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

Commit76de4b1

Browse files
committed
Correctly identify which EC members are computable at a plan node.
find_computable_ec_member() had the wrong mental model of whatits primary caller prepare_sort_from_pathkeys() would do withthe selected EquivalenceClass member expression. We will notcompute the EC expression in a plan node atop the one returningthe passed-in targetlist; rather, the EC expression will becomputed as an additional column of that targetlist. So anyVar or quasi-Var used in the given tlist is also available to theEC expression. In simple cases this makes no difference becausethe given tlist is just a list of Vars or quasi-Vars --- but ifwe are considering an appendrel member produced by flatteninga UNION ALL, the tlist may contain expressions, resulting infailure to match and a "could not find pathkey item to sort"error.To fix, we can flatten both the tlist and the EC members withpull_var_clause(), and then just check for subset-ness, sothat the code is actually shorter than before.While this bug is quite old, the present patch only works back tov13. We could possibly make it work in v12 by back-patching partsof3753982. On the whole though I don't like the risk/rewardratio of that idea. v12's final release is next month, meaningthere would be no chance to correct matters if the patch causes aregression. Since this failure has escaped notice for 14 years,it's likely nobody will hit it in the field with v12.Per bug #18652 from Alexander Lakhin.Andrei Lepikhov and Tom LaneDiscussion:https://postgr.es/m/18652-deaa782ebcca85d1@postgresql.org
1 parent79ca063 commit76de4b1

File tree

3 files changed

+71
-38
lines changed

3 files changed

+71
-38
lines changed

‎src/backend/optimizer/path/equivclass.c

Lines changed: 33 additions & 38 deletions
Original file line numberDiff line numberDiff line change
@@ -39,7 +39,6 @@
3939
staticEquivalenceMember*add_eq_member(EquivalenceClass*ec,
4040
Expr*expr,Relidsrelids,Relidsnullable_relids,
4141
boolis_child,Oiddatatype);
42-
staticboolis_exprlist_member(Expr*node,List*exprs);
4342
staticvoidgenerate_base_implied_equalities_const(PlannerInfo*root,
4443
EquivalenceClass*ec);
4544
staticvoidgenerate_base_implied_equalities_no_const(PlannerInfo*root,
@@ -833,9 +832,18 @@ find_ec_member_matching_expr(EquivalenceClass *ec,
833832
*expressions appearing in "exprs"; return NULL if no match.
834833
*
835834
* "exprs" can be either a list of bare expression trees, or a list of
836-
* TargetEntry nodes. Either way, it should contain Vars and possibly
837-
* Aggrefs and WindowFuncs, which are matched to the corresponding elements
838-
* of the EquivalenceClass's expressions.
835+
* TargetEntry nodes. Typically it will contain Vars and possibly Aggrefs
836+
* and WindowFuncs; however, when considering an appendrel member the list
837+
* could contain arbitrary expressions. We consider an EC member to be
838+
* computable if all the Vars, PlaceHolderVars, Aggrefs, and WindowFuncs
839+
* it needs are present in "exprs".
840+
*
841+
* There is some subtlety in that definition: for example, if an EC member is
842+
* Var_A + 1 while what is in "exprs" is Var_A + 2, it's still computable.
843+
* This works because in the final plan tree, the EC member's expression will
844+
* be computed as part of the same plan node targetlist that is currently
845+
* represented by "exprs". So if we have Var_A available for the existing
846+
* tlist member, it must be OK to use it in the EC expression too.
839847
*
840848
* Unlike find_ec_member_matching_expr, there's no special provision here
841849
* for binary-compatible relabeling. This is intentional: if we have to
@@ -855,12 +863,24 @@ find_computable_ec_member(PlannerInfo *root,
855863
Relidsrelids,
856864
boolrequire_parallel_safe)
857865
{
866+
List*exprvars;
858867
ListCell*lc;
859868

869+
/*
870+
* Pull out the Vars and quasi-Vars present in "exprs". In the typical
871+
* non-appendrel case, this is just another representation of the same
872+
* list. However, it does remove the distinction between the case of a
873+
* list of plain expressions and a list of TargetEntrys.
874+
*/
875+
exprvars=pull_var_clause((Node*)exprs,
876+
PVC_INCLUDE_AGGREGATES |
877+
PVC_INCLUDE_WINDOWFUNCS |
878+
PVC_INCLUDE_PLACEHOLDERS);
879+
860880
foreach(lc,ec->ec_members)
861881
{
862882
EquivalenceMember*em= (EquivalenceMember*)lfirst(lc);
863-
List*exprvars;
883+
List*emvars;
864884
ListCell*lc2;
865885

