Re: When can postgresql use a partial (NOT NULL) index? Seems to depend on size of IN clause (even with enable seqscan = off)

From: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>
To: Timothy Garnett <tgarnett(at)panjiva(dot)com>
Cc: pgsql-general(at)postgresql(dot)org
Subject: Re: When can postgresql use a partial (NOT NULL) index? Seems to depend on size of IN clause (even with enable seqscan = off)
Date: 2010-08-03 21:01:36
Message-ID: AANLkTi=wmUTgZY5fvGjuiDpG1L==pJkc_BQW_KrdL1s9@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Tue, Aug 3, 2010 at 2:03 PM, Timothy Garnett <tgarnett(at)panjiva(dot)com> wrote:
> Hi all,
>
> I'm debugging a performance issue that looks like it might actually be an
> issue/limitation/parameter/bug in the query planner, but since I couldn't
> find anything authoritative on when exactly postgresql is able to use
> partial not null indexes I'm not sure that that's the case and I was hopping
> someone could give some clarity around that or point to an option I could
> tweak that would change this behavior.  Anyways the table in question (with
> names changed) is below.  I'm running postgres 8.4.1

8.4.1 has some pretty nasty bugs that have since been fixed, and some
work was done on the planner as well since then. First upgrade and
see if your problem goes away.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Joshua D. Drake 2010-08-03 21:09:16 Re: problem with pg_standby
Previous Message Tom Lane 2010-08-03 21:01:15 Re: When can postgresql use a partial (NOT NULL) index? Seems to depend on size of IN clause (even with enable seqscan = off)