heap_form_tuple(tupdesc, values, nulls);
tuplestore_puttuple(_state->tuple_store, rettuple);
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
/* Done with hash for this object */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
hash_destroy(json_hash);
Fix handling of nested JSON objects in json_populate_recordset and friends. populate_recordset_object_start() improperly created a new hash table (overwriting the link to the existing one) if called at nest levels greater than one. This resulted in previous fields not appearing in the final output, as reported by Matti Hameister in bug #10728. In 9.4 the problem also affects json_to_recordset. This perhaps missed detection earlier because the default behavior is to throw an error for nested objects: you have to pass use_json_as_text = true to see the problem. In addition, fix query-lifespan leakage of the hashtable created by json_populate_record(). This is pretty much the same problem recently fixed in dblink: creating an intended-to-be-temporary context underneath the executor's per-tuple context isn't enough to make it go away at the end of the tuple cycle, because MemoryContextReset is not MemoryContextResetAndDeleteChildren. Michael Paquier and Tom Lane
2014-06-25 06:22:40 +02:00
_state->json_hash = NULL;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
populate_recordset_array_element_start(void *state, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level == 1 &&
_state->lex->token_type != JSON_TOKEN_OBJECT_START)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("argument of %s must be an array of objects",
_state->function_name)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
populate_recordset_array_start(void *state)
{
Remove use_json_as_text options from json_to_record/json_populate_record. The "false" case was really quite useless since all it did was to throw an error; a definition not helped in the least by making it the default. Instead let's just have the "true" case, which emits nested objects and arrays in JSON syntax. We might later want to provide the ability to emit sub-objects in Postgres record or array syntax, but we'd be best off to drive that off a check of the target field datatype, not a separate argument. For the functions newly added in 9.4, we can just remove the flag arguments outright. We can't do that for json_populate_record[set], which already existed in 9.3, but we can ignore the argument and always behave as if it were "true". It helps that the flag arguments were optional and not documented in any useful fashion anyway.
2014-06-29 19:50:58 +02:00
/* nothing to do */
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
}
static void
populate_recordset_scalar(void *state, char *token, JsonTokenType tokentype)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level == 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rationalize error messages within jsonfuncs.c. I noticed that the functions in jsonfuncs.c sometimes printed error messages that claimed I'd called some other function. Investigation showed that this was from repurposing code into "worker" functions without taking much care as to whether it would mention the right SQL-level function if it threw an error. Moreover, there was a weird mismash of messages that contained a fixed function name, messages that used %s for a function name, and messages that constructed a function name out of spare parts, like "json%s_populate_record" (which, quite aside from being ugly as sin, wasn't even sufficient to cover all the cases). This would put an undue burden on our long-suffering translators. Standardize on inserting the SQL function name with %s so as to reduce the number of translatable strings, and pass function names around as needed to make sure we can report the right one. Fix up some gratuitous variations in wording, too.
2014-06-26 00:25:22 +02:00
errmsg("cannot call %s on a scalar",
_state->function_name)));
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level == 2)
_state->saved_scalar = token;
}
static void
populate_recordset_object_field_start(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
if (_state->lex->lex_level > 2)
return;
if (_state->lex->token_type == JSON_TOKEN_ARRAY_START ||
_state->lex->token_type == JSON_TOKEN_OBJECT_START)
{
_state->save_json_start = _state->lex->token_start;
}
else
{
_state->save_json_start = NULL;
}
}
static void
populate_recordset_object_field_end(void *state, char *fname, bool isnull)
{
Clean up new JSON API typedefs The new JSON API uses a bit of an unusual typedef scheme, where for example OkeysState is a pointer to okeysState. And that's not applied consistently either. Change that to the more usual PostgreSQL style where struct typedefs are upper case, and use pointers explicitly.
2013-07-20 12:38:31 +02:00
PopulateRecordsetState *_state = (PopulateRecordsetState *) state;
JsonHashEntry *hashentry;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
bool found;
/*
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
* Ignore nested fields.
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
*/
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
if (_state->lex->lex_level > 2)
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
return;
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
/*
* Ignore field names >= NAMEDATALEN - they can't match a record field.
* (Note: without this test, the hash code would truncate the string at
* NAMEDATALEN-1, and could then match against a similarly-truncated
* record field name. That would be a reasonable behavior, but this code
* has previously insisted on exact equality, so we keep this behavior.)
*/
if (strlen(fname) >= NAMEDATALEN)
return;
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
Cosmetic improvements in jsonfuncs.c. Re-pgindent, remove a lot of random vertical whitespace, remove useless (if not counterproductive) inline markings, get rid of unnecessary zero-padding of strings for hashtable searches. No functional changes.
2014-06-25 20:22:18 +02:00
hashentry = hash_search(_state->json_hash, fname, HASH_ENTER, &found);
Add new JSON processing functions and parser API. The JSON parser is converted into a recursive descent parser, and exposed for use by other modules such as extensions. The API provides hooks for all the significant parser event such as the beginning and end of objects and arrays, and providing functions to handle these hooks allows for fairly simple construction of a wide variety of JSON processing functions. A set of new basic processing functions and operators is also added, which use this API, including operations to extract array elements, object fields, get the length of arrays and the set of keys of a field, deconstruct an object into a set of key/value pairs, and create records from JSON objects and arrays of objects. Catalog version bumped. Andrew Dunstan, with some documentation assistance from Merlin Moncure.
2013-03-29 19:12:13 +01:00
/*
* found being true indicates a duplicate. We don't do anything about
* that, a later field with the same name overrides the earlier field.
*/
hashentry->isnull = isnull;
if (_state->save_json_start != NULL)
{
int len = _state->lex->prev_token_terminator - _state->save_json_start;
char *val = palloc((len + 1) * sizeof(char));
memcpy(val, _state->save_json_start, len);
val[len] = '\0';
hashentry->val = val;
}
else
{
/* must have had a scalar instead */
hashentry->val = _state->saved_scalar;
}
}
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
/*
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
* findJsonbValueFromContainer() wrapper that sets up JsonbValue key string.
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
*/
static JsonbValue *
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
findJsonbValueFromContainerLen(JsonbContainer *container, uint32 flags,
char *key, uint32 keylen)
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
{
JsonbValue k;
k.type = jbvString;
De-anonymize the union in JsonbValue. Needed for strict C89 compliance.
2014-04-02 20:30:08 +02:00
k.val.string.val = key;
k.val.string.len = keylen;
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
Clean up jsonb code. The main target of this cleanup is the convertJsonb() function, but I also touched a lot of other things that I spotted into in the process. The new convertToJsonb() function uses an output buffer that's resized on demand, so the code to estimate of the size of JsonbValue is removed. The on-disk format was not changed, even though I refactored the structs used to handle it. The term "superheader" is replaced with "container". The jsonb_exists_any and jsonb_exists_all functions no longer sort the input array. That was a premature optimization, the idea being that if there are duplicates in the input array, you only need to check them once. Also, sorting the array saves some effort in the binary search used to find a key within an object. But there were drawbacks too: the sorting and deduplicating obviously isn't free, and in the typical case there are no duplicates to remove, and the gain in the binary search was minimal. Remove all that, which makes the code simpler too. This includes a bug-fix; the total length of the elements in a jsonb array or object mustn't exceed 2^28. That is now checked.
2014-05-07 22:16:19 +02:00
return findJsonbValueFromContainer(container, flags, &k);
Introduce jsonb, a structured format for storing json. The new format accepts exactly the same data as the json type. However, it is stored in a format that does not require reparsing the orgiginal text in order to process it, making it much more suitable for indexing and other operations. Insignificant whitespace is discarded, and the order of object keys is not preserved. Neither are duplicate object keys kept - the later value for a given key is the only one stored. The new type has all the functions and operators that the json type has, with the exception of the json generation functions (to_json, json_agg etc.) and with identical semantics. In addition, there are operator classes for hash and btree indexing, and two classes for GIN indexing, that have no equivalent in the json type. This feature grew out of previous work by Oleg Bartunov and Teodor Sigaev, which was intended to provide similar facilities to a nested hstore type, but which in the end proved to have some significant compatibility issues. Authors: Oleg Bartunov, Teodor Sigaev, Peter Geoghegan and Andrew Dunstan. Review: Andres Freund
2014-03-23 21:40:19 +01:00
}
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
/*
* Semantic actions for json_strip_nulls.
*
* Simply repeat the input on the output unless we encounter
* a null object field. State for this is set when the field
* is started and reset when the scalar action (which must be next)
* is called.
*/
static void
sn_object_start(void *state)
{
StripnullState *_state = (StripnullState *) state;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
appendStringInfoCharMacro(_state->strval, '{');
}
static void
sn_object_end(void *state)
{
StripnullState *_state = (StripnullState *) state;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
appendStringInfoCharMacro(_state->strval, '}');
}
static void
sn_array_start(void *state)
{
StripnullState *_state = (StripnullState *) state;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
appendStringInfoCharMacro(_state->strval, '[');
}
static void
sn_array_end(void *state)
{
StripnullState *_state = (StripnullState *) state;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
appendStringInfoCharMacro(_state->strval, ']');
}
static void
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
sn_object_field_start(void *state, char *fname, bool isnull)
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
{
StripnullState *_state = (StripnullState *) state;
if (isnull)
{
/*
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
* The next thing must be a scalar or isnull couldn't be true, so
* there is no danger of this state being carried down into a nested
* object or array. The flag will be reset in the scalar action.
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
*/
_state->skip_next_null = true;
return;
}
if (_state->strval->data[_state->strval->len - 1] != '{')
appendStringInfoCharMacro(_state->strval, ',');
/*
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
* Unfortunately we don't have the quoted and escaped string any more, so
* we have to re-escape it.
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
*/
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
escape_json(_state->strval, fname);
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
appendStringInfoCharMacro(_state->strval, ':');
}
static void
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
sn_array_element_start(void *state, bool isnull)
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
{
StripnullState *_state = (StripnullState *) state;
if (_state->strval->data[_state->strval->len - 1] != '[')
appendStringInfoCharMacro(_state->strval, ',');
}
static void
sn_scalar(void *state, char *token, JsonTokenType tokentype)
{
StripnullState *_state = (StripnullState *) state;
if (_state->skip_next_null)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Assert(tokentype == JSON_TOKEN_NULL);
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
_state->skip_next_null = false;
return;
}
if (tokentype == JSON_TOKEN_STRING)
escape_json(_state->strval, token);
else
appendStringInfoString(_state->strval, token);
}
/*
* SQL function json_strip_nulls(json) -> json
*/
Datum
json_strip_nulls(PG_FUNCTION_ARGS)
{
text *json = PG_GETARG_TEXT_P(0);
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
StripnullState *state;
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
JsonLexContext *lex;
JsonSemAction *sem;
lex = makeJsonLexContext(json, true);
state = palloc0(sizeof(StripnullState));
sem = palloc0(sizeof(JsonSemAction));
state->strval = makeStringInfo();
state->skip_next_null = false;
state->lex = lex;
sem->semstate = (void *) state;
sem->object_start = sn_object_start;
sem->object_end = sn_object_end;
sem->array_start = sn_array_start;
sem->array_end = sn_array_end;
sem->scalar = sn_scalar;
sem->array_element_start = sn_array_element_start;
sem->object_field_start = sn_object_field_start;
pg_parse_json(lex, sem);
PG_RETURN_TEXT_P(cstring_to_text_with_len(state->strval->data,
state->strval->len));
}
/*
* SQL function jsonb_strip_nulls(jsonb) -> jsonb
*/
Datum
jsonb_strip_nulls(PG_FUNCTION_ARGS)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Jsonb *jb = PG_GETARG_JSONB(0);
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
JsonbIterator *it;
JsonbParseState *parseState = NULL;
JsonbValue *res = NULL;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
JsonbValue v,
k;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken type;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
bool last_was_key = false;
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
if (JB_ROOT_IS_SCALAR(jb))
PG_RETURN_POINTER(jb);
it = JsonbIteratorInit(&jb->root);
while ((type = JsonbIteratorNext(&it, &v, false)) != WJB_DONE)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Assert(!(type == WJB_KEY && last_was_key));
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
if (type == WJB_KEY)
{
/* stash the key until we know if it has a null value */
k = v;
last_was_key = true;
continue;
}
if (last_was_key)
{
/* if the last element was a key this one can't be */
last_was_key = false;
/* skip this field if value is null */
if (type == WJB_VALUE && v.type == jbvNull)
continue;
/* otherwise, do a delayed push of the key */
Fix some jsonb issues found by Coverity in recent commits. Mostly these issues concern the non-use of function results. These have been changed to use (void) pushJsonbValue(...) instead of assigning the result to a variable that gets overwritten before it is used. There is a larger issue that we should possibly examine the API for pushJsonbValue(), so that instead of returning a value it modifies a state argument. The current idiom is rather clumsy. However, changing that requires quite a bit more work, so this change should do for the moment.
2014-12-16 16:32:06 +01:00
(void) pushJsonbValue(&parseState, WJB_KEY, &k);
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
}
if (type == WJB_VALUE || type == WJB_ELEM)
res = pushJsonbValue(&parseState, type, &v);
else
res = pushJsonbValue(&parseState, type, NULL);
}
Fix some jsonb issues found by Coverity in recent commits. Mostly these issues concern the non-use of function results. These have been changed to use (void) pushJsonbValue(...) instead of assigning the result to a variable that gets overwritten before it is used. There is a larger issue that we should possibly examine the API for pushJsonbValue(), so that instead of returning a value it modifies a state argument. The current idiom is rather clumsy. However, changing that requires quite a bit more work, so this change should do for the moment.
2014-12-16 16:32:06 +01:00
Assert(res != NULL);
Add json_strip_nulls and jsonb_strip_nulls functions. The functions remove object fields, including in nested objects, that have null as a value. In certain cases this can lead to considerably smaller datums, with no loss of semantic information. Andrew Dunstan, reviewed by Pavel Stehule.
2014-12-12 15:00:43 +01:00
PG_RETURN_POINTER(JsonbValueToJsonb(res));
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/*
* Add values from the jsonb to the parse state.
*
* If the parse state container is an object, the jsonb is pushed as
* a value, not a key.
*
Revert "Simplify addJsonbToParseState()" This reverts commit fba12c8c6c4159e1923958a4006b26f3cf873254. This relied on a commit that is also being reverted.
2015-05-27 04:54:11 +02:00
* This needs to be done using an iterator because pushJsonbValue doesn't
* like getting jbvBinary values, so we can't just push jb as a whole.
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
static void
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
addJsonbToParseState(JsonbParseState **jbps, Jsonb *jb)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
JsonbIterator *it;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
JsonbValue *o = &(*jbps)->contVal;
JsonbValue v;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken type;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Revert "Simplify addJsonbToParseState()" This reverts commit fba12c8c6c4159e1923958a4006b26f3cf873254. This relied on a commit that is also being reverted.
2015-05-27 04:54:11 +02:00
it = JsonbIteratorInit(&jb->root);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Assert(o->type == jbvArray || o->type == jbvObject);
if (JB_ROOT_IS_SCALAR(jb))
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
(void) JsonbIteratorNext(&it, &v, false); /* skip array header */
(void) JsonbIteratorNext(&it, &v, false); /* fetch scalar value */
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Revert "Simplify addJsonbToParseState()" This reverts commit fba12c8c6c4159e1923958a4006b26f3cf873254. This relied on a commit that is also being reverted.
2015-05-27 04:54:11 +02:00
switch (o->type)
{
case jbvArray:
(void) pushJsonbValue(jbps, WJB_ELEM, &v);
break;
case jbvObject:
(void) pushJsonbValue(jbps, WJB_VALUE, &v);
break;
default:
elog(ERROR, "unexpected parent of nested structure");
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
{
Revert "Simplify addJsonbToParseState()" This reverts commit fba12c8c6c4159e1923958a4006b26f3cf873254. This relied on a commit that is also being reverted.
2015-05-27 04:54:11 +02:00
while ((type = JsonbIteratorNext(&it, &v, false)) != WJB_DONE)
{
if (type == WJB_KEY || type == WJB_VALUE || type == WJB_ELEM)
(void) pushJsonbValue(jbps, type, &v);
else
(void) pushJsonbValue(jbps, type, NULL);
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
}
/*
* SQL function jsonb_pretty (jsonb)
*
* Pretty-printed text for the jsonb
*/
Datum
jsonb_pretty(PG_FUNCTION_ARGS)
{
Jsonb *jb = PG_GETARG_JSONB(0);
StringInfo str = makeStringInfo();
JsonbToCStringIndent(str, &jb->root, VARSIZE(jb));
PG_RETURN_TEXT_P(cstring_to_text_with_len(str->data, str->len));
}
/*
* SQL function jsonb_concat (jsonb, jsonb)
*
* function for || operator
*/
Datum
jsonb_concat(PG_FUNCTION_ARGS)
{
Jsonb *jb1 = PG_GETARG_JSONB(0);
Jsonb *jb2 = PG_GETARG_JSONB(1);
JsonbParseState *state = NULL;
JsonbValue *res;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
JsonbIterator *it1,
*it2;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/*
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
* If one of the jsonb is empty, just return the other if it's not scalar
* and both are of the same kind. If it's a scalar or they are of
* different kinds we need to perform the concatenation even if one is
Fix the fastpath rule for jsonb_concat with an empty operand. To prevent perverse results, we now only return the other operand if it's not scalar, and if both operands are of the same kind (array or object). Original bug complaint and patch from Oskari Saarenmaa, extended by me to cover the cases of different kinds of jsonb. Backpatch to 9.5 where jsonb_concat was introduced.
2015-09-13 23:06:45 +02:00
* empty.
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
Fix the fastpath rule for jsonb_concat with an empty operand. To prevent perverse results, we now only return the other operand if it's not scalar, and if both operands are of the same kind (array or object). Original bug complaint and patch from Oskari Saarenmaa, extended by me to cover the cases of different kinds of jsonb. Backpatch to 9.5 where jsonb_concat was introduced.
2015-09-13 23:06:45 +02:00
if (JB_ROOT_IS_OBJECT(jb1) == JB_ROOT_IS_OBJECT(jb2))
{
if (JB_ROOT_COUNT(jb1) == 0 && !JB_ROOT_IS_SCALAR(jb2))
PG_RETURN_JSONB(jb2);
else if (JB_ROOT_COUNT(jb2) == 0 && !JB_ROOT_IS_SCALAR(jb1))
PG_RETURN_JSONB(jb1);
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
it1 = JsonbIteratorInit(&jb1->root);
it2 = JsonbIteratorInit(&jb2->root);
res = IteratorConcat(&it1, &it2, &state);
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
Assert(res != NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
PG_RETURN_JSONB(JsonbValueToJsonb(res));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/*
* SQL function jsonb_delete (jsonb, text)
*
* return a copy of the jsonb with the indicated item
* removed.
