rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Ensure that RowExprs and whole-row Vars produce the expected column names. 

At one time it wasn't terribly important what column names were associated
with the fields of a composite Datum, but since the introduction of
operations like row_to_json(), it's important that looking up the rowtype
ID embedded in the Datum returns the column names that users would expect.
That did not work terribly well before this patch: you could get the column
names of the underlying table, or column aliases from any level of the
query, depending on minor details of the plan tree.  You could even get
totally empty field names, which is disastrous for cases like row_to_json().

To fix this for whole-row Vars, look to the RTE referenced by the Var, and
make sure its column aliases are applied to the rowtype associated with
the result Datums.  This is a tad scary because we might have to return
a transient RECORD type even though the Var is declared as having some
named rowtype.  In principle it should be all right because the record
type will still be physically compatible with the named rowtype; but
I had to weaken one Assert in ExecEvalConvertRowtype, and there might be
third-party code containing similar assumptions.

Similarly, RowExprs have to be willing to override the column names coming
from a named composite result type and produce a RECORD when the column
aliases visible at the site of the RowExpr differ from the underlying
table's column names.

In passing, revert the decision made in commit 398f70ec07 to add
an alias-list argument to ExecTypeFromExprList: better to provide that
functionality in a separate function.  This also reverts most of the code
changes in d685814835, which we don't need because we're no longer
depending on the tupdesc found in the child plan node's result slot to be
blessed.

Back-patch to 9.4, but not earlier, since this solution changes the results
in some cases that users might not have realized were buggy.  We'll apply a
more restricted form of this patch in older branches.
This commit is contained in:
Tom Lane 2014-11-10 15:21:09 -05:00
parent 1e0b4365c2
commit bf7ca15875
8 changed files with 340 additions and 73 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

