From: Richard Huxton <dev@archonet.com>
To: Robert Treat <xzilla@users.sourceforge.net>, Tom Lane <tgl@sss.pgh.pa.us>
Subject: Re: partial unique constraint
Date: Tue, 6 Apr 2004 20:54:07 +0100
User-Agent: KMail/1.5
Cc: pgsql-sql@postgresql.org
References: <1081261772.31785.48.camel@camel> <24888.1081264623@sss.pgh.pa.us>
	<1081269521.31791.73.camel@camel>
In-Reply-To: <1081269521.31791.73.camel@camel>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200404062054.07498.dev@archonet.com>

On Tuesday 06 April 2004 17:38, Robert Treat wrote:
> On Tue, 2004-04-06 at 11:17, Tom Lane wrote:
> > Robert Treat <xzilla@users.sourceforge.net> writes:
> > > Trying to come up with the proper syntax to meet the following
> > > criteria: create table foo (bar integer, baz boolean UNIQUE (bar, baz =
> > > true));
> >
> > The correct way to do it is with a separate CREATE INDEX command:
>
> I think I had initially abandoned looking at that type of solution after
> having run across this paragraph in the docs while looking for the
> proper constraint syntax:
>
> "Note:  The preferred way to add a unique constraint to a table is ALTER
> TABLE ... ADD CONSTRAINT. The use of indexes to enforce unique
> constraints could be considered an implementation detail that should not
> be accessed directly."
>
> The subliminal mind is a powerful force eh? Perhaps that paragraph
> should be modified... but I'm not sure if it should expanded to include
> thoughts along the lines of Stephan's response or maybe just drop the
> "should not be accessed directly" bit...

Anyone know if there is any way to do this in one of the later SQL standards? 
The CREATE INDEX thing is a bit of a hack, and I certainly wouldn't have 
thought of it either.

-- 
  Richard Huxton
  Archonet Ltd