From: | Etsuro Fujita <fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp> |
---|---|
To: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Unneeded parallel safety tests in grouping_planner |
Date: | 2019-02-27 12:45:29 |
Message-ID: | 5C768669.8070207@lab.ntt.co.jp |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi,
Yet another thing I noticed while working on [1] is this in
grouping_planner:
/*
* If the input rel is marked consider_parallel and there's nothing
that's
* not parallel-safe in the LIMIT clause, then the final_rel can be
marked
* consider_parallel as well. Note that if the query has rowMarks or is
* not a SELECT, consider_parallel will be false for every relation
in the
* query.
*/
if (current_rel->consider_parallel &&
is_parallel_safe(root, parse->limitOffset) &&
is_parallel_safe(root, parse->limitCount))
final_rel->consider_parallel = true;
If there is a need to add a LIMIT node, we don't consider generating
partial paths for the final relation below (see commit
0927d2f46ddd4cf7d6bf2cc84b3be923e0aedc52), so it seems unnecessary
anymore to assess the parallel-safety of the LIMIT and OFFSET clauses.
To save cycles, why not remove those tests from that function like the
attached?
Best regards,
Etsuro Fujita
Attachment | Content-Type | Size |
---|---|---|
unneeded-parallel-safety-tests.patch | text/x-patch | 2.1 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Prabhat Sahu | 2019-02-27 13:38:53 | pg_dump/pg_restore fail for TAR_DUMP and CUSTOM_DUMP from v94/v95/v96 to v11/master. |
Previous Message | Alvaro Herrera | 2019-02-27 12:32:17 | Re: Segfault when restoring -Fd dump on current HEAD |