Horton

Re: Fix BUG #17335: Duplicate result rows in Gather node

David Rowley <dgrowleyml@gmail.com>

View thread View on postgr.es
From: David Rowley <dgrowleyml@gmail.com>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Yura Sokolov <y.sokolov@postgrespro.ru>, PostgreSQL Developers <pgsql-hackers@lists.postgresql.org>
Date: 2022-02-08T21:10:51Z
Lists: pgsql-hackers

Attachments

Thanks for having a look at this.

On Fri, 4 Feb 2022 at 13:48, Robert Haas <robertmhaas@gmail.com> wrote:
> I think the actual rule is: every path under a Gather or GatherMerge
> must be parallel-safe.

I've adjusted the patch so that it counts parallel_aware and
parallel_safe Paths independently and verifies everything below a
Gather[Merge] is parallel_safe.

The diff stat currently looks like:

src/backend/optimizer/plan/createplan.c | 230
1 file changed, 230 insertions(+)

I still feel this is quite a bit of code for what we're getting here.
I'd be more for it if the path traversal function existed for some
other reason and I was just adding the callback functions and Asserts.

I'm keen to hear what others think about that.

David

Commits

  1. Consider parallel awareness when removing single-child Appends

  2. Suppress Append and MergeAppend plan nodes that have a single child.