To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: "Correia, Carla" <Carla.Correia@logicacmg.com>,
	"'pgsql-sql@postgresql.org'" <pgsql-sql@postgresql.org>
Subject: Re: Complex outer joins?
References: <0F232CC93A58D6119C1600B0D0799B8102338822@hamsrvmx03.logica.co.uk>
	<12785.1048519090@sss.pgh.pa.us>
In-Reply-To: <12785.1048519090@sss.pgh.pa.us>
From: Greg Stark <gsstark@mit.edu>
Organization: The Emacs Conspiracy; member since 1992
Date: 26 Mar 2003 09:36:11 -0500
Message-ID: <87smtakxes.fsf@stark.dyndns.tv>
Lines: 30
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Tom Lane <tgl@sss.pgh.pa.us> writes:

> The SQL-standard way of writing this would presumably be either
> 
>     from G left join L on (G.SELID = L.SELID)
>          left join C on (L.SELID = C.SELID and L.SELLEVEL = C.SELLEVEL)

I would think of it as this one.

>     from G left join
>          (L left join C on (L.SELID = C.SELID and L.SELLEVEL = C.SELLEVEL))
>          on (G.SELID = L.SELID)

I don't see how that would be at all different.

> depending on which join you think ought to be done first.  It might be
> that the results are the same in this case, but I'm not convinced of
> that.  In general the results of outer joins definitely depend on join
> order.

I'm pretty sure Oracle actually builds an abstract join representation where
the two queries above would actually be represented the same way. Then decides
the order from amongst the equivalent choices based on performance decisions.

Can you show an example where the join order would affect the result set? I
can't think of any.


-- 
greg