From: Tom Lane <tgl@sss.pgh.pa.us>
To: Richard Broersma Jr <rabroersma@yahoo.com>
Cc: General Postgresql List <pgsql-sql@postgresql.org>
Subject: Re: Alternative to Select in table check constraint
Date: Sat, 01 Jul 2006 00:11:22 -0400
Message-ID: <18959.1151727082@sss.pgh.pa.us> (raw)
In-Reply-To: <20060630215512.18287.qmail@web31813.mail.mud.yahoo.com>
References: <20060630215512.18287.qmail@web31813.mail.mud.yahoo.com>
Richard Broersma Jr <rabroersma@yahoo.com> writes:
> The following codes doesn't work on PostgreSQL 8.1.4 but according to
> the book does conform to SQL-92.
> CHECK ( 1 = ALL ( SELECT COUNT(STATUS)
> FROM BADGES
> WHERE STATUS = 'A'
> GROUP BY EMPNO))
Yeah, the spec does allow sub-SELECTs in CHECK constraints but PG
doesn't implement that. The problem with it is that there's no clear
way to make it perform reasonably, because the CHECK doesn't simply
implicate the row you're currently inserting/updating --- every other
row is potentially referenced by the sub-SELECT, and so changing row
X might make the CHECK condition fail at row Y. A brute-force
implementation would be that every update of any sort to BADGES causes
us to re-evaluate the CHECK constraint (hence, re-scan BADGES entirely
to compute the sub-SELECT) for every row of BADGES (hence, O(N^2) work
if there are N rows in the table). That is certainly unworkable :-(.
A bright person can think of ways to optimize particular cases but
it's not easy to see how the machine might figure it out for arbitrary
SELECTs.
The unique-index hack that Michael suggested amounts to hand-optimizing
the sub-SELECT constraint into something that's efficiently checkable.
regards, tom lane
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-sql@postgresql.org
Cc: tgl@sss.pgh.pa.us, rabroersma@yahoo.com
Subject: Re: Alternative to Select in table check constraint
In-Reply-To: <18959.1151727082@sss.pgh.pa.us>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox