From: | Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com> |
---|---|
To: | "Fujii(dot)Yuki(at)df(dot)MitsubishiElectric(dot)co(dot)jp" <Fujii(dot)Yuki(at)df(dot)MitsubishiElectric(dot)co(dot)jp>, "pgsql-hackers(at)lists(dot)postgresql(dot)org" <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Cc: | "david(at)pgmasters(dot)net" <david(at)pgmasters(dot)net>, "ah(at)cybertec(dot)at" <ah(at)cybertec(dot)at>, "tgl(at)sss(dot)pgh(dot)pa(dot)us" <tgl(at)sss(dot)pgh(dot)pa(dot)us>, "zhihui(dot)fan1213(at)gmail(dot)com" <zhihui(dot)fan1213(at)gmail(dot)com>, "legrand_legrand(at)hotmail(dot)com" <legrand_legrand(at)hotmail(dot)com>, "daniel(at)yesql(dot)se" <daniel(at)yesql(dot)se> |
Subject: | Re: WIP: Aggregation push-down - take2 |
Date: | 2022-10-31 19:00:20 |
Message-ID: | a2f704ab-aa50-b6af-5e92-0fece5ad07e5@enterprisedb.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi,
On 7/12/22 08:49, Fujii(dot)Yuki(at)df(dot)MitsubishiElectric(dot)co(dot)jp wrote:
> Hi everyone.
>
> I rebased the following patches which were submitted in [1].
> v17-0001-Introduce-RelInfoList-structure.patch
> v17-0002-Aggregate-push-down-basic-functionality.patch
> v17-0003-Use-also-partial-paths-as-the-input-for-grouped-path.patch
>
> I checked I can apply the rebased patch to commit 2cd2569c72b8920048e35c31c9be30a6170e1410.
>
> I'm going to register the rebased patch in next commitfest.
>
I've started looking at this patch series again, but I wonder what's the
plan. The last patch version no longer applies, so I rebased it - see
the attachment. The failures were pretty minor, but there're two warnings:
pathnode.c:3174:11: warning: variable 'agg_exprs' set but not used
[-Wunused-but-set-variable]
Node *agg_exprs;
^
pathnode.c:3252:11: warning: variable 'agg_exprs' set but not used
[-Wunused-but-set-variable]
Node *agg_exprs;
^
so there seem to be some loose ends. Moreover, there are two failures in
make check, due to plan changes like this:
+ Finalize GroupAggregate
Group Key: p.i
- -> Nested Loop
- -> Partial HashAggregate
- Group Key: c1.parent
- -> Seq Scan on agg_pushdown_child1 c1
- -> Index Scan using agg_pushdown_parent_pkey on ...
- Index Cond: (i = c1.parent)
-(8 rows)
+ -> Sort
+ Sort Key: p.i
+ -> Nested Loop
+ -> Partial HashAggregate
+ Group Key: c1.parent
+ -> Seq Scan on agg_pushdown_child1 c1
+ -> Index Scan using agg_pushdown_parent_pkey on ...
+ Index Cond: (i = c1.parent)
+(10 rows)
This seems somewhat strange - maybe the plan is correct, but the extra
sort seems unnecessary.
However, maybe I'm confused/missing something? The above message says
v17 having parts 0001-0003, but there's only one patch in v18. So maybe
I failed to apply some prior patch?
regards
--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
Attachment | Content-Type | Size |
---|---|---|
v19-Push-aggregation-down-to-base-relations-and-joins.patch | text/x-patch | 130.3 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Matthias van de Meent | 2022-10-31 19:42:17 | Re: [PATCHES] Post-special page storage TDE support |
Previous Message | Robert Haas | 2022-10-31 18:40:28 | Re: heavily contended lwlocks with long wait queues scale badly |