Re: missing indexes in indexlist with partitioned tables

Arne Roland <a.roland@index.de>

From: Arne Roland <A.Roland@index.de>
To: Amit Langote <amitlangote09@gmail.com>, David Rowley <dgrowleyml@gmail.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>, Zhihong Yu <zyu@yugabyte.com>, Alvaro Herrera <alvherre@alvh.no-ip.org>, Julien Rouhaud <rjuju123@gmail.com>
Date: 2022-10-01T16:34:16Z
Lists: pgsql-hackers

Attachments

On Tue, Sep 20, 2022 at 4:53 PM David Rowley <dgrowleyml@gmail.com> wrote:
> Arne sent me an off-list
> message to say he's planning on working on the patch that uses the
> existing field instead of the new one he originally added. Let's hold
> off for that patch.

I wouldn't say, I explicitly stated that. But I ended up doing something, that resulted in the attached patch. :)


For my own sanity I greped one last time for the usage of indexlist.

Most of the (untouched) usages have comments that, they are only called for baserels/plain tables.

Namely all but the cluster of partitioned tables. I had to reread that section. There we are just traversing the tree and omitting partitioned tables.


There is now a test section in join.sql for partitioned tables, that tests very similar to the

baserel case. That's more thorough, than what I originally went for.

Further feedback would be appreciated!


Regards
Arne

Commits

  1. Allow left join removals and unique joins on partitioned tables

  2. Add index on pg_publication_rel.prpubid