From: | "Dave Page" <dpage(at)postgresql(dot)org> |
---|---|
To: | "Simon Riggs" <simon(at)2ndquadrant(dot)com> |
Cc: | pgadmin-hackers <pgadmin-hackers(at)postgresql(dot)org> |
Subject: | Re: ALTER INDEX set fillfactor |
Date: | 2007-10-06 16:45:52 |
Message-ID: | 200710061745520000@074520261 |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgadmin-hackers |
> ------- Original Message -------
> From: Simon Riggs <simon(at)2ndquadrant(dot)com>
> To: Dave Page <dpage(at)postgresql(dot)org>
> Sent: 06/10/07, 08:33:30
> Subject: Re: [pgadmin-hackers] ALTER INDEX set fillfactor
>
> On Fri, 2007-10-05 at 21:24 +0100, Dave Page wrote:
> > Simon Riggs wrote:
>
> > > It would be even better if there was an option to do that CONCURRENTLY,
> > > i.e. add the new index and then drop the old one afterwards. (It's
> > > unfortunate that there isn't a REINDEX concurrently, but there isn't
> > > yet).
> >
> > Funny - I suggested that to Greg just the other day...
>
> Sorry, I meant in the absence of REINDEX concurrently, we can issue:
>
> CREATE INDEX CONCURRENTLY newindex
> DROP INDEX oldindex
>
> on each index on the table in turn, which does the same thing.
CIC can't be run in a transaction block, so how do we queue up the drop & rename to run after it completes? Given the changes to VACUUM & CREATE DATABASE/TABLESPACE in 8.3 I would imagine it won't run in a multi-statement query either now.
/D
From | Date | Subject | |
---|---|---|---|
Next Message | Simon Riggs | 2007-10-06 17:53:15 | Re: ALTER INDEX set fillfactor |
Previous Message | Simon Riggs | 2007-10-06 07:33:30 | Re: ALTER INDEX set fillfactor |