*/
Datum
jsonb_delete(PG_FUNCTION_ARGS)
{
Jsonb *in = PG_GETARG_JSONB(0);
text *key = PG_GETARG_TEXT_PP(1);
char *keyptr = VARDATA_ANY(key);
int keylen = VARSIZE_ANY_EXHDR(key);
JsonbParseState *state = NULL;
JsonbIterator *it;
JsonbValue v,
*res = NULL;
bool skipNested = false;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot delete from scalar")));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (JB_ROOT_COUNT(in) == 0)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
it = JsonbIteratorInit(&in->root);
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != 0)
{
skipNested = true;
if ((r == WJB_ELEM || r == WJB_KEY) &&
(v.type == jbvString && keylen == v.val.string.len &&
memcmp(keyptr, v.val.string.val, keylen) == 0))
{
/* skip corresponding value as well */
if (r == WJB_KEY)
JsonbIteratorNext(&it, &v, true);
continue;
}
res = pushJsonbValue(&state, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
}
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
Assert(res != NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(JsonbValueToJsonb(res));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
Implement array version of jsonb_delete and operator This makes it possible to delete multiple keys from a jsonb value by passing in an array of text values, which makes the operaiton much faster than individually deleting the keys (which would require copying the jsonb structure over and over again. Reviewed by Dmitry Dolgov and Michael Paquier
2017-01-18 21:37:59 +01:00
/*
* SQL function jsonb_delete (jsonb, variadic text[])
*
* return a copy of the jsonb with the indicated items
* removed.
*/
Datum
jsonb_delete_array(PG_FUNCTION_ARGS)
{
Jsonb *in = PG_GETARG_JSONB(0);
ArrayType *keys = PG_GETARG_ARRAYTYPE_P(1);
Datum *keys_elems;
bool *keys_nulls;
int keys_len;
JsonbParseState *state = NULL;
JsonbIterator *it;
JsonbValue v,
*res = NULL;
bool skipNested = false;
JsonbIteratorToken r;
if (ARR_NDIM(keys) > 1)
ereport(ERROR,
(errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR),
errmsg("wrong number of array subscripts")));
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot delete from scalar")));
if (JB_ROOT_COUNT(in) == 0)
PG_RETURN_JSONB(in);
deconstruct_array(keys, TEXTOID, -1, false, 'i',
&keys_elems, &keys_nulls, &keys_len);
if (keys_len == 0)
PG_RETURN_JSONB(in);
it = JsonbIteratorInit(&in->root);
while ((r = JsonbIteratorNext(&it, &v, skipNested)) != 0)
{
skipNested = true;
if ((r == WJB_ELEM || r == WJB_KEY) && v.type == jbvString)
{
int i;
bool found = false;
for (i = 0; i < keys_len; i++)
{
char *keyptr;
int keylen;
if (keys_nulls[i])
continue;
keyptr = VARDATA_ANY(keys_elems[i]);
keylen = VARSIZE_ANY_EXHDR(keys_elems[i]);
if (keylen == v.val.string.len &&
memcmp(keyptr, v.val.string.val, keylen) == 0)
{
found = true;
break;
}
}
if (found)
{
/* skip corresponding value as well */
if (r == WJB_KEY)
JsonbIteratorNext(&it, &v, true);
continue;
}
}
res = pushJsonbValue(&state, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
}
Assert(res != NULL);
PG_RETURN_JSONB(JsonbValueToJsonb(res));
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/*
* SQL function jsonb_delete (jsonb, int)
*
* return a copy of the jsonb with the indicated item
* removed. Negative int means count back from the
* end of the items.
*/
Datum
jsonb_delete_idx(PG_FUNCTION_ARGS)
{
Jsonb *in = PG_GETARG_JSONB(0);
int idx = PG_GETARG_INT32(1);
JsonbParseState *state = NULL;
JsonbIterator *it;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
uint32 i = 0,
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
n;
JsonbValue v,
*res = NULL;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot delete from scalar")));
Desupport jsonb subscript deletion on objects Supporting deletion of JSON pairs within jsonb objects using an array-style integer subscript allowed for surprising outcomes. This was mostly due to the implementation-defined ordering of pairs within objects for jsonb. It also seems desirable to make jsonb integer subscript deletion consistent with the 9.4 era general purpose integer subscripting operator for jsonb (although that operator returns NULL when an object is encountered, while we prefer here to throw an error). Peter Geoghegan, following discussion on -hackers.
2015-06-08 02:46:00 +02:00
if (JB_ROOT_IS_OBJECT(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Improve some messages
2015-12-11 04:05:27 +01:00
errmsg("cannot delete from object using integer index")));
Desupport jsonb subscript deletion on objects Supporting deletion of JSON pairs within jsonb objects using an array-style integer subscript allowed for surprising outcomes. This was mostly due to the implementation-defined ordering of pairs within objects for jsonb. It also seems desirable to make jsonb integer subscript deletion consistent with the 9.4 era general purpose integer subscripting operator for jsonb (although that operator returns NULL when an object is encountered, while we prefer here to throw an error). Peter Geoghegan, following discussion on -hackers.
2015-06-08 02:46:00 +02:00
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (JB_ROOT_COUNT(in) == 0)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
it = JsonbIteratorInit(&in->root);
r = JsonbIteratorNext(&it, &v, false);
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
Assert(r == WJB_BEGIN_ARRAY);
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
n = v.val.array.nElems;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (idx < 0)
{
if (-idx > n)
idx = n;
else
idx = n + idx;
}
if (idx >= n)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Remove dead code. Defect noticed by Coverity.
