postgresql/contrib/ltree/ltree.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

318 lines
11 KiB
C
Raw Permalink Normal View History

2010-09-20 22:08:53 +02:00
/* contrib/ltree/ltree.h */
#ifndef __LTREE_H__
#define __LTREE_H__
#include "fmgr.h"
#include "tsearch/ts_locale.h"
#include "utils/memutils.h"
/* ltree */
/*
* We want the maximum length of a label to be encoding-independent, so
* set it somewhat arbitrarily at 1000 characters (not bytes), while using
* uint16 fields to hold the byte length.
*/
#define LTREE_LABEL_MAX_CHARS 1000
/*
* LOWER_NODE used to be defined in the Makefile via the compile flags.
* However the MSVC build scripts neglected to do the same which resulted in
* MSVC builds not using LOWER_NODE. Since then, the MSVC scripts have been
* modified to look for -D compile flags in Makefiles, so here, in order to
* get the historic behavior of LOWER_NODE not being defined on MSVC, we only
* define it when not building in that environment. This is important as we
* want to maintain the same LOWER_NODE behavior after a pg_upgrade.
*/
#ifndef _MSC_VER
#define LOWER_NODE
#endif
typedef struct
{
uint16 len; /* label string length in bytes */
char name[FLEXIBLE_ARRAY_MEMBER];
} ltree_level;
#define LEVEL_HDRSIZE (offsetof(ltree_level,name))
#define LEVEL_NEXT(x) ( (ltree_level*)( ((char*)(x)) + MAXALIGN(((ltree_level*)(x))->len + LEVEL_HDRSIZE) ) )
typedef struct
{
int32 vl_len_; /* varlena header (do not touch directly!) */
uint16 numlevel; /* number of labels */
/* Array of maxalign'd ltree_level structs follows: */
char data[FLEXIBLE_ARRAY_MEMBER];
} ltree;
#define LTREE_HDRSIZE MAXALIGN( offsetof(ltree, data) )
#define LTREE_FIRST(x) ( (ltree_level*)( ((char*)(x))+LTREE_HDRSIZE ) )
#define LTREE_MAX_LEVELS PG_UINT16_MAX /* ltree.numlevel is uint16 */
/* lquery */
/* lquery_variant: one branch of some OR'ed alternatives */
typedef struct
{
int32 val; /* CRC of label string */
uint16 len; /* label string length in bytes */
uint8 flag; /* see LVAR_xxx flags below */
char name[FLEXIBLE_ARRAY_MEMBER];
} lquery_variant;
/*
* Note: these macros contain too many MAXALIGN calls and so will sometimes
* overestimate the space needed for an lquery_variant. However, we can't
* change it without breaking on-disk compatibility for lquery.
*/
#define LVAR_HDRSIZE MAXALIGN(offsetof(lquery_variant, name))
#define LVAR_NEXT(x) ( (lquery_variant*)( ((char*)(x)) + MAXALIGN(((lquery_variant*)(x))->len) + LVAR_HDRSIZE ) )
#define LVAR_ANYEND 0x01 /* '*' flag: prefix match */
#define LVAR_INCASE 0x02 /* '@' flag: case-insensitive match */
#define LVAR_SUBLEXEME 0x04 /* '%' flag: word-wise match */
/*
* In an lquery_level, "flag" contains the union of the variants' flags
* along with possible LQL_xxx flags; so those bit sets can't overlap.
Fix lquery's NOT handling, and add ability to quantify non-'*' items. The existing implementation of the ltree ~ lquery match operator is sufficiently complex and undocumented that it's hard to tell exactly what it does. But one thing it clearly gets wrong is the combination of NOT symbols (!) and '*' symbols. A pattern such as '*.!foo.*' should, by any ordinary understanding of regular expression behavior, match any ltree that has at least one label that's not "foo". As best we can tell by experimentation, what it's actually matching is any ltree in which *no* label is "foo". That's surprising, and not at all what the documentation says. Now, that's arguably a useful behavior, so if we rewrite to fix the bug we should provide some other way to get it. To do so, add the ability to attach lquery quantifiers to non-'*' items as well as '*'s. Then the pattern '!foo{,}' expresses "any ltree in which no label is foo". For backwards compatibility, the default quantifier for non-'*' items has to be "{1}", although the default for '*' items is '{,}'. I wouldn't have done it like that in a green field, but it's not totally horrible. Armed with that, rewrite checkCond() from scratch. Treating '*' and non-'*' items alike makes it simpler, not more complicated, so that the function actually gets a lot shorter than it was. Filip Rembiałkowski, Tom Lane, Nikita Glukhov, per a very ancient bug report from M. Palm Discussion: https://postgr.es/m/CAP_rww=waX2Oo6q+MbMSiZ9ktdj6eaJj0cQzNu=Ry2cCDij5fw@mail.gmail.com
2020-03-31 17:14:30 +02:00
*
* "low" and "high" are nominally the minimum and maximum number of matches.
