From: Ron Mayer <rm_pg@cheapcomplexdevices.com>
Subject: Re: [SQL] OFFSET impact on Performance???
Date: Thu, 20 Jan 2005 08:49:39 -0800
Organization: Hub.Org Networking Services
Lines: 23
Message-ID: <41EFE123.9020209@cheapcomplexdevices.com>
References: <019f01c4fee1$1b58d600$0b00a8c0@forge>
	<41EF9FD3.5020400@archonet.com>
	<020d01c4fef6$595bb9d0$0b00a8c0@forge>
	<41EFCC5B.5010302@archonet.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
To: Richard Huxton <dev@archonet.com>
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
In-Reply-To: <41EFCC5B.5010302@archonet.com>
To: pgsql-performance@postgresql.org

Richard Huxton wrote:
> 
> If you've got a web-application then you'll probably want to insert the 
> results into a cache table for later use.
> 

If I have quite a bit of activity like this (people selecting 10000 out
of a few million rows and paging through them in a web browser), would
it be good to have a single table with a userid column shared by all
users, or a separate table for each user that can be truncated/dropped?

I started out with one table; but with people doing 10s of thousand
of inserts and deletes per session, I had a pretty hard time figuring
out a reasonable vacuum strategy.

Eventually I started doing a whole bunch of create table tmp_XXXX
tables where XXXX is a userid; and a script to drop these tables - but
that's quite ugly in a different way.

With 8.0 I guess I'll try the single table again - perhaps what I
want may be to always have a I/O throttled vacuum running...  hmm.

Any suggestions?