Re: Indexes not working (bug in 7.0.2?) - Mailing list pgsql-general

From Zlatko Calusic
Subject Re: Indexes not working (bug in 7.0.2?)
Date
Msg-id 87snrfx43q.fsf@atlas.iskon.hr
Whole thread Raw
In response to Indexes not working (bug in 7.0.2?)  (Zlatko Calusic <zlatko@iskon.hr>)
Responses Driver  ("Enrico Comini" <algobit@algobit.com>)
Re: Indexes not working (bug in 7.0.2?)  (Zeljko Trogrlic <zeljko@technologist.com>)
List pgsql-general
Tom Lane <tgl@sss.pgh.pa.us> writes:

> Zlatko Calusic <zlatko@iskon.hr> writes:
> > It is now NOT using the index, and I don't understand why? Queries are
> > practically the same, tables are practically the same, why is postgres
> > using indexes in the first case and not in the second?
>
> Because it has substantially different ideas about the sizes of the
> two tables --- notice the different estimated row counts.  If you
> haven't "vacuum analyzed" these tables recently, do so to bring the
> planner's statistics up-to-date, and then see what you get.  You may
> also care to read the user's manual chapter about EXPLAIN,
> http://www.postgresql.org/users-lounge/docs/7.0/postgres/c4888.htm
>

Yes, thanks to all who helped. 'vacuum analyze' was the magical
incantation that helped.

I still have one uncertainty. Is it possible that after some time
postgres once again decides not to use indices (assuming I haven't run
'vacuum analyze' again)?

--
Zlatko

pgsql-general by date:

Previous
From: Michaël Fiey
Date:
Subject: NULL value withe cursor
Next
From: "Enrico Comini"
Date:
Subject: Driver