* However, for backwards compatibility with pre-v13 on-disk lqueries,
* non-'*' levels (those with numvar > 0) only have valid low/high if the
* LQL_COUNT flag is set; otherwise those fields are zero, but the behavior
* is as if they were both 1.
*/
typedef struct
{
uint16 totallen; /* total length of this level, in bytes */
uint16 flag; /* see LQL_xxx and LVAR_xxx flags */
uint16 numvar; /* number of variants; 0 means '*' */
Fix lquery's NOT handling, and add ability to quantify non-'*' items. The existing implementation of the ltree ~ lquery match operator is sufficiently complex and undocumented that it's hard to tell exactly what it does. But one thing it clearly gets wrong is the combination of NOT symbols (!) and '*' symbols. A pattern such as '*.!foo.*' should, by any ordinary understanding of regular expression behavior, match any ltree that has at least one label that's not "foo". As best we can tell by experimentation, what it's actually matching is any ltree in which *no* label is "foo". That's surprising, and not at all what the documentation says. Now, that's arguably a useful behavior, so if we rewrite to fix the bug we should provide some other way to get it. To do so, add the ability to attach lquery quantifiers to non-'*' items as well as '*'s. Then the pattern '!foo{,}' expresses "any ltree in which no label is foo". For backwards compatibility, the default quantifier for non-'*' items has to be "{1}", although the default for '*' items is '{,}'. I wouldn't have done it like that in a green field, but it's not totally horrible. Armed with that, rewrite checkCond() from scratch. Treating '*' and non-'*' items alike makes it simpler, not more complicated, so that the function actually gets a lot shorter than it was. Filip Rembiałkowski, Tom Lane, Nikita Glukhov, per a very ancient bug report from M. Palm Discussion: https://postgr.es/m/CAP_rww=waX2Oo6q+MbMSiZ9ktdj6eaJj0cQzNu=Ry2cCDij5fw@mail.gmail.com
2020-03-31 17:14:30 +02:00
uint16 low; /* minimum repeat count */
uint16 high; /* maximum repeat count */
/* Array of maxalign'd lquery_variant structs follows: */
char variants[FLEXIBLE_ARRAY_MEMBER];
} lquery_level;
2002-09-04 22:31:48 +02:00
#define LQL_HDRSIZE MAXALIGN( offsetof(lquery_level,variants) )
#define LQL_NEXT(x) ( (lquery_level*)( ((char*)(x)) + MAXALIGN(((lquery_level*)(x))->totallen) ) )
#define LQL_FIRST(x) ( (lquery_variant*)( ((char*)(x))+LQL_HDRSIZE ) )
#define LQL_NOT 0x10 /* level has '!' (NOT) prefix */
Fix lquery's NOT handling, and add ability to quantify non-'*' items. The existing implementation of the ltree ~ lquery match operator is sufficiently complex and undocumented that it's hard to tell exactly what it does. But one thing it clearly gets wrong is the combination of NOT symbols (!) and '*' symbols. A pattern such as '*.!foo.*' should, by any ordinary understanding of regular expression behavior, match any ltree that has at least one label that's not "foo". As best we can tell by experimentation, what it's actually matching is any ltree in which *no* label is "foo". That's surprising, and not at all what the documentation says. Now, that's arguably a useful behavior, so if we rewrite to fix the bug we should provide some other way to get it. To do so, add the ability to attach lquery quantifiers to non-'*' items as well as '*'s. Then the pattern '!foo{,}' expresses "any ltree in which no label is foo". For backwards compatibility, the default quantifier for non-'*' items has to be "{1}", although the default for '*' items is '{,}'. I wouldn't have done it like that in a green field, but it's not totally horrible. Armed with that, rewrite checkCond() from scratch. Treating '*' and non-'*' items alike makes it simpler, not more complicated, so that the function actually gets a lot shorter than it was. Filip Rembiałkowski, Tom Lane, Nikita Glukhov, per a very ancient bug report from M. Palm Discussion: https://postgr.es/m/CAP_rww=waX2Oo6q+MbMSiZ9ktdj6eaJj0cQzNu=Ry2cCDij5fw@mail.gmail.com
