Date: Thu, 13 Oct 2005 14:43:03 -0400
From: Andrew Sullivan <ajs@crankycanuck.ca>
To: pgsql-sql@postgresql.org
Subject: Re: pg, mysql comparison with "group by" clause
Message-ID: <20051013184303.GQ16317@phlogiston.dyndns.org>
References: <001001c5cea8$77cabc90$0300a8c0@dell8200>
	<20051011142658.C54076@megazone.bigpanda.com>
	<87br1u1jvq.fsf@stark.xeocode.com>
	<1129155875.29961.168.camel@state.g2switchworks.com>
	<87zmpez0au.fsf@stark.xeocode.com>
	<1129219046.29961.184.camel@state.g2switchworks.com>
	<878xwxi87h.fsf@stark.xeocode.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <878xwxi87h.fsf@stark.xeocode.com>
User-Agent: Mutt/1.5.9i

On Thu, Oct 13, 2005 at 02:26:58PM -0400, Greg Stark wrote:
> Scott Marlowe <smarlowe@g2switchworks.com> writes:
> > could run this query at the same time and get different data from the
> > same set and the same point in time. 
> 
> I'm pretty unsympathetic to the "we should make a language less powerful and
> more awkward because someone might use it wrong" argument.

That's not what Scott's saying.  Scott is saying that the syntax
you're talking about is _formally wrong_.  That's surely not "more
powerful", except in the sense that stepping on a land mine is more
powerful than many other ways you could shoot yourself in the foot.  

> path. In an ideal world the user should be guaranteed that
> equivalent queries would always result in the same plan regardless
> of how they're written.

And again, I say it sounds like you're actually arguing for "the
optimiser needs to get better".  Special-purpose, formally wrong
syntax is surely not better than making the optimiser get the right
syntax right every time, is it?

A

-- 
Andrew Sullivan  | ajs@crankycanuck.ca
When my information changes, I alter my conclusions.  What do you do sir?
		--attr. John Maynard Keynes