From: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
---|---|
To: | Bruce Momjian <bruce(at)momjian(dot)us> |
Cc: | Rod Taylor <pg(at)rbt(dot)ca>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-hackers(at)postgresql(dot)org, Greg Stark <gsstark(at)mit(dot)edu> |
Subject: | Re: Better name/syntax for "online" index creation |
Date: | 2006-07-25 03:41:10 |
Message-ID: | 20060725034110.GD11023@surnet.cl |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Bruce Momjian wrote:
> > Anyway, if you want psql to easily identify it, just return NOLOCK as
> > part of the command string returned:
> >
> > test=> create index i on test(x);
> > CREATE INDEX NOLOCK
>
> Oh, psql needs to know before the command is sent? How do we handle it
> now with CLUSTER?
We don't, which is exactly the problem. If I'm not mistaken, currently
psql in autocommit off mode, CLUSTER doesn't start a transaction block,
which is arguably wrong because some forms of CLUSTER (single-table) are
able to work within a transaction. But since not all of them are, then
we must act like they all were, because otherwise we would send spurious
error messages to the user.
> Whatever psql is trying to prevent doesn't seem to
> warrant mucking up the logical order of the CREATE INDEX command.
Personally I'm not sure if this is too serious an issue.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
From | Date | Subject | |
---|---|---|---|
Next Message | Alvaro Herrera | 2006-07-25 03:46:58 | Re: pgstattuple extension for indexes |
Previous Message | Bruce Momjian | 2006-07-25 02:12:00 | Re: Making config file parser available to add-ins |