866886
/*
@@ -878,18 +898,18 @@ find_computable_ec_member(PlannerInfo *root,
878898
continue;
879899

880900
/*
881-
* Match if all Vars and quasi-Vars areavailable in "exprs".
901+
* Match if all Vars and quasi-Vars arepresent in "exprs".
882902
*/
883-
exprvars=pull_var_clause((Node*)em->em_expr,
884-
PVC_INCLUDE_AGGREGATES |
885-
PVC_INCLUDE_WINDOWFUNCS |
886-
PVC_INCLUDE_PLACEHOLDERS);
887-
foreach(lc2,exprvars)
903+
emvars=pull_var_clause((Node*)em->em_expr,
904+
PVC_INCLUDE_AGGREGATES |
905+
PVC_INCLUDE_WINDOWFUNCS |
906+
PVC_INCLUDE_PLACEHOLDERS);
907+
foreach(lc2,emvars)
888908
{
889-
if (!is_exprlist_member(lfirst(lc2),exprs))
909+
if (!list_member(exprvars,lfirst(lc2)))
890910
break;
891911
}
892-
list_free(exprvars);
912+
list_free(emvars);
893913
if (lc2)
894914
continue;/* we hit a non-available Var */
895915

@@ -907,31 +927,6 @@ find_computable_ec_member(PlannerInfo *root,
907927
returnNULL;
908928
}
909929

910-
/*
911-
* is_exprlist_member
912-
* Subroutine for find_computable_ec_member: is "node" in "exprs"?
913-
*
914-
* Per the requirements of that function, "exprs" might or might not have
915-
* TargetEntry superstructure.
916-
*/
917-
staticbool
918-
is_exprlist_member(Expr*node,List*exprs)
919-
{
920-
ListCell*lc;
921-
922-
foreach(lc,exprs)
923-
{
924-
Expr*expr= (Expr*)lfirst(lc);
925-
926-
if (expr&&IsA(expr,TargetEntry))
927-
expr= ((TargetEntry*)expr)->expr;
928-
929-
if (equal(node,expr))
930-
return true;
931-
}
932-
return false;
933-
}
934-
935930
/*
936931
* Find an equivalence class member expression, all of whose Vars, come from
937932
* the indicated relation.

‎src/test/regress/expected/inherit.out

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1543,6 +1543,36 @@ select min(1-id) from matest0;
15431543

15441544
reset enable_seqscan;
15451545
reset enable_parallel_append;
1546+
explain (verbose, costs off) -- bug #18652
1547+
select 1 - id as c from
1548+
(select id from matest3 t1 union all select id * 2 from matest3 t2) ss
1549+
order by c;
1550+
QUERY PLAN
1551+
------------------------------------------------------------
1552+
Result
1553+
Output: ((1 - t1.id))
1554+
-> Merge Append
1555+
Sort Key: ((1 - t1.id))
1556+
-> Index Scan using matest3i on public.matest3 t1
1557+
Output: t1.id, (1 - t1.id)
1558+
-> Sort
1559+
Output: ((t2.id * 2)), ((1 - (t2.id * 2)))
1560+
Sort Key: ((1 - (t2.id * 2)))
1561+
-> Seq Scan on public.matest3 t2
1562+
Output: (t2.id * 2), (1 - (t2.id * 2))
1563+
(11 rows)
1564+
1565+
select 1 - id as c from
1566+
(select id from matest3 t1 union all select id * 2 from matest3 t2) ss
1567+
order by c;
1568+
c
1569+
-----
1570+
-11
1571+
-9
1572+
-5
1573+
-4
1574+
(4 rows)
1575+
15461576
drop table matest0 cascade;
15471577
NOTICE: drop cascades to 3 other objects
15481578
DETAIL: drop cascades to table matest1

‎src/test/regress/sql/inherit.sql

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -543,6 +543,14 @@ select min(1-id) from matest0;
543543
reset enable_seqscan;
544544
reset enable_parallel_append;
545545

546+
explain (verbose, costs off)-- bug #18652
547+
select1- idas cfrom
548+
(select idfrom matest3 t1union allselect id*2from matest3 t2) ss
549+
order by c;
550+
select1- idas cfrom
551+
(select idfrom matest3 t1union allselect id*2from matest3 t2) ss
552+
order by c;
553+
546554
droptable matest0 cascade;
547555

548556
--

0 commit comments

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