From: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>, Greg Stark <gsstark(at)mit(dot)edu>, Michael Paesold <mpaesold(at)gmx(dot)at>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: code question: storing INTO relation |
Date: | 2004-11-14 11:06:38 |
Message-ID: | 1100430398.2950.129.camel@localhost.localdomain |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Fri, 2004-11-12 at 23:13, Tom Lane wrote:
> Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > Greg Stark wrote:
> >> I think that's already done for CREATE INDEX/REINDEX.
>
> > I don't think so. Can someone confirm?
>
> Greg is correct --- at least for btree build, which is the only index
> type we have WAL-ified at all :-(
[well...at least they're optimized then... :) ]
With regard to the other index types, my opinion was:
HASH - works OK, but a pain to administer, no huge benefit in using
R-TREE - slightly broken in places, limited in usablity
GiST - index of choice for PostGIS, TSearch2, in need of optimization
Following recent optimization work on GiST, it now seems worth the
trouble to add WAL logging to it. ISTM that the other two aren't widely
used enough to make it worthwhile to spend time on, evidence for which
is also that no one ever has, up 'til now.
Time-management seems to be the key to making progress in the most
important areas...
--
Best Regards, Simon Riggs
From | Date | Subject | |
---|---|---|---|
Next Message | Simon Riggs | 2004-11-14 11:24:34 | Re: postmaster segfaults with HUGE table |
Previous Message | Simon Riggs | 2004-11-14 10:52:08 | Re: Increasing the length of |