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

Commit1e9a6ba

Browse files
committed
Don't try to remove duplicate OR-subclauses in create_bitmap_subplan and
make_restrictinfo_from_bitmapqual. The likelihood of finding duplicatesseems much less than in the AND-subclause case, and the cost much higher,because OR lists with hundreds or even thousands of subclauses are notuncommon. Per discussion with Ilia Kantor and andrew@supernews.
1 parent23e2f9e commit1e9a6ba

File tree

2 files changed

+24
-16
lines changed

2 files changed

+24
-16
lines changed

‎src/backend/optimizer/plan/createplan.c

Lines changed: 12 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -10,7 +10,7 @@
1010
*
1111
*
1212
* IDENTIFICATION
13-
* $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.199 2005/10/06 16:01:54 tgl Exp $
13+
* $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.200 2005/10/13 00:06:46 tgl Exp $
1414
*
1515
*-------------------------------------------------------------------------
1616
*/
@@ -1046,9 +1046,13 @@ create_bitmap_subplan(PlannerInfo *root, Path *bitmapqual,
10461046
ListCell*l;
10471047

10481048
/*
1049-
* Here, we detect both obvious redundancies and qual-free subplans.
1050-
* A qual-free subplan would cause us to generate "... OR true ..."
1051-
* which we may as well reduce to just "true".
1049+
* Here, we only detect qual-free subplans. A qual-free subplan would
1050+
* cause us to generate "... OR true ..." which we may as well reduce
1051+
* to just "true". We do not try to eliminate redundant subclauses
1052+
* because (a) it's not as likely as in the AND case, and (b) we might
1053+
* well be working with hundreds or even thousands of OR conditions,
1054+
* perhaps from a long IN list. The performance of list_append_unique
1055+
* would be unacceptable.
10521056
*/
10531057
foreach(l,opath->bitmapquals)
10541058
{
@@ -1062,13 +1066,13 @@ create_bitmap_subplan(PlannerInfo *root, Path *bitmapqual,
10621066
if (subqual==NIL)
10631067
const_true_subqual= true;
10641068
elseif (!const_true_subqual)
1065-
subquals=list_append_unique(subquals,
1066-
make_ands_explicit(subqual));
1069+
subquals=lappend(subquals,
1070+
make_ands_explicit(subqual));
10671071
if (subindexqual==NIL)
10681072
const_true_subindexqual= true;
10691073
elseif (!const_true_subindexqual)
1070-
subindexquals=list_append_unique(subindexquals,
1071-
make_ands_explicit(subindexqual));
1074+
subindexquals=lappend(subindexquals,
1075+
make_ands_explicit(subindexqual));
10721076
}
10731077
plan= (Plan*)make_bitmap_or(subplans);
10741078
plan->startup_cost=opath->path.startup_cost;

‎src/backend/optimizer/util/restrictinfo.c

Lines changed: 12 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
*
99
*
1010
* IDENTIFICATION
11-
* $PostgreSQL: pgsql/src/backend/optimizer/util/restrictinfo.c,v 1.39 2005/07/28 20:26:21 tgl Exp $
11+
* $PostgreSQL: pgsql/src/backend/optimizer/util/restrictinfo.c,v 1.40 2005/10/13 00:06:46 tgl Exp $
1212
*
1313
*-------------------------------------------------------------------------
1414
*/
@@ -123,9 +123,13 @@ make_restrictinfo_from_bitmapqual(Path *bitmapqual,
123123
List*withoutris=NIL;
124124

125125
/*
126-
* Here, we detect both obvious redundancies and qual-free subplans.
127-
* A qual-free subplan would cause us to generate "... OR true ..."
128-
* which we may as well reduce to just "true".
126+
* Here, we only detect qual-free subplans. A qual-free subplan would
127+
* cause us to generate "... OR true ..." which we may as well reduce
128+
* to just "true". We do not try to eliminate redundant subclauses
129+
* because (a) it's not as likely as in the AND case, and (b) we might
130+
* well be working with hundreds or even thousands of OR conditions,
131+
* perhaps from a long IN list. The performance of list_append_unique
132+
* would be unacceptable.
129133
*/
130134
foreach(l,opath->bitmapquals)
131135
{
@@ -144,12 +148,12 @@ make_restrictinfo_from_bitmapqual(Path *bitmapqual,
144148
returnNIL;
145149
}
146150
/* Create AND subclause with RestrictInfos */
147-
withris=list_append_unique(withris,
148-
make_ands_explicit(sublist));
151+
withris=lappend(withris,
152+
make_ands_explicit(sublist));
149153
/* And one without */
150154
sublist=get_actual_clauses(sublist);
151-
withoutris=list_append_unique(withoutris,
152-
make_ands_explicit(sublist));
155+
withoutris=lappend(withoutris,
156+
make_ands_explicit(sublist));
153157
}
154158

155159
/*

0 commit comments

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