Date: Sat, 7 May 2005 18:12:30 +0200
From: Vortex <vortex25@gmx.de>
To: Bruno Wolff III <bruno@wolff.to>
Cc: pgsql-sql@postgresql.org
Subject: Re: select within aggregate?
Message-Id: <20050507181230.4044d498.vortex25@gmx.de>
In-Reply-To: <20050506124923.GB14417@wolff.to>
References: <20050506092535.060bca35.vortex25@gmx.de>
	<20050506124923.GB14417@wolff.to>
Organization: private
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Fri, 6 May 2005 07:49:23 -0500
Bruno Wolff III <bruno@wolff.to> wrote:

> You could use max(ts) as long as you can have two requests occur at the
> same time. This might be a reasonable assumption, but it might have been
> better to include a sequence so that each record would be guaranteed to
> have a unique key.
Yes, of course. So far it was more a theoretical question.

> Using DISTINCT ON to get the latest URIs is probably better. You can
> join that query to the one doing the count. This will probably be more
> efficient than executing a subquery for each remote host.
The DISTINCT ON construction looks quite interesting. It seems to allow
some sorting within an group (even althought it's no real aggregate) in
general. I can't say if it's really more eficient than the solution suggested
by R.Muralidharan. But i will do some tests!

Thank you very much!

Klaus