From: Richard Neill <rn214@richardneill.org>
Cc: pgsql-performance@postgresql.org <pgsql-performance@postgresql.org>
Subject: Re: Why does the query planner use two full indexes, when a dedicated partial index exists?
Date: Thu, 20 Dec 2012 05:57:14 +0000
Message-ID: <50D2A8BA.2070801@richardneill.org> (raw)
In-Reply-To: <CAMkU=1wYVjf63j+4ttZrVMmYfst7Q7ngez5e6X4s+keYbpcwdg@mail.gmail.com>
References: <50D22DE2.4090301@richardneill.org>
<CAL_0b1s6_YZDdzt3NizwC8X+dYA37T-4aZOTzw3KJ3-iXuSjjw@mail.gmail.com>
<50D2528E.7000800@richardneill.org>
<CAL_0b1t5_h-VUE+sL2cfd1E78xYOq9T7P3n31iZhLT3vEa=mOA@mail.gmail.com>
<50D25A59.3060105@richardneill.org>
<CAMkU=1wYVjf63j+4ttZrVMmYfst7Q7ngez5e6X4s+keYbpcwdg@mail.gmail.com>
List-Unsubscribe: <mailto:majordomo@postgresql.org?body=unsub%20pgsql-performance>
Dear Jeff,
Thanks for your help,
> * The reindex solution doesn't work. I just tried it, and
> the query planner
> is still using the wrong indexes.
>
>
> It switched to a better one of the wrong indices, though, and got
> several times faster.
>
I think that this is a red herring. The switching between the two
"wrong" indices seems to be caused by non-uniformity in the
parcel_id_code: although it's distributed fairly well across 1-99999,
it's not perfect.
As for the speed-up, I think that's mostly caused by the fact that
running "Analyse" is pulling the entire table (and the relevant index)
into RAM and flushing other things out of that cache.
> How did it get so bloated in the first place? Is the table being
> updated so rapidly that the statistics might be wrong even immediately
> after analyze finishes?
I don't think it is. We're doing about 10 inserts and 20 updates per
second on that table. But when I tested it, production had stopped for
the night - so the system was quiescent between the analyse and the select.
> In any case, I can't get it to prefer the full index in 9.1.6 at all.
> The partial index wins hands down unless the table is physically
> clustered by the parcel_id_code column. In which that case, the partial
> index wins by only a little bit.
Interesting that you should say that... the original setup script did
choose to cluster the table on that column.
Also, I wonder whether it matters which order the indexes are created in?
Best wishes,
Richard
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: rn214@richardneill.org
Subject: Re: Why does the query planner use two full indexes, when a dedicated partial index exists?
In-Reply-To: <50D2A8BA.2070801@richardneill.org>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox