Re: query optimiser changes 6.5->7.0 - Mailing list pgsql-general

From Tom Lane
Subject Re: query optimiser changes 6.5->7.0
Date
Msg-id 9879.959913587@sss.pgh.pa.us
Whole thread Raw
In response to query optimiser changes 6.5->7.0  ("Simon Hardingham" <simon@netxtra.net>)
List pgsql-general
"Simon Hardingham" <simon@netxtra.net> writes:
> I have run explain on the query and it shows that it is just
> performed a sequential scan on version 7.0

> Seq Scan on gazet  (cost.....)

> On the old version (6.5.1) it reports

> Index Scan using gazet_index on gazet  (cost=....

> Any suggestions as to how I can improve performance on this databases new
> server?

Unfortunately you have been careful to suppress any information that
might actually let someone give you useful advice ;-).  There are
several threads in the archives about undesirable index-vs-sequential-
scan choices in 7.0; check pgsql-sql as well as pgsql-general for the
last month or so.

Also, 7.0.1, propagating now to an archive near you, contains some
fudge-factor twiddling to make it more willing to choose an indexscan.
We shall soon find out whether that made things better or worse for
typical uses...

            regards, tom lane

pgsql-general by date:

Previous
From: Michael Blakeley
Date:
Subject: interval questions
Next
From: Tom Lane
Date:
Subject: Re: interval questions