Horton

Re: Index-only scan returns incorrect results when using a composite GIST index with a gist_trgm_ops column.

Michael Paquier <michael.paquier@gmail.com>

View thread View on postgr.es
From: Michael Paquier <michael.paquier@gmail.com>
To: Andrey Borodin <x4mmm@yandex-team.ru>
Cc: Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>, sk@zsrv.org, david@nlpgo.com, pgsql-bugs@lists.postgresql.org, pgsql-hackers@postgresql.org
Date: 2018-01-18T08:25:05Z
Lists: pgsql-bugs, pgsql-hackers 
On Thu, Jan 18, 2018 at 12:57:38PM +0500, Andrey Borodin wrote:
>> 18 янв. 2018 г., в 10:48, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp> написал(а):
>> 
>> Gist imposes the ninth strategy to perform index only scan but
>> planner is not considering that
>> ....
>> Please find the attached patch.
> I agree with you that current behavior is a bug and your patch seems correct.
> I'm a bit worried about ninth strategy words: fetch is not necessary
>> now, if opclass lacks compress methods - index only scan is
>> possible. See
>> https://github.com/postgres/postgres/commit/d3a4f89d8a3e500bd7c0b7a8a8a5ce1b47859128
>> for details.
>
> Though there are tests in cube and seg for that, if your patch passes
> check-world, than this behavior is not affected. 

The proposed patch has no regression tests. If the current set is not
enough to stress the problem, you surely should add some (haven't
checked the patch in detail, sorry ;p ).
--
Michael

Commits

  1. Improve stability of recently-added regression test case.

  2. Fix IOS planning when only some index columns can return an attribute.

  3. Ability to advance replication slots

  4. Allow no-op GiST support functions to be omitted.