Re: Inconsistent Behavior of GROUP BY ROLLUP in v17 vs master - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Inconsistent Behavior of GROUP BY ROLLUP in v17 vs master
Date
Msg-id 3676522.1758770647@sss.pgh.pa.us
Whole thread Raw
In response to Re: Inconsistent Behavior of GROUP BY ROLLUP in v17 vs master  (David Rowley <dgrowleyml@gmail.com>)
Responses Re: Inconsistent Behavior of GROUP BY ROLLUP in v17 vs master
List pgsql-hackers
David Rowley <dgrowleyml@gmail.com> writes:
> On Thu, 25 Sept 2025 at 13:01, Richard Guo <guofenglinux@gmail.com> wrote:
>> I plan to push this patch soon, unless there are any objections.

> What's your confidence levels on the logic now being correct? 100%?
> 90%? Hopeful?

FWIW, my confidence in it is rather low.  I've not had time to think
this through carefully, but it seems to me that the test ought to
involve whether there is an empty grouping set, yet the proposed
patch does no such thing --- or at least, if it manages to achieve
that effect, it's not obvious how.

18.1 will not be coming out till November, so I feel no need to
rush to judgment on what to do here.

            regards, tom lane



pgsql-hackers by date:

Previous
From: Fujii Masao
Date:
Subject: Re: Documentation fix on pgbench \aset command
Next
From: Chao Li
Date:
Subject: Re: Add support for specifying tables in pg_createsubscriber.