Re: Performance issue in foreign-key-aware join estimation

From: David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
Subject: Re: Performance issue in foreign-key-aware join estimation
Date: 2019-03-09 00:18:57
Message-ID: CAKJS1f-wWz6mkn1ccYD5d4YY1gUSaak6vS=W_5GOQ3zU_LSkzQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sat, 9 Mar 2019 at 08:05, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Setting the CF entry to WOA for now. I wonder though if we should
> just push it out to v13 immediately --- are you intending to do more
> with it in the near future?

Thanks a lot for going over this and providing feedback. I put the
patch out there mostly to see if such a thing is something we might
want, and it's good to see no objections to the idea. I didn't want to
waste too much time if there was going to be some serious objections
to the idea.

This is something I'd like to look into for v13. I think it'll be
much easier to do if we can get your List reimplementation patch in
first. That's going to allow list_nth on PlannerInfo->eq_classes to
work quickly and will remove the need for having to build an array to
store the classes, and as you mention, RelOptInfo could store a
Bitmapset to store which ECs have members for this rel. I've
modified the CF entry to target v13.

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Corey Huinker 2019-03-09 00:19:45 Re: \describe*
Previous Message Michael Paquier 2019-03-09 00:09:24 Re: Fwd: Add tablespace tap test to pg_rewind