115
src/backend/executor/execQual.c
View File

@ -50,6 +50,7 @@
#include "nodes/nodeFuncs.h"
#include "optimizer/planner.h"
#include "parser/parse_coerce.h"
#include "parser/parsetree.h"
#include "pgstat.h"
#include "utils/acl.h"
#include "utils/builtins.h"
@ -712,6 +713,8 @@ ExecEvalWholeRowVar(WholeRowVarExprState *wrvstate, ExprContext *econtext,
{
Var *variable = (Var *) wrvstate->xprstate.expr;
TupleTableSlot *slot;
TupleDesc output_tupdesc;
MemoryContext oldcontext;
bool needslow = false;
if (isDone)
@ -787,8 +790,6 @@ ExecEvalWholeRowVar(WholeRowVarExprState *wrvstate, ExprContext *econtext,
/* If so, build the junkfilter in the query memory context */
if (junk_filter_needed)
{
MemoryContext oldcontext;
oldcontext = MemoryContextSwitchTo(econtext->ecxt_per_query_memory);
wrvstate->wrv_junkFilter =
ExecInitJunkFilter(subplan->plan->targetlist,
@ -860,10 +861,60 @@ ExecEvalWholeRowVar(WholeRowVarExprState *wrvstate, ExprContext *econtext,
needslow = true; /* need runtime check for null */
}
/*
* Use the variable's declared rowtype as the descriptor for the
* output values, modulo possibly assigning new column names below. In
* particular, we *must* absorb any attisdropped markings.
*/
oldcontext = MemoryContextSwitchTo(econtext->ecxt_per_query_memory);
output_tupdesc = CreateTupleDescCopy(var_tupdesc);
MemoryContextSwitchTo(oldcontext);
ReleaseTupleDesc(var_tupdesc);
}
else
{
/*
* In the RECORD case, we use the input slot's rowtype as the
* descriptor for the output values, modulo possibly assigning new
* column names below.
*/
oldcontext = MemoryContextSwitchTo(econtext->ecxt_per_query_memory);
output_tupdesc = CreateTupleDescCopy(slot->tts_tupleDescriptor);
MemoryContextSwitchTo(oldcontext);
}
/* Skip the checking on future executions of node */
/*
* Construct a tuple descriptor for the composite values we'll produce,
* and make sure its record type is "blessed". The main reason to do this
* is to be sure that operations such as row_to_json() will see the
* desired column names when they look up the descriptor from the type
* information embedded in the composite values.
*
* We already got the correct physical datatype info above, but now we
* should try to find the source RTE and adopt its column aliases, in case
* they are different from the original rowtype's names. For example, in
* "SELECT foo(t) FROM tab t(x,y)", the first two columns in the composite
* output should be named "x" and "y" regardless of tab's column names.
*
* If we can't locate the RTE, assume the column names we've got are OK.
* (As of this writing, the only cases where we can't locate the RTE are
* in execution of trigger WHEN clauses, and then the Var will have the
* trigger's relation's rowtype, so its names are fine.)
*/
if (econtext->ecxt_estate &&
variable->varno <= list_length(econtext->ecxt_estate->es_range_table))
{
RangeTblEntry *rte = rt_fetch(variable->varno,
econtext->ecxt_estate->es_range_table);
ExecTypeSetColNames(output_tupdesc, rte->eref->colnames);
}
/* Bless the tupdesc if needed, and save it in the execution state */
wrvstate->wrv_tupdesc = BlessTupleDesc(output_tupdesc);
/* Skip all the above on future executions of node */
if (needslow)
wrvstate->xprstate.evalfunc = (ExprStateEvalFunc) ExecEvalWholeRowSlow;
else
@ -886,7 +937,6 @@ ExecEvalWholeRowFast(WholeRowVarExprState *wrvstate, ExprContext *econtext,
{
Var *variable = (Var *) wrvstate->xprstate.expr;
TupleTableSlot *slot;
TupleDesc slot_tupdesc;
HeapTupleHeader dtuple;
if (isDone)
@ -916,34 +966,16 @@ ExecEvalWholeRowFast(WholeRowVarExprState *wrvstate, ExprContext *econtext,
if (wrvstate->wrv_junkFilter != NULL)
slot = ExecFilterJunk(wrvstate->wrv_junkFilter, slot);
/*
* If it's a RECORD Var, we'll use the slot's type ID info. It's likely
* that the slot's type is also RECORD; if so, make sure it's been
* "blessed", so that the Datum can be interpreted later. (Note: we must
* do this here, not in ExecEvalWholeRowVar, because some plan trees may