2020-03-31 17:14:30 +02:00
#define LQL_COUNT 0x20 /* level is non-'*' and has repeat counts */
#ifdef LOWER_NODE
#define FLG_CANLOOKSIGN(x) ( ( (x) & ( LQL_NOT | LVAR_ANYEND | LVAR_SUBLEXEME ) ) == 0 )
#else
#define FLG_CANLOOKSIGN(x) ( ( (x) & ( LQL_NOT | LVAR_ANYEND | LVAR_SUBLEXEME | LVAR_INCASE ) ) == 0 )
#endif
#define LQL_CANLOOKSIGN(x) FLG_CANLOOKSIGN( ((lquery_level*)(x))->flag )
typedef struct
2002-09-04 22:31:48 +02:00
{
int32 vl_len_; /* varlena header (do not touch directly!) */
uint16 numlevel; /* number of lquery_levels */
uint16 firstgood; /* number of leading simple-match levels */
uint16 flag; /* see LQUERY_xxx flags below */
/* Array of maxalign'd lquery_level structs follows: */
char data[FLEXIBLE_ARRAY_MEMBER];
} lquery;
2002-09-04 22:31:48 +02:00
#define LQUERY_HDRSIZE MAXALIGN( offsetof(lquery, data) )
#define LQUERY_FIRST(x) ( (lquery_level*)( ((char*)(x))+LQUERY_HDRSIZE ) )
#define LQUERY_MAX_LEVELS PG_UINT16_MAX /* lquery.numlevel is uint16 */
#define LQUERY_HASNOT 0x01
/* valid label chars are alphanumerics, underscores and hyphens */
#define ISLABEL(x) ( t_isalnum(x) || t_iseq(x, '_') || t_iseq(x, '-') )
/* full text query */
/*
* item in polish notation with back link
* to left operand
*/
typedef struct ITEM
{
int16 type;
int16 left;
int32 val;
uint8 flag;
/* user-friendly value */
uint8 length;
uint16 distance;
} ITEM;
/*
*Storage:
* (len)(size)(array of ITEM)(array of operand in user-friendly form)
*/
typedef struct
{
int32 vl_len_; /* varlena header (do not touch directly!) */
int32 size;
char data[FLEXIBLE_ARRAY_MEMBER];
} ltxtquery;
#define HDRSIZEQT MAXALIGN(VARHDRSZ + sizeof(int32))
#define COMPUTESIZE(size,lenofoperand) ( HDRSIZEQT + (size) * sizeof(ITEM) + (lenofoperand) )
#define LTXTQUERY_TOO_BIG(size,lenofoperand) \
((size) > (MaxAllocSize - HDRSIZEQT - (lenofoperand)) / sizeof(ITEM))
#define GETQUERY(x) (ITEM*)( (char*)(x)+HDRSIZEQT )
#define GETOPERAND(x) ( (char*)GETQUERY(x) + ((ltxtquery*)x)->size * sizeof(ITEM) )
#define ISOPERATOR(x) ( (x)=='!' || (x)=='&' || (x)=='|' || (x)=='(' || (x)==')' )
#define END 0
#define ERR 1
#define VAL 2
#define OPR 3
#define OPEN 4
#define CLOSE 5
#define VALTRUE 6 /* for stop words */
#define VALFALSE 7
/* use in array iterator */
PGDLLEXPORT Datum ltree_isparent(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltree_risparent(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltq_regex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltq_rregex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum lt_q_regex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum lt_q_rregex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltxtq_exec(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltxtq_rexec(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltq_regex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltq_rregex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _lt_q_regex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _lt_q_rregex(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltxtq_exec(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltxtq_rexec(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltree_isparent(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum _ltree_risparent(PG_FUNCTION_ARGS);
