Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit39df0f1

Browse files
committed
Allow planner to use expression-index stats for function calls in WHERE.
Previously, a function call appearing at the top level of WHERE had ahard-wired selectivity estimate of 0.3333333, a kludge conveniently datedin the source code itself to July 1992. The expectation at the time wasthat somebody would soon implement estimator support functions analogousto those for operators; but no such code has appeared, nor does it seemlikely to in the near future. We do have an alternative solution though,at least for immutable functions on single relations: creating anexpression index on the function call will allow ANALYZE to gather statsabout the function's selectivity. But the code in clause_selectivity()failed to make use of such data even if it exists.Refactor so that that will happen. I chose to make it try this techniquefor any clause type for which clause_selectivity() doesn't have a specialcase, not just functions. To avoid adding unnecessary overhead in thecommon case where we don't learn anything new, make selfuncs.c provide anAPI that hooks directly to examine_variable() and then var_eq_const(),rather than the previous coding which laboriously constructed an OpExpronly so that it could be expensively deconstructed again.I preserved the behavior that the default estimate for a function callis 0.3333333. (For any other expression node type, it's 0.5, as before.)I had originally thought to make the default be 0.5 across the board, butchanging a default estimate that's survived for twenty-three years seemslike something not to do without a lot more testing than I care to putinto it right now.Per a complaint from Jehan-Guillaume de Rorthais. Back-patch into 9.5,but not further, at least for the moment.
1 parent76f965f commit39df0f1

File tree

3 files changed

+59
-32
lines changed

3 files changed

+59
-32
lines changed

‎src/backend/optimizer/path/clausesel.c

Lines changed: 13 additions & 32 deletions
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,6 @@
1414
*/
1515
#include"postgres.h"
1616

17-
#include"catalog/pg_operator.h"
1817
#include"nodes/makefuncs.h"
1918
#include"optimizer/clauses.h"
2019
#include"optimizer/cost.h"
@@ -568,18 +567,8 @@ clause_selectivity(PlannerInfo *root,
568567
if (var->varlevelsup==0&&
569568
(varRelid==0||varRelid== (int)var->varno))
570569
{
571-
/*
572-
* A Var at the top of a clause must be a bool Var. This is
573-
* equivalent to the clause reln.attribute = 't', so we compute
574-
* the selectivity as if that is what we have.
575-
*/
576-
s1=restriction_selectivity(root,
577-
BooleanEqualOperator,
578-
list_make2(var,
579-
makeBoolConst(true,
580-
false)),
581-
InvalidOid,
582-
varRelid);
570+
/* Use the restriction selectivity function for a bool Var */
571+
s1=boolvarsel(root, (Node*)var,varRelid);
583572
}
584573
}
585574
elseif (IsA(clause,Const))
@@ -680,25 +669,6 @@ clause_selectivity(PlannerInfo *root,
680669
if (IsA(clause,DistinctExpr))
681670
s1=1.0-s1;
682671
}
683-
elseif (is_funcclause(clause))
684-
{
685-
/*
686-
* This is not an operator, so we guess at the selectivity. THIS IS A
687-
* HACK TO GET V4 OUT THE DOOR. FUNCS SHOULD BE ABLE TO HAVE
688-
* SELECTIVITIES THEMSELVES. -- JMH 7/9/92
689-
*/
690-
s1= (Selectivity)0.3333333;
691-
}
692-
#ifdefNOT_USED
693-
elseif (IsA(clause,SubPlan)||
694-
IsA(clause,AlternativeSubPlan))
695-
{
696-
/*
697-
* Just for the moment! FIX ME! - vadim 02/04/98
698-
*/
699-
s1= (Selectivity)0.5;
700-
}
701-
#endif
702672
elseif (IsA(clause,ScalarArrayOpExpr))
703673
{
704674
/* Use node specific selectivity calculation function */
@@ -766,6 +736,17 @@ clause_selectivity(PlannerInfo *root,
766736
jointype,
767737
sjinfo);
768738
}
739+
else
740+
{
741+
/*
742+
* For anything else, see if we can consider it as a boolean variable.
743+
* This only works if it's an immutable expression in Vars of a single
744+
* relation; but there's no point in us checking that here because
745+
* boolvarsel() will do it internally, and return a suitable default
746+
* selectivity if not.
747+
*/
748+
s1=boolvarsel(root,clause,varRelid);
749+
}
769750

770751
/* Cache the result if possible */
771752
if (cacheable)

‎src/backend/utils/adt/selfuncs.c

Lines changed: 45 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -105,6 +105,7 @@
105105
#include"access/sysattr.h"
106106
#include"catalog/index.h"
107107
#include"catalog/pg_collation.h"
108+
#include"catalog/pg_operator.h"
108109
#include"catalog/pg_opfamily.h"
109110
#include"catalog/pg_statistic.h"
110111
#include"catalog/pg_type.h"
@@ -1439,6 +1440,50 @@ icnlikesel(PG_FUNCTION_ARGS)
14391440
PG_RETURN_FLOAT8(patternsel(fcinfo,Pattern_Type_Like_IC, true));
14401441
}
14411442

1443+
/*
1444+
*boolvarsel- Selectivity of Boolean variable.
1445+
*
1446+
* This can actually be called on any boolean-valued expression. If it
1447+
* involves only Vars of the specified relation, and if there are statistics
1448+
* about the Var or expression (the latter is possible if it's indexed) then
1449+
* we'll produce a real estimate; otherwise it's just a default.
1450+
*/
1451+
Selectivity
1452+
boolvarsel(PlannerInfo*root,Node*arg,intvarRelid)
1453+
{
1454+
VariableStatDatavardata;
1455+
doubleselec;
1456+
1457+
examine_variable(root,arg,varRelid,&vardata);
1458+
if (HeapTupleIsValid(vardata.statsTuple))
1459+
{
1460+
/*
1461+
* A boolean variable V is equivalent to the clause V = 't', so we
1462+
* compute the selectivity as if that is what we have.
1463+
*/
1464+
selec=var_eq_const(&vardata,BooleanEqualOperator,
1465+
BoolGetDatum(true), false, true);
1466+
}
1467+
elseif (is_funcclause(arg))
1468+
{
1469+
/*
1470+
* If we have no stats and it's a function call, estimate 0.3333333.
1471+
* This seems a pretty unprincipled choice, but Postgres has been
1472+
* using that estimate for function calls since 1992. The hoariness
1473+
* of this behavior suggests that we should not be in too much hurry
1474+
* to use another value.
1475+
*/
1476+
selec=0.3333333;
1477+
}
1478+
else
1479+
{
1480+
/* Otherwise, the default estimate is 0.5 */
1481+
selec=0.5;
1482+
}
1483+
ReleaseVariableStats(vardata);
1484+
returnselec;
1485+
}
1486+
14421487
/*
14431488
*booltestsel- Selectivity of BooleanTest Node.
14441489
*/

‎src/include/utils/selfuncs.h

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -164,6 +164,7 @@ extern Datum icregexnejoinsel(PG_FUNCTION_ARGS);
164164
externDatumnlikejoinsel(PG_FUNCTION_ARGS);
165165
externDatumicnlikejoinsel(PG_FUNCTION_ARGS);
166166

167+
externSelectivityboolvarsel(PlannerInfo*root,Node*arg,intvarRelid);
167168
externSelectivitybooltestsel(PlannerInfo*root,BoolTestTypebooltesttype,
168169
Node*arg,intvarRelid,
169170
JoinTypejointype,SpecialJoinInfo*sjinfo);

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp