From: Josh Berkus <josh@agliodbs.com>
Reply-To: josh@agliodbs.com
Organization: Aglio Database Solutions
To: elein <elein@varlena.com>
Subject: Re: [SQL] SQL Spec Compliance Questions
Date: Fri, 4 Jun 2004 12:20:47 -0700
User-Agent: KMail/1.5.4
Cc: postgresql advocacy <pgsql-advocacy@postgresql.org>
References: <200403302050.12415.josh@agliodbs.com>
	<200406031609.51415.josh@agliodbs.com>
	<20040604113316.K11485@cookie.varlena.com>
In-Reply-To: <20040604113316.K11485@cookie.varlena.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200406041220.47452.josh@agliodbs.com>

Elein,

> In postgres and Illustra, references were possible.
> That is, storing the OID (REF) of a table in a column of another
> table achieving the same effect.  SQL syntax support
> was needed for the REF type.  I believe this was removed
> or suppressed in PostgreSQL.

Yeah, I'd like to see a return of table references as FKs.   It would allow me 
to eliminate all of these rather non-relational integer-based surrogate keys.

I don't think our OIDs would be the way to go on this, given the problems 
already discussed with PG OIDs.   I kinda like SyBase's solution, where they 
have a system-acessable hash key of the data in the row.

-- 
-Josh Berkus
 Aglio Database Solutions
 San Francisco