Re: Strange behaviour of SELECT ... IN - Mailing list pgsql-general

From Stephan Szabo
Subject Re: Strange behaviour of SELECT ... IN
Date
Msg-id 20020626103955.P93117-100000@megazone23.bigpanda.com
Whole thread Raw
In response to Strange behaviour of SELECT ... IN  (Jorge Sarmiento <jsarmiento@ccom.org>)
Responses Re: Strange behaviour of SELECT ... IN  (Jorge Sarmiento <jsarmiento@ccom.org>)
List pgsql-general
On Wed, 26 Jun 2002, Jorge Sarmiento wrote:

> when I do a:
>
> explain select * from table1 where name in ('JORGE');
>
> NOTICE:  QUERY PLAN:
>
> Index Scan using idx_table1 on table  (cost=0.00..10901.87 rows=3184 width=48)
>
> but if do a:
>
> explain select * from table1 where name in ('JORGE',  'JUAN',  'JOSE');
>
> NOTICE:  QUERY PLAN:
>
> Seq Scan on table1  (cost=0.00..16689.73 rows=9506 width=48)
>
>
> why??? is there any way to make postgresql use index in the second type of
> query?

Well, the number of rows in the second is about 3 times greater.  Have
you done a vacuum analyze of the table in question?  How many rows does
the table actually have?




pgsql-general by date:

Previous
From: Richard Huxton
Date:
Subject: Re: unlimited undo/journaling
Next
From: Bruce Momjian
Date:
Subject: Re: Shared Memory Sizing