* return different slots at different times. We have to be ready to
* bless additional slots during the run.)
*/
slot_tupdesc = slot->tts_tupleDescriptor;
if (variable->vartype == RECORDOID &&
slot_tupdesc->tdtypeid == RECORDOID &&
slot_tupdesc->tdtypmod < 0)
assign_record_type_typmod(slot_tupdesc);
/*
* Copy the slot tuple and make sure any toasted fields get detoasted.
*/
dtuple = DatumGetHeapTupleHeader(ExecFetchSlotTupleDatum(slot));
/*
* If the Var identifies a named composite type, label the datum with that
* type; otherwise we'll use the slot's info.
* Label the datum with the composite type info we identified before.
*/
if (variable->vartype != RECORDOID)
{
HeapTupleHeaderSetTypeId(dtuple, variable->vartype);
HeapTupleHeaderSetTypMod(dtuple, variable->vartypmod);
}
HeapTupleHeaderSetTypeId(dtuple, wrvstate->wrv_tupdesc->tdtypeid);
HeapTupleHeaderSetTypMod(dtuple, wrvstate->wrv_tupdesc->tdtypmod);
return PointerGetDatum(dtuple);
}
@ -997,8 +1029,9 @@ ExecEvalWholeRowSlow(WholeRowVarExprState *wrvstate, ExprContext *econtext,
tuple = ExecFetchSlotTuple(slot);
tupleDesc = slot->tts_tupleDescriptor;
/* wrv_tupdesc is a good enough representation of the Var's rowtype */
Assert(variable->vartype != RECORDOID);
var_tupdesc = lookup_rowtype_tupdesc(variable->vartype, -1);
var_tupdesc = wrvstate->wrv_tupdesc;
/* Check to see if any dropped attributes are non-null */
for (i = 0; i < var_tupdesc->natts; i++)
@ -1025,12 +1058,10 @@ ExecEvalWholeRowSlow(WholeRowVarExprState *wrvstate, ExprContext *econtext,
dtuple = DatumGetHeapTupleHeader(ExecFetchSlotTupleDatum(slot));
/*
* Reset datum's type ID fields to match the Var.
* Label the datum with the composite type info we identified before.
*/
HeapTupleHeaderSetTypeId(dtuple, variable->vartype);
HeapTupleHeaderSetTypMod(dtuple, variable->vartypmod);
ReleaseTupleDesc(var_tupdesc);
HeapTupleHeaderSetTypeId(dtuple, wrvstate->wrv_tupdesc->tdtypeid);
HeapTupleHeaderSetTypMod(dtuple, wrvstate->wrv_tupdesc->tdtypmod);
return PointerGetDatum(dtuple);
}
@ -2850,8 +2881,14 @@ ExecEvalConvertRowtype(ConvertRowtypeExprState *cstate,
cstate->initialized = false;
}
Assert(HeapTupleHeaderGetTypeId(tuple) == cstate->indesc->tdtypeid);
Assert(HeapTupleHeaderGetTypMod(tuple) == cstate->indesc->tdtypmod);
/*
* We used to be able to assert that incoming tuples are marked with
* exactly the rowtype of cstate->indesc. However, now that
* ExecEvalWholeRowVar might change the tuples' marking to plain RECORD
* due to inserting aliases, we can only make this weak test:
*/
Assert(HeapTupleHeaderGetTypeId(tuple) == cstate->indesc->tdtypeid ||
HeapTupleHeaderGetTypeId(tuple) == RECORDOID);
/* if first time through, initialize conversion map */
if (!cstate->initialized)
@ -4375,6 +4412,7 @@ ExecInitExpr(Expr *node, PlanState *parent)
WholeRowVarExprState *wstate = makeNode(WholeRowVarExprState);
wstate->parent = parent;
wstate->wrv_tupdesc = NULL;
wstate->wrv_junkFilter = NULL;
state = (ExprState *) wstate;
state->evalfunc = (ExprStateEvalFunc) ExecEvalWholeRowVar;
@ -4778,17 +4816,18 @@ ExecInitExpr(Expr *node, PlanState *parent)
/* Build tupdesc to describe result tuples */
if (rowexpr->row_typeid == RECORDOID)
{
/* generic record, use runtime type assignment */
rstate->tupdesc = ExecTypeFromExprList(rowexpr->args,
rowexpr->colnames);
BlessTupleDesc(rstate->tupdesc);
/* we won't need to redo this at runtime */
/* generic record, use types of given expressions */
rstate->tupdesc = ExecTypeFromExprList(rowexpr->args);
}
else
{
/* it's been cast to a named type, use that */
rstate->tupdesc = lookup_rowtype_tupdesc_copy(rowexpr->row_typeid, -1);
}
/* In either case, adopt RowExpr's column aliases */
ExecTypeSetColNames(rstate->tupdesc, rowexpr->colnames);
/* Bless the tupdesc in case it's now of type RECORD */
BlessTupleDesc(rstate->tupdesc);
/* Set up evaluation, skipping any deleted columns */
Assert(list_length(rowexpr->args) <= rstate->tupdesc->natts);
attrs = rstate->tupdesc->attrs;