/* Concatenation functions */
PGDLLEXPORT Datum ltree_addltree(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltree_addtext(PG_FUNCTION_ARGS);
PGDLLEXPORT Datum ltree_textadd(PG_FUNCTION_ARGS);
/* Util function */
PGDLLEXPORT Datum ltree_in(PG_FUNCTION_ARGS);
bool ltree_execute(ITEM *curitem, void *checkval,
bool calcnot, bool (*chkcond) (void *checkval, ITEM *val));
int ltree_compare(const ltree *a, const ltree *b);
bool inner_isparent(const ltree *c, const ltree *p);
bool compare_subnode(ltree_level *t, char *qn, int len,
int (*cmpptr) (const char *, const char *, size_t), bool anyend);
ltree *lca_inner(ltree **a, int len);
int ltree_strncasecmp(const char *a, const char *b, size_t s);
/* fmgr macros for ltree objects */
#define DatumGetLtreeP(X) ((ltree *) PG_DETOAST_DATUM(X))
#define DatumGetLtreePCopy(X) ((ltree *) PG_DETOAST_DATUM_COPY(X))
#define PG_GETARG_LTREE_P(n) DatumGetLtreeP(PG_GETARG_DATUM(n))
#define PG_GETARG_LTREE_P_COPY(n) DatumGetLtreePCopy(PG_GETARG_DATUM(n))
#define DatumGetLqueryP(X) ((lquery *) PG_DETOAST_DATUM(X))
#define DatumGetLqueryPCopy(X) ((lquery *) PG_DETOAST_DATUM_COPY(X))
#define PG_GETARG_LQUERY_P(n) DatumGetLqueryP(PG_GETARG_DATUM(n))
#define PG_GETARG_LQUERY_P_COPY(n) DatumGetLqueryPCopy(PG_GETARG_DATUM(n))
#define DatumGetLtxtqueryP(X) ((ltxtquery *) PG_DETOAST_DATUM(X))
#define DatumGetLtxtqueryPCopy(X) ((ltxtquery *) PG_DETOAST_DATUM_COPY(X))
#define PG_GETARG_LTXTQUERY_P(n) DatumGetLtxtqueryP(PG_GETARG_DATUM(n))
#define PG_GETARG_LTXTQUERY_P_COPY(n) DatumGetLtxtqueryPCopy(PG_GETARG_DATUM(n))
/* GiST support for ltree */
#define BITBYTE 8
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define SIGLENBIT(siglen) ((siglen) * BITBYTE)
#define LTREE_SIGLEN_DEFAULT (2 * sizeof(int32))
#define LTREE_SIGLEN_MAX GISTMaxIndexKeySize
#define LTREE_GET_SIGLEN() (PG_HAS_OPCLASS_OPTIONS() ? \
((LtreeGistOptions *) PG_GET_OPCLASS_OPTIONS())->siglen : \
LTREE_SIGLEN_DEFAULT)
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
typedef unsigned char *BITVECP;
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define LOOPBYTE(siglen) \
for(i = 0; i < (siglen); i++)
#define GETBYTE(x,i) ( *( (BITVECP)(x) + (int)( (i) / BITBYTE ) ) )
#define GETBITBYTE(x,i) ( ((unsigned char)(x)) >> i & 0x01 )
#define CLRBIT(x,i) GETBYTE(x,i) &= ~( 0x01 << ( (i) % BITBYTE ) )
#define SETBIT(x,i) GETBYTE(x,i) |= ( 0x01 << ( (i) % BITBYTE ) )
#define GETBIT(x,i) ( (GETBYTE(x,i) >> ( (i) % BITBYTE )) & 0x01 )
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define HASHVAL(val, siglen) (((unsigned int)(val)) % SIGLENBIT(siglen))
#define HASH(sign, val, siglen) SETBIT((sign), HASHVAL(val, siglen))
/*
* type of index key for ltree. Tree are combined B-Tree and R-Tree
* Storage:
* Leaf pages
* (len)(flag)(ltree)
* Non-Leaf
* (len)(flag)(sign)(left_ltree)(right_ltree)
* ALLTRUE: (len)(flag)(left_ltree)(right_ltree)
*
*/
typedef struct
{
int32 vl_len_; /* varlena header (do not touch directly!) */
uint32 flag;
char data[FLEXIBLE_ARRAY_MEMBER];
} ltree_gist;
2002-09-04 22:31:48 +02:00
#define LTG_ONENODE 0x01
#define LTG_ALLTRUE 0x02
#define LTG_NORIGHT 0x04
2002-09-04 22:31:48 +02:00
#define LTG_HDRSIZE MAXALIGN(VARHDRSZ + sizeof(uint32))
#define LTG_SIGN(x) ( (BITVECP)( ((char*)(x))+LTG_HDRSIZE ) )
#define LTG_NODE(x) ( (ltree*)( ((char*)(x))+LTG_HDRSIZE ) )
#define LTG_ISONENODE(x) ( ((ltree_gist*)(x))->flag & LTG_ONENODE )
