From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com> |
Cc: | James Coleman <jtc331(at)gmail(dot)com>, Luc Vlaming <luc(at)swarm64(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Robert Haas <robertmhaas(at)gmail(dot)com> |
Subject: | Re: "could not find pathkey item to sort" for TPC-DS queries 94-96 |
Date: | 2021-04-18 17:21:55 |
Message-ID: | 90179.1618766515@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I wrote:
> I think it's time for some refactoring of this code so that we can
> actually share the logic. Accordingly, I propose the attached.
After sleeping on it, here's an improved version that gets rid of
an unnecessary assumption about ECs usually not containing both
parallel-safe and parallel-unsafe members. I'd tried to do this
yesterday but didn't like the amount of side-effects on createplan.c
(caused by the direct call sites not being passed the "root" pointer).
However, we can avoid refactoring createplan.c APIs by saying that it's
okay to pass root = NULL to find_computable_ec_member if you're not
asking it to check parallel safety. And there's not really a need to
put a parallel-safety check into find_ec_member_matching_expr at all;
that task can be left with the one caller that cares.
regards, tom lane
Attachment | Content-Type | Size |
---|---|---|
0001-Fix-find_em_expr_usable_for_sorting_rel-3.patch | text/x-diff | 20.5 KB |
0002-change-function-name-3.patch | text/x-diff | 4.9 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2021-04-18 17:46:46 | Re: 回复: Core dump happens when execute sql CREATE VIEW v1(c1) AS (SELECT ('4' COLLATE "C")::INT FROM generate_series(1, 10)); |
Previous Message | Yulin PEI | 2021-04-18 16:31:07 | 回复: Core dump happens when execute sql CREATE VIEW v1(c1) AS (SELECT ('4' COLLATE "C")::INT FROM generate_series(1, 10)); |