2015-07-19 19:19:38 +02:00
pushJsonbValue(&state, r, NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
while ((r = JsonbIteratorNext(&it, &v, true)) != 0)
{
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (r == WJB_ELEM)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
if (i++ == idx)
continue;
}
res = pushJsonbValue(&state, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
}
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Assert(res != NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(JsonbValueToJsonb(res));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/*
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
* SQL function jsonb_set(jsonb, text[], jsonb, boolean)
*
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
Datum
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
jsonb_set(PG_FUNCTION_ARGS)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
Jsonb *in = PG_GETARG_JSONB(0);
ArrayType *path = PG_GETARG_ARRAYTYPE_P(1);
Jsonb *newval = PG_GETARG_JSONB(2);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
bool create = PG_GETARG_BOOL(3);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
JsonbValue *res = NULL;
Datum *path_elems;
bool *path_nulls;
int path_len;
JsonbIterator *it;
JsonbParseState *st = NULL;
if (ARR_NDIM(path) > 1)
ereport(ERROR,
(errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR),
errmsg("wrong number of array subscripts")));
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
errmsg("cannot set path in scalar")));
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
if (JB_ROOT_COUNT(in) == 0 && !create)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
deconstruct_array(path, TEXTOID, -1, false, 'i',
&path_elems, &path_nulls, &path_len);
if (path_len == 0)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
it = JsonbIteratorInit(&in->root);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
res = setPath(&it, path_elems, path_nulls, path_len, &st,
Fix broken jsonb_set() logic for replacing array elements. Commit 0b62fd036 did a fairly sloppy job of refactoring setPath() to support jsonb_insert() along with jsonb_set(). In its defense, though, there was no regression test case exercising the case of replacing an existing element in a jsonb array. Per bug #14366 from Peng Sun. Back-patch to 9.6 where bug was introduced. Report: <20161012065349.1412.47858@wrigleys.postgresql.org>
2016-10-13 06:25:28 +02:00
0, newval, create ? JB_PATH_CREATE : JB_PATH_REPLACE);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Assert(res != NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(JsonbValueToJsonb(res));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/*
Fix typo in comment The function was renamed in 908e23473, but the comment never learned about it.
2016-11-14 17:31:35 +01:00
* SQL function jsonb_delete_path(jsonb, text[])
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
Datum
jsonb_delete_path(PG_FUNCTION_ARGS)
{
Jsonb *in = PG_GETARG_JSONB(0);
ArrayType *path = PG_GETARG_ARRAYTYPE_P(1);
JsonbValue *res = NULL;
Datum *path_elems;
bool *path_nulls;
int path_len;
JsonbIterator *it;
JsonbParseState *st = NULL;
if (ARR_NDIM(path) > 1)
ereport(ERROR,
(errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR),
errmsg("wrong number of array subscripts")));
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot delete path in scalar")));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (JB_ROOT_COUNT(in) == 0)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
deconstruct_array(path, TEXTOID, -1, false, 'i',
&path_elems, &path_nulls, &path_len);
if (path_len == 0)
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(in);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
it = JsonbIteratorInit(&in->root);
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
res = setPath(&it, path_elems, path_nulls, path_len, &st,
0, NULL, JB_PATH_DELETE);
Assert(res != NULL);
PG_RETURN_JSONB(JsonbValueToJsonb(res));
}
/*
* SQL function jsonb_insert(jsonb, text[], jsonb, boolean)
*
*/
Datum
jsonb_insert(PG_FUNCTION_ARGS)
{
Jsonb *in = PG_GETARG_JSONB(0);
ArrayType *path = PG_GETARG_ARRAYTYPE_P(1);
Jsonb *newval = PG_GETARG_JSONB(2);
bool after = PG_GETARG_BOOL(3);
JsonbValue *res = NULL;
Datum *path_elems;
bool *path_nulls;
int path_len;
JsonbIterator *it;
JsonbParseState *st = NULL;
if (ARR_NDIM(path) > 1)
ereport(ERROR,
(errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR),
errmsg("wrong number of array subscripts")));
if (JB_ROOT_IS_SCALAR(in))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot set path in scalar")));
deconstruct_array(path, TEXTOID, -1, false, 'i',
&path_elems, &path_nulls, &path_len);
if (path_len == 0)
PG_RETURN_JSONB(in);
it = JsonbIteratorInit(&in->root);
res = setPath(&it, path_elems, path_nulls, path_len, &st, 0, newval,
after ? JB_PATH_INSERT_AFTER : JB_PATH_INSERT_BEFORE);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Assert(res != NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Fix jsonb replace and delete on scalars and empty structures These operations now error out if attempted on scalars, and simply return the input if attempted on empty arrays or objects. Along the way we remove the unnecessary cloning of the input when it's known to be unchanged. Regression tests covering these cases are added.
2015-05-13 19:52:08 +02:00
PG_RETURN_JSONB(JsonbValueToJsonb(res));
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/*
* Iterate over all jsonb objects and merge them into one.
* The logic of this function copied from the same hstore function,
* except the case, when it1 & it2 represents jbvObject.
* In that case we just append the content of it2 to it1 without any
* verifications.
*/
static JsonbValue *
IteratorConcat(JsonbIterator **it1, JsonbIterator **it2,
JsonbParseState **state)
{
JsonbValue v1,
v2,
*res = NULL;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r1,
r2,
rk1,
rk2;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
r1 = rk1 = JsonbIteratorNext(it1, &v1, false);
r2 = rk2 = JsonbIteratorNext(it2, &v2, false);
/*
* Both elements are objects.
*/
if (rk1 == WJB_BEGIN_OBJECT && rk2 == WJB_BEGIN_OBJECT)
{
/*
Fix typos in comments Dmitriy Olshevskiy
2015-05-17 14:57:30 +02:00
* Append the all tokens from v1 to res, except last WJB_END_OBJECT
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
* (because res will not be finished yet).
*/
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
pushJsonbValue(state, r1, NULL);
while ((r1 = JsonbIteratorNext(it1, &v1, true)) != WJB_END_OBJECT)
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
pushJsonbValue(state, r1, &v1);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/*
* Append the all tokens from v2 to res, include last WJB_END_OBJECT
* (the concatenation will be completed).
*/
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r2 = JsonbIteratorNext(it2, &v2, true)) != 0)
res = pushJsonbValue(state, r2, r2 != WJB_END_OBJECT ? &v2 : NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/*
* Both elements are arrays (either can be scalar).
