Re: subquery and table join, index not use for table

From: Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com>
To: CoL <col(at)mportal(dot)hu>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: subquery and table join, index not use for table
Date: 2004-01-16 17:48:18
Message-ID: 20040116094239.A66040@megazone.bigpanda.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance


On Wed, 14 Jan 2004, CoL wrote:
[plan1]
> -> Seq Scan on menutable b (cost=0.00..13.01 rows=38 width=22)
> (actual time=0.02..0.38 rows=38 loops=1)

[plan2]
> -> Index Scan using menutable_pkey on menutable b
> (cost=0.00..29.36 rows=38 width=22) (actual time=0.02..0.12 rows=38 loops=1)

It's estimating a cost of 13 for the sequence scan and 29 for the index
scan so it's choosing the sequence scan.

The value of random_page_cost may be too high for your system and
especially so if the database is small and likely to fit in ram. In
addition, if the table in question is small, you'll likely find that at
some point for a larger data set that the system switches over to an index
scan.

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Stephan Szabo 2004-01-16 18:17:50 Re: Join optimisation Quandry
Previous Message scott.marlowe 2004-01-16 17:20:24 Re: shared_buffer value