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

Commitf38fbf3

Browse files
committed
If we expect a hash join to be performed in multiple batches, suppress
"physical tlist" optimization on the outer relation (ie, force a projectionstep to occur in its scan). This avoids storing useless column values whenthe outer relation's tuples are written to temporary batch files.Modified version of a patch by Michael Henderson and Ramon Lawrence.
1 parentee4c187 commitf38fbf3

File tree

5 files changed

+25
-6
lines changed

5 files changed

+25
-6
lines changed

‎src/backend/nodes/outfuncs.c

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
*
99
*
1010
* IDENTIFICATION
11-
* $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.355 2009/03/21 00:04:39 tgl Exp $
11+
* $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.356 2009/03/26 17:15:34 tgl Exp $
1212
*
1313
* NOTES
1414
* Every node type that can appear in stored rules' parsetrees *must*
@@ -1448,6 +1448,7 @@ _outHashPath(StringInfo str, HashPath *node)
14481448
_outJoinPathInfo(str, (JoinPath*)node);
14491449

14501450
WRITE_NODE_FIELD(path_hashclauses);
1451+
WRITE_INT_FIELD(num_batches);
14511452
}
14521453

14531454
staticvoid

‎src/backend/optimizer/path/costsize.c

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -54,7 +54,7 @@
5454
* Portions Copyright (c) 1994, Regents of the University of California
5555
*
5656
* IDENTIFICATION
57-
* $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.205 2009/03/21 00:04:39 tgl Exp $
57+
* $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.206 2009/03/26 17:15:35 tgl Exp $
5858
*
5959
*-------------------------------------------------------------------------
6060
*/
@@ -1880,6 +1880,8 @@ cost_hashjoin(HashPath *path, PlannerInfo *root, SpecialJoinInfo *sjinfo)
18801880
&numbatches,
18811881
&num_skew_mcvs);
18821882
virtualbuckets= (double)numbuckets*(double)numbatches;
1883+
/* mark the path with estimated # of batches */
1884+
path->num_batches=numbatches;
18831885

18841886
/*
18851887
* Determine bucketsize fraction for inner relation. We use the smallest

‎src/backend/optimizer/plan/createplan.c

Lines changed: 5 additions & 1 deletion
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 1.256 2009/03/21 00:04:39 tgl Exp $
13+
* $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.257 2009/03/26 17:15:35 tgl Exp $
1414
*
1515
*-------------------------------------------------------------------------
1616
*/
@@ -1910,6 +1910,10 @@ create_hashjoin_plan(PlannerInfo *root,
19101910
/* We don't want any excess columns in the hashed tuples */
19111911
disuse_physical_tlist(inner_plan,best_path->jpath.innerjoinpath);
19121912

1913+
/* If we expect batching, suppress excess columns in outer tuples too */
1914+
if (best_path->num_batches>1)
1915+
disuse_physical_tlist(outer_plan,best_path->jpath.outerjoinpath);
1916+
19131917
/*
19141918
* If there is a single join clause and we can identify the outer
19151919
* variable as a simple column reference, supply its identity for

‎src/backend/optimizer/util/pathnode.c

Lines changed: 13 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
*
99
*
1010
* IDENTIFICATION
11-
* $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.150 2009/02/27 00:06:27 tgl Exp $
11+
* $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.151 2009/03/26 17:15:35 tgl Exp $
1212
*
1313
*-------------------------------------------------------------------------
1414
*/
@@ -1480,9 +1480,20 @@ create_hashjoin_path(PlannerInfo *root,
14801480
pathnode->jpath.outerjoinpath=outer_path;
14811481
pathnode->jpath.innerjoinpath=inner_path;
14821482
pathnode->jpath.joinrestrictinfo=restrict_clauses;
1483-
/* A hashjoin never has pathkeys, since its ordering is unpredictable */
1483+
/*
1484+
* A hashjoin never has pathkeys, since its output ordering is
1485+
* unpredictable due to possible batching. XXX If the inner relation is
1486+
* small enough, we could instruct the executor that it must not batch,
1487+
* and then we could assume that the output inherits the outer relation's
1488+
* ordering, which might save a sort step. However there is considerable
1489+
* downside if our estimate of the inner relation size is badly off.
1490+
* For the moment we don't risk it. (Note also that if we wanted to take
1491+
* this seriously, joinpath.c would have to consider many more paths for
1492+
* the outer rel than it does now.)
1493+
*/
14841494
pathnode->jpath.path.pathkeys=NIL;
14851495
pathnode->path_hashclauses=hashclauses;
1496+
/* cost_hashjoin will fill in pathnode->num_batches */
14861497

14871498
cost_hashjoin(pathnode,root,sjinfo);
14881499

‎src/include/nodes/relation.h

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -7,7 +7,7 @@
77
* Portions Copyright (c) 1996-2009, PostgreSQL Global Development Group
88
* Portions Copyright (c) 1994, Regents of the University of California
99
*
10-
* $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.170 2009/03/05 23:06:45 tgl Exp $
10+
* $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.171 2009/03/26 17:15:35 tgl Exp $
1111
*
1212
*-------------------------------------------------------------------------
1313
*/
@@ -845,6 +845,7 @@ typedef struct HashPath
845845
{
846846
JoinPathjpath;
847847
List*path_hashclauses;/* join clauses used for hashing */
848+
intnum_batches;/* number of batches expected */
848849
}HashPath;
849850

850851
/*

0 commit comments

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