Horton

Re: 'update returning *' returns 0 columns instead of empty row with 2 columns when (i) no rows updated and (ii) when applied to a partitioned table with sub-partition

Amit Langote <langote_amit_f8@lab.ntt.co.jp>

View thread View on postgr.es
From: Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Petr Fedorov <petr.fedorov@phystech.edu>, pgsql-bugs@postgresql.org
Date: 2019-02-22T05:33:55Z
Lists: pgsql-bugs

Attachments

On 2019/02/22 13:43, Tom Lane wrote:
> Amit Langote <Langote_Amit_f8@lab.ntt.co.jp> writes:
>> Also, I noticed regression test failure having to do with statement
>> triggers not firing, which makes sense, as there's no ModifyTable to
>> invoke them.
> 
> Oh!  You are right, that's a separate bug.  So really we can't have
> this fast-path exit at all, we should produce a ModifyTable node in
> every case.
> 
> I'm too tired to work on that anymore today, do you want to run
> with it?

Sure, see attached a patch.

To fix the trigger bug, we'll be putting a minimally valid-looking
ModifyTable node on top of a dummy plan.  So, the bug reported on this
thread is taken care of automatically, because ModifyTable plan's
targetlist is already set correctly.

Thanks,
Amit

Commits

  1. Fix plan created for inherited UPDATE/DELETE with all tables excluded.