65
src/backend/executor/execTuples.c
View File

@ -943,28 +943,25 @@ ExecTypeFromTLInternal(List *targetList, bool hasoid, bool skipjunk)
/*
* ExecTypeFromExprList - build a tuple descriptor from a list of Exprs
*
* Caller must also supply a list of field names (String nodes).
* This is roughly like ExecTypeFromTL, but we work from bare expressions
* not TargetEntrys. No names are attached to the tupledesc's columns.
*/
TupleDesc
ExecTypeFromExprList(List *exprList, List *namesList)
ExecTypeFromExprList(List *exprList)
{
TupleDesc typeInfo;
ListCell *le;
ListCell *ln;
ListCell *lc;
int cur_resno = 1;
Assert(list_length(exprList) == list_length(namesList));
typeInfo = CreateTemplateTupleDesc(list_length(exprList), false);
forboth(le, exprList, ln, namesList)
foreach(lc, exprList)
{
Node *e = lfirst(le);
char *n = strVal(lfirst(ln));
Node *e = lfirst(lc);
TupleDescInitEntry(typeInfo,
cur_resno,
n,
NULL,
exprType(e),
exprTypmod(e),
0);
@ -977,6 +974,54 @@ ExecTypeFromExprList(List *exprList, List *namesList)
return typeInfo;
}
/*
* ExecTypeSetColNames - set column names in a TupleDesc
*
* Column names must be provided as an alias list (list of String nodes).
*
* For some callers, the supplied tupdesc has a named rowtype (not RECORD)
* and it is moderately likely that the alias list matches the column names
* already present in the tupdesc. If we do change any column names then
* we must reset the tupdesc's type to anonymous RECORD; but we avoid doing
* so if no names change.
*/
void
ExecTypeSetColNames(TupleDesc typeInfo, List *namesList)
{
bool modified = false;
int colno = 0;
ListCell *lc;
foreach(lc, namesList)
{
char *cname = strVal(lfirst(lc));
Form_pg_attribute attr;
/* Guard against too-long names list */
if (colno >= typeInfo->natts)
break;
attr = typeInfo->attrs[colno++];
/* Ignore empty aliases (these must be for dropped columns) */
if (cname[0] == '\0')
continue;
/* Change tupdesc only if alias is actually different */
if (strcmp(cname, NameStr(attr->attname)) != 0)
{
namestrcpy(&(attr->attname), cname);
modified = true;
}
}
/* If we modified the tupdesc, it's now a new record type */
if (modified)
{
typeInfo->tdtypeid = RECORDOID;
typeInfo->tdtypmod = -1;
}
}
/*
* BlessTupleDesc - make a completed tuple descriptor useful for SRFs
*

19
src/backend/executor/nodeFunctionscan.c
View File

@ -26,7 +26,6 @@
#include "executor/nodeFunctionscan.h"
#include "funcapi.h"
#include "nodes/nodeFuncs.h"
#include "parser/parsetree.h"
#include "utils/builtins.h"
#include "utils/memutils.h"
@ -279,8 +278,6 @@ FunctionScanState *
ExecInitFunctionScan(FunctionScan *node, EState *estate, int eflags)
{
FunctionScanState *scanstate;
RangeTblEntry *rte = rt_fetch(node->scan.scanrelid,
estate->es_range_table);
int nfuncs = list_length(node->functions);
TupleDesc scan_tupdesc;
int i,
@ -494,22 +491,6 @@ ExecInitFunctionScan(FunctionScan *node, EState *estate, int eflags)
Assert(attno == natts);
}
/*
* Make sure the scan result tupdesc has the column names the query
* expects. This affects the output of constructs like row_to_json which
* read the column names from the passed-in tupdesc.
*/
i = 0;
foreach(lc, rte->eref->colnames)
{
char *attname = strVal(lfirst(lc));
if (i >= scan_tupdesc->natts)
break; /* shouldn't happen, but just in case */
namestrcpy(&(scan_tupdesc->attrs[i]->attname), attname);
i++;
}
ExecAssignScanType(&scanstate->ss, scan_tupdesc);
/*

6
src/backend/executor/nodeValuesscan.c
View File

@ -25,7 +25,6 @@
#include "executor/executor.h"
#include "executor/nodeValuesscan.h"
#include "parser/parsetree.h"
static TupleTableSlot *ValuesNext(ValuesScanState *node);
@ -189,8 +188,6 @@ ValuesScanState *
ExecInitValuesScan(ValuesScan *node, EState *estate, int eflags)
{
ValuesScanState *scanstate;
RangeTblEntry *rte = rt_fetch(node->scan.scanrelid,
estate->es_range_table);
TupleDesc tupdesc;
ListCell *vtl;
int i;
@ -242,8 +239,7 @@ ExecInitValuesScan(ValuesScan *node, EState *estate, int eflags)
/*
* get info about values list
*/
tupdesc = ExecTypeFromExprList((List *) linitial(node->values_lists),
rte->eref->colnames);
tupdesc = ExecTypeFromExprList((List *) linitial(node->values_lists));
ExecAssignScanType(&scanstate->ss, tupdesc);