*/
else if (rk1 == WJB_BEGIN_ARRAY && rk2 == WJB_BEGIN_ARRAY)
{
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
pushJsonbValue(state, r1, NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r1 = JsonbIteratorNext(it1, &v1, true)) != WJB_END_ARRAY)
{
Assert(r1 == WJB_ELEM);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
pushJsonbValue(state, r1, &v1);
}
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
while ((r2 = JsonbIteratorNext(it2, &v2, true)) != WJB_END_ARRAY)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
Assert(r2 == WJB_ELEM);
pushJsonbValue(state, WJB_ELEM, &v2);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
res = pushJsonbValue(state, WJB_END_ARRAY, NULL /* signal to sort */ );
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
/* have we got array || object or object || array? */
else if (((rk1 == WJB_BEGIN_ARRAY && !(*it1)->isScalar) && rk2 == WJB_BEGIN_OBJECT) ||
(rk1 == WJB_BEGIN_OBJECT && (rk2 == WJB_BEGIN_ARRAY && !(*it2)->isScalar)))
{
JsonbIterator **it_array = rk1 == WJB_BEGIN_ARRAY ? it1 : it2;
JsonbIterator **it_object = rk1 == WJB_BEGIN_OBJECT ? it1 : it2;
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
bool prepend = (rk1 == WJB_BEGIN_OBJECT);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
pushJsonbValue(state, WJB_BEGIN_ARRAY, NULL);
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (prepend)
{
pushJsonbValue(state, WJB_BEGIN_OBJECT, NULL);
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r1 = JsonbIteratorNext(it_object, &v1, true)) != 0)
pushJsonbValue(state, r1, r1 != WJB_END_OBJECT ? &v1 : NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r2 = JsonbIteratorNext(it_array, &v2, true)) != 0)
res = pushJsonbValue(state, r2, r2 != WJB_END_ARRAY ? &v2 : NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
{
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r1 = JsonbIteratorNext(it_array, &v1, true)) != WJB_END_ARRAY)
pushJsonbValue(state, r1, &v1);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
pushJsonbValue(state, WJB_BEGIN_OBJECT, NULL);
Clean up and simplify jsonb_concat code. Some of this is made possible by commit 9b74f32cdbff8b9be47fc69164eae552050509ff which lets pushJsonbValue handle binary Jsonb values, meaning that clients no longer have to, and some is just doing things in simpler and more straightforward ways.
2015-05-25 17:43:06 +02:00
while ((r2 = JsonbIteratorNext(it_object, &v2, true)) != 0)
pushJsonbValue(state, r2, r2 != WJB_END_OBJECT ? &v2 : NULL);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
res = pushJsonbValue(state, WJB_END_ARRAY, NULL);
}
}
else
{
/*
* This must be scalar || object or object || scalar, as that's all
* that's left. Both of these make no sense, so error out.
*/
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid concatenation of jsonb objects")));
}
return res;
}
/*
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
* Do most of the heavy work for jsonb_set/jsonb_insert
*
* If JB_PATH_DELETE bit is set in op_type, the element is to be removed.
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
*
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
* If any bit mentioned in JB_PATH_CREATE_OR_INSERT is set in op_type,
* we create the new value if the key or array index does not exist.
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
*
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
* Bits JB_PATH_INSERT_BEFORE and JB_PATH_INSERT_AFTER in op_type
* behave as JB_PATH_CREATE if new value is inserted in JsonbObject.
*
* All path elements before the last must already exist
* whatever bits in op_type are set, or nothing is done.
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
static JsonbValue *
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPath(JsonbIterator **it, Datum *path_elems,
bool *path_nulls, int path_len,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
JsonbParseState **st, int level, Jsonb *newval, int op_type)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
JsonbValue v;
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Code review for error reports in jsonb_set(). User-facing (even tested by regression tests) error conditions were thrown with elog(), hence had wrong SQLSTATE and were untranslatable. And the error message texts weren't up to project style, either.
2016-03-23 16:00:39 +01:00
JsonbValue *res;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Prevent stack overflow in json-related functions. Sufficiently-deep recursion heretofore elicited a SIGSEGV. If an application constructs PostgreSQL json or jsonb values from arbitrary user input, application users could have exploited this to terminate all active database connections. That applies to 9.3, where the json parser adopted recursive descent, and later versions. Only row_to_json() and array_to_json() were at risk in 9.2, both in a non-security capacity. Back-patch to 9.2, where the json type was introduced. Oskari Saarenmaa, reviewed by Michael Paquier. Security: CVE-2015-5289
2015-10-05 16:06:29 +02:00
check_stack_depth();
Disallow invalid path elements in jsonb_set Null path elements and, where the object is an array, invalid integer elements now cause an error. Incorrect behaviour noted by Thom Brown, patch from Dmitry Dolgov. Backpatch to 9.5 where jsonb_set was introduced
2015-10-04 19:28:16 +02:00
if (path_nulls[level])
Code review for error reports in jsonb_set(). User-facing (even tested by regression tests) error conditions were thrown with elog(), hence had wrong SQLSTATE and were untranslatable. And the error message texts weren't up to project style, either.
2016-03-23 16:00:39 +01:00
ereport(ERROR,
(errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
errmsg("path element at position %d is null",
level + 1)));
Disallow invalid path elements in jsonb_set Null path elements and, where the object is an array, invalid integer elements now cause an error. Incorrect behaviour noted by Thom Brown, patch from Dmitry Dolgov. Backpatch to 9.5 where jsonb_set was introduced
2015-10-04 19:28:16 +02:00
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
r = JsonbIteratorNext(it, &v, false);
switch (r)
{
case WJB_BEGIN_ARRAY:
(void) pushJsonbValue(st, r, NULL);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPathArray(it, path_elems, path_nulls, path_len, st, level,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
newval, v.val.array.nElems, op_type);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
r = JsonbIteratorNext(it, &v, false);
Assert(r == WJB_END_ARRAY);
res = pushJsonbValue(st, r, NULL);
break;
case WJB_BEGIN_OBJECT:
(void) pushJsonbValue(st, r, NULL);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPathObject(it, path_elems, path_nulls, path_len, st, level,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
newval, v.val.object.nPairs, op_type);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
r = JsonbIteratorNext(it, &v, true);
Assert(r == WJB_END_OBJECT);
res = pushJsonbValue(st, r, NULL);
break;
case WJB_ELEM:
case WJB_VALUE:
res = pushJsonbValue(st, r, &v);
break;
default:
Code review for error reports in jsonb_set(). User-facing (even tested by regression tests) error conditions were thrown with elog(), hence had wrong SQLSTATE and were untranslatable. And the error message texts weren't up to project style, either.
2016-03-23 16:00:39 +01:00
elog(ERROR, "unrecognized iterator result: %d", (int) r);
res = NULL; /* keep compiler quiet */
break;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
return res;
}
/*
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
* Object walker for setPath
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
static void
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPathObject(JsonbIterator **it, Datum *path_elems, bool *path_nulls,
int path_len, JsonbParseState **st, int level,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
Jsonb *newval, uint32 npairs, int op_type)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
JsonbValue v;
int i;
JsonbValue k;
bool done = false;
if (level >= path_len || path_nulls[level])
done = true;
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
/* empty object is a special case for create */
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
if ((npairs == 0) && (op_type & JB_PATH_CREATE_OR_INSERT) &&
(level == path_len - 1))
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
{
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
JsonbValue newkey;
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
newkey.type = jbvString;
newkey.val.string.len = VARSIZE_ANY_EXHDR(path_elems[level]);
newkey.val.string.val = VARDATA_ANY(path_elems[level]);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
(void) pushJsonbValue(st, WJB_KEY, &newkey);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
addJsonbToParseState(st, newval);
}
for (i = 0; i < npairs; i++)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r = JsonbIteratorNext(it, &k, true);
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
Assert(r == WJB_KEY);
if (!done &&
k.val.string.len == VARSIZE_ANY_EXHDR(path_elems[level]) &&
memcmp(k.val.string.val, VARDATA_ANY(path_elems[level]),
k.val.string.len) == 0)
{
if (level == path_len - 1)
{
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
/*
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
* called from jsonb_insert(), it forbids redefining an
Fix typo
2016-07-26 04:07:53 +02:00
* existing value
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
*/
if (op_type & (JB_PATH_INSERT_BEFORE | JB_PATH_INSERT_AFTER))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("cannot replace existing key"),
errhint("Try using the function jsonb_set "
"to replace key value.")));
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
r = JsonbIteratorNext(it, &v, true); /* skip value */
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
if (!(op_type & JB_PATH_DELETE))
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
(void) pushJsonbValue(st, WJB_KEY, &k);
addJsonbToParseState(st, newval);
}
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
done = true;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
{
(void) pushJsonbValue(st, r, &k);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPath(it, path_elems, path_nulls, path_len,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
st, level + 1, newval, op_type);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
}
else
{
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
if ((op_type & JB_PATH_CREATE_OR_INSERT) && !done &&
level == path_len - 1 && i == npairs - 1)
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
{
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
JsonbValue newkey;
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
newkey.type = jbvString;
newkey.val.string.len = VARSIZE_ANY_EXHDR(path_elems[level]);
newkey.val.string.val = VARDATA_ANY(path_elems[level]);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
Avoid naming a variable "new", and remove bogus initializer. Per gripe from Tom Lane.
2015-06-01 04:56:53 +02:00
(void) pushJsonbValue(st, WJB_KEY, &newkey);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
addJsonbToParseState(st, newval);
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
(void) pushJsonbValue(st, r, &k);
r = JsonbIteratorNext(it, &v, false);
(void) pushJsonbValue(st, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
if (r == WJB_BEGIN_ARRAY || r == WJB_BEGIN_OBJECT)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
int walking_level = 1;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
while (walking_level != 0)
{
r = JsonbIteratorNext(it, &v, false);
if (r == WJB_BEGIN_ARRAY || r == WJB_BEGIN_OBJECT)
++walking_level;
if (r == WJB_END_ARRAY || r == WJB_END_OBJECT)
--walking_level;
(void) pushJsonbValue(st, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
}
}
}
}
}
/*
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
* Array walker for setPath
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
*/
static void
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
setPathArray(JsonbIterator **it, Datum *path_elems, bool *path_nulls,
int path_len, JsonbParseState **st, int level,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
Jsonb *newval, uint32 nelems, int op_type)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
JsonbValue v;
int idx,
i;
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
bool done = false;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/* pick correct index */
if (level < path_len && !path_nulls[level])
{
Fix unsafe use of strtol() on a non-null-terminated Text datum. jsonb_set() could produce wrong answers or incorrect error reports, or in the worst case even crash, when trying to convert a path-array element into an integer for use as an array subscript. Per report from Vitaly Burovoy. Back-patch to 9.5 where the faulty code was introduced (in commit c6947010ceb42143). Michael Paquier
2016-03-23 15:43:13 +01:00
char *c = TextDatumGetCString(path_elems[level]);
Fix "path" infrastructure bug affecting jsonb_set() jsonb_set() and other clients of the setPathArray() utility function could get spurious results when an array integer subscript is provided that is not within the range of int. To fix, ensure that the value returned by strtol() within setPathArray() is within the range of int; when it isn't, assume an invalid input in line with existing, similar cases. The path-orientated operators that appeared in PostgreSQL 9.3 and 9.4 do not call setPathArray(), and already independently take this precaution, so no change there. Peter Geoghegan
2015-06-13 01:26:03 +02:00
long lindex;
Code review for error reports in jsonb_set(). User-facing (even tested by regression tests) error conditions were thrown with elog(), hence had wrong SQLSTATE and were untranslatable. And the error message texts weren't up to project style, either.
2016-03-23 16:00:39 +01:00
char *badp;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
errno = 0;
Fix "path" infrastructure bug affecting jsonb_set() jsonb_set() and other clients of the setPathArray() utility function could get spurious results when an array integer subscript is provided that is not within the range of int. To fix, ensure that the value returned by strtol() within setPathArray() is within the range of int; when it isn't, assume an invalid input in line with existing, similar cases. The path-orientated operators that appeared in PostgreSQL 9.3 and 9.4 do not call setPathArray(), and already independently take this precaution, so no change there. Peter Geoghegan
2015-06-13 01:26:03 +02:00
lindex = strtol(c, &badp, 10);
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
if (errno != 0 || badp == c || *badp != '\0' || lindex > INT_MAX ||
lindex < INT_MIN)
Code review for error reports in jsonb_set(). User-facing (even tested by regression tests) error conditions were thrown with elog(), hence had wrong SQLSTATE and were untranslatable. And the error message texts weren't up to project style, either.
