Re: quick question abt pg_dump and restore - Mailing list pgsql-general

From Josh Harrison
Subject Re: quick question abt pg_dump and restore
Date
Msg-id 8d89ea1d0801090851v8b777c0lcdfac4a2973af937@mail.gmail.com
Whole thread Raw
In response to Re: quick question abt pg_dump and restore  (Andrew Sullivan <ajs@crankycanuck.ca>)
Responses Re: quick question abt pg_dump and restore  (Andrew Sullivan <ajs@crankycanuck.ca>)
List pgsql-general


On Jan 9, 2008 11:39 AM, Andrew Sullivan <ajs@crankycanuck.ca> wrote:
On Wed, Jan 09, 2008 at 05:28:15PM +0100, Ivan Sergio Borgonovo wrote:
> Does it make any sense *knowing* how the implementation works to load
> records in a table in a specific order to improve performances?

Well, this is more or less what CLUSTER does.  There are some cases where
happening to know about the order the table is in will yield happy effects,
yes.

You are right. Sometimes when i cluster the table according to the frequently accessed indexes then it makes queries pretty fast. But its not a feasible solution always since some tables have more indexes which are accessed frequently. So clustering the table according to one index will yield poor performance to queries involving other indexes. Index-only scan is a good solution for this I guess for queries involving indexed columns (like in oracle) !!!

josh

pgsql-general by date:

Previous
From: Andrew Sullivan
Date:
Subject: Re: count(*) and bad design was: Experiences with extensibility
Next
From: Andreas Kretschmer
Date:
Subject: Re: quick question abt pg_dump and restore