From: "Anibal David Acosta" <aa@devshock.com>
To: <pgsql-performance@postgresql.org>
Subject: postgres not use index, IN statement
Date: Mon, 23 May 2011 17:30:34 -0400
Message-ID: <006801cc1990$a8b34e10$fa19ea30$@devshock.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
Thread-Index: AcwZj/ZbdY4Qnu25TSiA99EiLYow9A==
Content-Language: es-py

I have a strange situation.
I have a table of detail with millones of rows and a table of items with
thousands of rows

When I do..

select count(*) from wiz_application_response where application_item_id in
(select id from wiz_application_item where application_id=110)

This query NOT use the index on column application_item_id, instead is doing
a sequential scan

BUT, when I add the range of min and max id of the subquery, the postgres
uses the INDEX
This is the second query...

select count(*) from wiz_application_response where application_item_id
between 908 and 1030 and application_item_id in(select id from
wiz_application_item where application_id=110)

908 and 1030 are limits (lower and upper) of the subquery, the subquery
returns 100 elements aprox.

So, this is some bug?

Thanks!

Anibal