Re: Adding a stored generated column without long-lived locks

From: Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at>
To: Matthias van de Meent <boekewurm+postgres(at)gmail(dot)com>
Cc: Alberto Piai <alberto(dot)piai(at)gmail(dot)com>, Álvaro Herrera <alvherre(at)kurilemu(dot)de>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: Adding a stored generated column without long-lived locks
Date: 2026-10-03 01:25:49
Message-ID: 27824d022b9ddf7edcaa11f450a6cb29b4f3fe7a.camel@cybertec.at
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, 2026-10-02 at 18:58 +0200, Matthias van de Meent wrote:
> On Fri, 2 Oct 2026 at 09:30, Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at> wrote:
> >
> > You could simply require the constraint to be
> >
> >   ROW(col)::record *= ROW(expression)::record
> >
> > That would also behave as desired for NULL values.
> >
> > True, it looks hacky, but as you said, it is a constraint created
> > for this special purpose.
>
> I don't think that's a good solution:  Whilst it may work, it also
> adds significant and potentially unnecessary overhead to the execution
> of the CHECK constraint.  The added complexity in the expression
> itself also increases the chance of a user defining the constraint
> incorrectly.

All of that is correct, but the constraint is not permanent, so you don't
have to pay the slight overhead for the more complicated expression
permanently. If the user defines the constraint wrongly, nothing bad
can happen: ALTER TABLE will complain that there is no appropriate
constraint. Probably annoying, but no damage done.

Yours,
Laurenz Albe

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Manu 2026-10-03 01:48:37 Re: Fix reindexdb with parallel index-level conrurrent run
Previous Message Michael Paquier 2026-10-02 23:48:00 Re: injection_points: canceled or terminated waiters leak their wait slots