Date: Fri, 27 Jun 2003 10:10:25 -0500
From: Bruno Wolff III <bruno@wolff.to>
To: =?iso-8859-1?Q?Egy=FCd?= Csaba <csegyud@freemail.hu>
Cc: pgsql-sql@postgresql.org
Subject: Re: Getting all rows even if not a member of any groups
Message-ID: <20030627151025.GE15269@wolff.to>
Mail-Followup-To: =?iso-8859-1?Q?Egy=FCd?= Csaba <csegyud@freemail.hu>,
	pgsql-sql@postgresql.org
References: <11362.1056313692@www59.gmx.net>
	<00bd01c33c9d$1b584e90$230a0a0a@compaq>
	<20030627140854.GA15269@wolff.to>
	<00e501c33cb6$f4cc4cd0$230a0a0a@compaq>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <00e501c33cb6$f4cc4cd0$230a0a0a@compaq>
User-Agent: Mutt/1.5.4i

On Fri, Jun 27, 2003 at 16:18:10 +0200,
  Együd Csaba <csegyud@freemail.hu> wrote:
> 
> This is absolutelly what I want, but I can't understand how it is working.
> Where can I find a descriptive (tale-like, for kids ... :) ) documentation
> about using joins?

If you look at the documentation for the select command and page down a
bit there is a description of join syntax. Note that in 7.4 using
the explicit join syntax won't force join order. (This really only
affects cross joins and inner joins; left and right joins normally
can't be reordered.)

You need a left join to pick up products that aren't in any group.
The parenthesis changed the join order so that group names were attached
to group ids before group ids were joined to products. This can have
performance implications. I think that this is probably the faster
way, but the other option would to have been to make the second join
a left join as well.