3
src/include/executor/executor.h
View File

@ -268,7 +268,8 @@ extern TupleTableSlot *ExecInitNullTupleSlot(EState *estate,
TupleDesc tupType);
extern TupleDesc ExecTypeFromTL(List *targetList, bool hasoid);
extern TupleDesc ExecCleanTypeFromTL(List *targetList, bool hasoid);
extern TupleDesc ExecTypeFromExprList(List *exprList, List *namesList);
extern TupleDesc ExecTypeFromExprList(List *exprList);
extern void ExecTypeSetColNames(TupleDesc typeInfo, List *namesList);
extern void UpdateChangedParamSet(PlanState *node, Bitmapset *newchg);
typedef struct TupOutputState

1
src/include/nodes/execnodes.h
View File

@ -578,6 +578,7 @@ typedef struct WholeRowVarExprState
{
ExprState xprstate;
struct PlanState *parent; /* parent PlanState, or NULL if none */
TupleDesc wrv_tupdesc; /* descriptor for resulting tuples */
JunkFilter *wrv_junkFilter; /* JunkFilter to remove resjunk cols */
} WholeRowVarExprState;

160
src/test/regress/expected/rowtypes.out
View File

@ -474,3 +474,163 @@ select (row('Jim', 'Beam')).text; -- error
ERROR: could not identify column "text" in record data type
LINE 1: select (row('Jim', 'Beam')).text;
^
--
-- Test that composite values are seen to have the correct column names
-- (bug #11210 and other reports)
--
select row_to_json(i) from int8_tbl i;
row_to_json
------------------------------------------------
{"q1":123,"q2":456}
{"q1":123,"q2":4567890123456789}
{"q1":4567890123456789,"q2":123}
{"q1":4567890123456789,"q2":4567890123456789}
{"q1":4567890123456789,"q2":-4567890123456789}
(5 rows)
select row_to_json(i) from int8_tbl i(x,y);
row_to_json
----------------------------------------------
{"x":123,"y":456}
{"x":123,"y":4567890123456789}
{"x":4567890123456789,"y":123}
{"x":4567890123456789,"y":4567890123456789}
{"x":4567890123456789,"y":-4567890123456789}
(5 rows)
create temp view vv1 as select * from int8_tbl;
select row_to_json(i) from vv1 i;
row_to_json
------------------------------------------------
{"q1":123,"q2":456}
{"q1":123,"q2":4567890123456789}
{"q1":4567890123456789,"q2":123}
{"q1":4567890123456789,"q2":4567890123456789}
{"q1":4567890123456789,"q2":-4567890123456789}
(5 rows)
select row_to_json(i) from vv1 i(x,y);
row_to_json
----------------------------------------------
{"x":123,"y":456}
{"x":123,"y":4567890123456789}
{"x":4567890123456789,"y":123}
{"x":4567890123456789,"y":4567890123456789}
{"x":4567890123456789,"y":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1, q2 from int8_tbl) as ss;
row_to_json
------------------------------------------------
{"q1":123,"q2":456}
{"q1":123,"q2":4567890123456789}
{"q1":4567890123456789,"q2":123}
{"q1":4567890123456789,"q2":4567890123456789}
{"q1":4567890123456789,"q2":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1, q2 from int8_tbl offset 0) as ss;
row_to_json
------------------------------------------------
{"q1":123,"q2":456}
{"q1":123,"q2":4567890123456789}
{"q1":4567890123456789,"q2":123}
{"q1":4567890123456789,"q2":4567890123456789}
{"q1":4567890123456789,"q2":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl) as ss;
row_to_json
----------------------------------------------
{"a":123,"b":456}
{"a":123,"b":4567890123456789}
{"a":4567890123456789,"b":123}
{"a":4567890123456789,"b":4567890123456789}
{"a":4567890123456789,"b":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl offset 0) as ss;
row_to_json
----------------------------------------------
{"a":123,"b":456}
{"a":123,"b":4567890123456789}
{"a":4567890123456789,"b":123}
{"a":4567890123456789,"b":4567890123456789}
