Date: Fri, 26 Aug 2005 04:06:35 +0200
From: Tobias Brox <tobias@nordicbet.com>
To: "Jeffrey W. Baker" <jwbaker@acm.org>
Cc: Tobias Brox <tobias@nordicbet.com>,
	pgsql-performance@postgresql.org
Subject: Re: Limit + group + join
Message-ID: <20050826020635.GM10328@tobias.lan>
References: <20050826002709.GK10328@tobias.lan>
	<1125021419.16451.0.camel@noodles>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1125021419.16451.0.camel@noodles>
Organization: Group Nordicbet
User-Agent: Mutt/1.5.10i

[Jeffrey W. Baker - Thu at 06:56:59PM -0700]
> > explain select c.id from c join b on c_id=c.id  group by c.id order by c.id
> > desc limit 5;
> 
> Where's b in this join clause?

"join b on c_id=c.id"

It just a funny way of writing:

select c.id from c,b where c_id=c.id  group by c.id order by c.id desc limit 5;

> It looks like a cartesian product to me.

No.  The query will return exactly the same as the simplest query:

  select c.id from c order by c.id  desc limit 5;   

As said, this is a gross oversimplification of the production envorinment.
In the production environment, I really need to use both join, group and
limit.  I tested a bit with subqueries, it was not a good solution
(selecting really a lot of rows and aggregates from many of the tables).

The next idea is to hack it up by manually finding out where the "limit"
will cut, and place a restriction in the where-part of the query.

-- 
Notice of Confidentiality: This information may be confidential, and
blah-blah-blah - so please keep your eyes closed.  Please delete and destroy
this email.  Failure to comply will cause my lawyer to yawn.