Re: Gist consistent and compression

From: Dimitri Fontaine <dfontaine(at)hi-media(dot)com>
To: Yeb Havinga <y(dot)t(dot)havinga(at)mgrid(dot)net>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Gist consistent and compression
Date: 2009-05-02 19:02:59
Message-ID: E56421B9-93A5-493D-8F8F-BCFED042452C@hi-media.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

Le 1 mai 09 à 23:04, Yeb Havinga a écrit :
> An idea for possible improvement of query speed on gist indexes with
> a costly compression function. We have a gist index that uses a
> compressed internal datatype. The compression is with some cost (it
> involves a syscache lookup).

[...]

> small one. So it would make sense to add the caching to the gist
> support in the backend, instead of having to build one in the user
> defined functions gist functions for every datatype.
>
> Any thoughts?

It seems to me that what you're asking for is addressed indirectly in
the possibility to make your internal data type a full SQL visible
datatype. Then you store this new datatype directly in the table and
index that. Instead of converting from external to internal type at
consistent() time in a query, provide an implicit CAST for external to
internal for queries to "just work" without editing. The CAST will get
called once per literal in the query.

See prefix and the prefix_range datatype as an example of this:
http://blog.tapoueh.org/prefix.html

I'm not sure if the caching facility should be added to -core
PostgreSQL or not, I'm just proposing another way to solve your
problem without requiring it...

Regards,
--
dim

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2009-05-02 20:38:48 Re: windows doesn't notice backend death
Previous Message Andrew Dunstan 2009-05-02 18:56:56 windows doesn't notice backend death