{"a":4567890123456789,"b":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl) as ss(x,y);
row_to_json
----------------------------------------------
{"x":123,"y":456}
{"x":123,"y":4567890123456789}
{"x":4567890123456789,"y":123}
{"x":4567890123456789,"y":4567890123456789}
{"x":4567890123456789,"y":-4567890123456789}
(5 rows)
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl offset 0) as ss(x,y);
row_to_json
----------------------------------------------
{"x":123,"y":456}
{"x":123,"y":4567890123456789}
{"x":4567890123456789,"y":123}
{"x":4567890123456789,"y":4567890123456789}
{"x":4567890123456789,"y":-4567890123456789}
(5 rows)
explain (costs off)
select row_to_json(q) from
(select thousand, tenthous from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
QUERY PLAN
-------------------------------------------------------------
Subquery Scan on q
-> Index Only Scan using tenk1_thous_tenthous on tenk1
Index Cond: ((thousand = 42) AND (tenthous < 2000))
(3 rows)
select row_to_json(q) from
(select thousand, tenthous from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
row_to_json
---------------------------------
{"thousand":42,"tenthous":42}
{"thousand":42,"tenthous":1042}
(2 rows)
select row_to_json(q) from
(select thousand as x, tenthous as y from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
row_to_json
-------------------
{"x":42,"y":42}
{"x":42,"y":1042}
(2 rows)
select row_to_json(q) from
(select thousand as x, tenthous as y from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q(a,b);
row_to_json
-------------------
{"a":42,"b":42}
{"a":42,"b":1042}
(2 rows)
create temp table tt1 as select * from int8_tbl limit 2;
create temp table tt2 () inherits(tt1);
insert into tt2 values(0,0);
select row_to_json(r) from (select q2,q1 from tt1 offset 0) r;
row_to_json
----------------------------------
{"q2":456,"q1":123}
{"q2":4567890123456789,"q1":123}
{"q2":0,"q1":0}
(3 rows)

44
src/test/regress/sql/rowtypes.sql
View File

@ -227,3 +227,47 @@ select cast (row('Jim', 'Beam') as text);
select (row('Jim', 'Beam'))::text;
select text(row('Jim', 'Beam')); -- error
select (row('Jim', 'Beam')).text; -- error
--
-- Test that composite values are seen to have the correct column names
-- (bug #11210 and other reports)
--
select row_to_json(i) from int8_tbl i;
select row_to_json(i) from int8_tbl i(x,y);
create temp view vv1 as select * from int8_tbl;
select row_to_json(i) from vv1 i;
select row_to_json(i) from vv1 i(x,y);
select row_to_json(ss) from
(select q1, q2 from int8_tbl) as ss;
select row_to_json(ss) from
(select q1, q2 from int8_tbl offset 0) as ss;
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl) as ss;
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl offset 0) as ss;
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl) as ss(x,y);
select row_to_json(ss) from
(select q1 as a, q2 as b from int8_tbl offset 0) as ss(x,y);
explain (costs off)
select row_to_json(q) from
(select thousand, tenthous from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
select row_to_json(q) from
(select thousand, tenthous from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
select row_to_json(q) from
(select thousand as x, tenthous as y from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q;
select row_to_json(q) from
(select thousand as x, tenthous as y from tenk1
where thousand = 42 and tenthous < 2000 offset 0) q(a,b);
create temp table tt1 as select * from int8_tbl limit 2;
create temp table tt2 () inherits(tt1);
insert into tt2 values(0,0);
select row_to_json(r) from (select q2,q1 from tt1 offset 0) r;