2016-03-23 16:00:39 +01:00
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("path element at position %d is not an integer: \"%s\"",
level + 1, c)));
idx = lindex;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
idx = nelems;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (idx < 0)
{
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
if (-idx > nelems)
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
idx = INT_MIN;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
else
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
idx = nelems + idx;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
if (idx > 0 && idx > nelems)
idx = nelems;
/*
Support JSON negative array subscripts everywhere Previously, there was an inconsistency across json/jsonb operators that operate on datums containing JSON arrays -- only some operators supported negative array count-from-the-end subscripting. Specifically, only a new-to-9.5 jsonb deletion operator had support (the new "jsonb - integer" operator). This inconsistency seemed likely to be counter-intuitive to users. To fix, allow all places where the user can supply an integer subscript to accept a negative subscript value, including path-orientated operators and functions, as well as other extraction operators. This will need to be called out as an incompatibility in the 9.5 release notes, since it's possible that users are relying on certain established extraction operators changed here yielding NULL in the event of a negative subscript. For the json type, this requires adding a way of cheaply getting the total JSON array element count ahead of time when parsing arrays with a negative subscript involved, necessitating an ad-hoc lex and parse. This is followed by a "conversion" from a negative subscript to its equivalent positive-wise value using the count. From there on, it's as if a positive-wise value was originally provided. Note that there is still a minor inconsistency here across jsonb deletion operators. Unlike the aforementioned new "-" deletion operator that accepts an integer on its right hand side, the new "#-" path orientated deletion variant does not throw an error when it appears like an array subscript (input that could be recognized by as an integer literal) is being used on an object, which is wrong-headed. The reason for not being stricter is that it could be the case that an object pair happens to have a key value that looks like an integer; in general, these two possibilities are impossible to differentiate with rhs path text[] argument elements. However, we still don't allow the "#-" path-orientated deletion operator to perform array-style subscripting. Rather, we just return the original left operand value in the event of a negative subscript (which seems analogous to how the established "jsonb/json #> text[]" path-orientated operator may yield NULL in the event of an invalid subscript). In passing, make SetArrayPath() stricter about not accepting cases where there is trailing non-numeric garbage bytes rather than a clean NUL byte. This means, for example, that strings like "10e10" are now not accepted as an array subscript of 10 by some new-to-9.5 path-orientated jsonb operators (e.g. the new #- operator). Finally, remove dead code for jsonb subscript deletion; arguably, this should have been done in commit b81c7b409. Peter Geoghegan and Andrew Dunstan
2015-07-18 02:56:13 +02:00
* if we're creating, and idx == INT_MIN, we prepend the new value to the
* array also if the array is empty - in which case we don't really care
* what the idx value is
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
*/
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
if ((idx == INT_MIN || nelems == 0) && (level == path_len - 1) &&
(op_type & JB_PATH_CREATE_OR_INSERT))
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
{
Assert(newval != NULL);
addJsonbToParseState(st, newval);
done = true;
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
/* iterate over the array elements */
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
for (i = 0; i < nelems; i++)
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
{
Use JsonbIteratorToken consistently in automatic variable declarations. Many functions stored JsonbIteratorToken values in variables of other integer types. Also, standardize order relative to other declarations. Expect compilers to generate the same code before and after this change.
2015-10-12 05:53:35 +02:00
JsonbIteratorToken r;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
if (i == idx && level < path_len)
{
if (level == path_len - 1)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
r = JsonbIteratorNext(it, &v, true); /* skip */
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
if (op_type & (JB_PATH_INSERT_BEFORE | JB_PATH_CREATE))
addJsonbToParseState(st, newval);
/*
* We should keep current value only in case of
pgindent run for 9.6
2016-06-10 00:02:36 +02:00
* JB_PATH_INSERT_BEFORE or JB_PATH_INSERT_AFTER because
* otherwise it should be deleted or replaced
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
*/
if (op_type & (JB_PATH_INSERT_AFTER | JB_PATH_INSERT_BEFORE))
(void) pushJsonbValue(st, r, &v);
Fix broken jsonb_set() logic for replacing array elements. Commit 0b62fd036 did a fairly sloppy job of refactoring setPath() to support jsonb_insert() along with jsonb_set(). In its defense, though, there was no regression test case exercising the case of replacing an existing element in a jsonb array. Per bug #14366 from Peng Sun. Back-patch to 9.6 where bug was introduced. Report: <20161012065349.1412.47858@wrigleys.postgresql.org>
2016-10-13 06:25:28 +02:00
if (op_type & (JB_PATH_INSERT_AFTER | JB_PATH_REPLACE))
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
addJsonbToParseState(st, newval);
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
done = true;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
(void) setPath(it, path_elems, path_nulls, path_len,
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
st, level + 1, newval, op_type);
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
else
{
r = JsonbIteratorNext(it, &v, false);
(void) pushJsonbValue(st, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
if (r == WJB_BEGIN_ARRAY || r == WJB_BEGIN_OBJECT)
{
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
int walking_level = 1;
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
while (walking_level != 0)
{
r = JsonbIteratorNext(it, &v, false);
if (r == WJB_BEGIN_ARRAY || r == WJB_BEGIN_OBJECT)
++walking_level;
if (r == WJB_END_ARRAY || r == WJB_END_OBJECT)
--walking_level;
(void) pushJsonbValue(st, r, r < WJB_BEGIN_ARRAY ? &v : NULL);
}
}
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
Fix broken jsonb_set() logic for replacing array elements. Commit 0b62fd036 did a fairly sloppy job of refactoring setPath() to support jsonb_insert() along with jsonb_set(). In its defense, though, there was no regression test case exercising the case of replacing an existing element in a jsonb array. Per bug #14366 from Peng Sun. Back-patch to 9.6 where bug was introduced. Report: <20161012065349.1412.47858@wrigleys.postgresql.org>
2016-10-13 06:25:28 +02:00
if ((op_type & JB_PATH_CREATE_OR_INSERT) && !done &&
Add jsonb_insert It inserts a new value into an jsonb array at arbitrary position or a new key to jsonb object. Author: Dmitry Dolgov Reviewers: Petr Jelinek, Vitaly Burovoy, Andrew Dunstan
2016-04-06 18:20:17 +02:00
level == path_len - 1 && i == nelems - 1)
Rename jsonb_replace to jsonb_set and allow it to add new values The function is given a fourth parameter, which defaults to true. When this parameter is true, if the last element of the path is missing in the original json, jsonb_set creates it in the result and assigns it the new value. If it is false then the function does nothing unless all elements of the path are present, including the last. Based on some original code from Dmitry Dolgov, heavily modified by me. Catalog version bumped.
2015-06-01 02:34:10 +02:00
{
addJsonbToParseState(st, newval);
}
Additional functions and operators for jsonb jsonb_pretty(jsonb) produces nicely indented json output. jsonb || jsonb concatenates two jsonb values. jsonb - text removes a key and its associated value from the json jsonb - int removes the designated array element jsonb - text[] removes a key and associated value or array element at the designated path jsonb_replace(jsonb,text[],jsonb) replaces the array element designated by the path or the value associated with the key designated by the path with the given value. Original work by Dmitry Dolgov, adapted and reworked for PostgreSQL core by Andrew Dunstan, reviewed and tidied up by Petr Jelinek.
2015-05-12 21:52:45 +02:00
}
}
}