#define LTG_ISALLTRUE(x) ( ((ltree_gist*)(x))->flag & LTG_ALLTRUE )
#define LTG_ISNORIGHT(x) ( ((ltree_gist*)(x))->flag & LTG_NORIGHT )
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define LTG_LNODE(x, siglen) ( (ltree*)( ( ((char*)(x))+LTG_HDRSIZE ) + ( LTG_ISALLTRUE(x) ? 0 : (siglen) ) ) )
#define LTG_RENODE(x, siglen) ( (ltree*)( ((char*)LTG_LNODE(x, siglen)) + VARSIZE(LTG_LNODE(x, siglen))) )
#define LTG_RNODE(x, siglen) ( LTG_ISNORIGHT(x) ? LTG_LNODE(x, siglen) : LTG_RENODE(x, siglen) )
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define LTG_GETLNODE(x, siglen) ( LTG_ISONENODE(x) ? LTG_NODE(x) : LTG_LNODE(x, siglen) )
#define LTG_GETRNODE(x, siglen) ( LTG_ISONENODE(x) ? LTG_NODE(x) : LTG_RNODE(x, siglen) )
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
extern ltree_gist *ltree_gist_alloc(bool isalltrue, BITVECP sign, int siglen,
ltree *left, ltree *right);
/* GiST support for ltree[] */
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define LTREE_ASIGLEN_DEFAULT (7 * sizeof(int32))
#define LTREE_ASIGLEN_MAX GISTMaxIndexKeySize
#define LTREE_GET_ASIGLEN() (PG_HAS_OPCLASS_OPTIONS() ? \
((LtreeGistOptions *) PG_GET_OPCLASS_OPTIONS())->siglen : \
LTREE_ASIGLEN_DEFAULT)
#define ASIGLENBIT(siglen) ((siglen) * BITBYTE)
#define ALOOPBYTE(siglen) \
for (i = 0; i < (siglen); i++)
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
#define AHASHVAL(val, siglen) (((unsigned int)(val)) % ASIGLENBIT(siglen))
#define AHASH(sign, val, siglen) SETBIT((sign), AHASHVAL(val, siglen))
Implement operator class parameters PostgreSQL provides set of template index access methods, where opclasses have much freedom in the semantics of indexing. These index AMs are GiST, GIN, SP-GiST and BRIN. There opclasses define representation of keys, operations on them and supported search strategies. So, it's natural that opclasses may be faced some tradeoffs, which require user-side decision. This commit implements opclass parameters allowing users to set some values, which tell opclass how to index the particular dataset. This commit doesn't introduce new storage in system catalog. Instead it uses pg_attribute.attoptions, which is used for table column storage options but unused for index attributes. In order to evade changing signature of each opclass support function, we implement unified way to pass options to opclass support functions. Options are set to fn_expr as the constant bytea expression. It's possible due to the fact that opclass support functions are executed outside of expressions, so fn_expr is unused for them. This commit comes with some examples of opclass options usage. We parametrize signature length in GiST. That applies to multiple opclasses: tsvector_ops, gist__intbig_ops, gist_ltree_ops, gist__ltree_ops, gist_trgm_ops and gist_hstore_ops. Also we parametrize maximum number of integer ranges for gist__int_ops. However, the main future usage of this feature is expected to be json, where users would be able to specify which way to index particular json parts. Catversion is bumped. Discussion: https://postgr.es/m/d22c3a18-31c7-1879-fc11-4c1ce2f5e5af%40postgrespro.ru Author: Nikita Glukhov, revised by me Reviwed-by: Nikolay Shaplov, Robert Haas, Tom Lane, Tomas Vondra, Alvaro Herrera
2020-03-30 18:17:11 +02:00
/* gist_ltree_ops and gist__ltree_ops opclass options */
typedef struct
{
int32 vl_len_; /* varlena header (do not touch directly!) */
int siglen; /* signature length in bytes */
} LtreeGistOptions;
/* type of key is the same to ltree_gist */
#endif