Re: MERGE loses UPDATE transition rows when moving rows between partitions

From: Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com>
To: Tender Wang <tndrwang(at)gmail(dot)com>
Cc: shihao zhong <zhong950419(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Michael Ries <mries(at)salesforce(dot)com>
Subject: Re: MERGE loses UPDATE transition rows when moving rows between partitions
Date: 2026-11-18 16:23:40
Message-ID: CAEZATCVWEW3K7HrVZq_Qp9nDS5cCj2OhuOA=-HqJxHu=HXdHCQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Tue, 10 Nov 2026 at 04:45, Tender Wang <tndrwang(at)gmail(dot)com> wrote:
>
> Thanks for rebasing the patch, preparing the backport versions, and
> identifying the commit responsible.
>
> Since this is a regression from a4f7d33a904, I've CC'd Dean, along
> with Michael Ries, who also reported the issue.

Thanks. That looked good to me, so I've pushed and back-patched it to
v15, with minor cosmetic adjustments.

I renamed the variables ar_insert_trig_tcs and ar_delete_trig_tcs
because the TransitionCaptureState structure is not command specific,
but applies to all command types, so code like
"ar_insert_trig_tcs->tcs_update_new_table" looked odd.

Regards,
Dean

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Alberto Piai 2026-11-18 16:40:47 Re: Adding a stored generated column without long-lived locks
Previous Message Manu 2026-11-18 16:16:21 Re: Assertion failure with temporary relations in single-user mode