OID and PL/pgSQL trigger :-(

From: Alke <alke(at)iol(dot)it>
To: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: OID and PL/pgSQL trigger :-(
Date: 1999-08-23 13:40:58
Message-ID: 37C14F6A.EEB14AC4@iol.it
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hy all!

I posted this mail on the general and sql mailing-list, but I got no
aswer, so I try to post it here;
if the question is bad-explained, please ask me more details...

I would like to use OIDs as a primary key on every table of my db,
because this seem to me a very smart thing to do..... but after some
experiment, I've decided to create a SERIAL field on every table, and to

use it as the primary key, and to abandon OIDs...

This is an example of what I whould like to write (and I can't):

CREATE FUNCTION firstTable_DelCascade() RETURNS OPAQUE AS '
BEGIN
DELETE FROM anotherTable WHERE anotherTable.OIDfromFirstTable =
OLD.OID;
RETURN OLD;
END;
' LANGUAGE 'plpgsql';

Well, this can't work because OLD is of type RECORD, and type record
have no OID :-(
(this is what I found in the docs, about RECORD:
"Only the user attributes of a table row are accessible in the row, no
Oid or other system attributes (hence the row could be from a view and
view rows don't have useful system attributes)" )

The main question is: can I get the OID of the OLD/NEW record in
triggers?
A more general question is: what is the usefulness of OIDs, if I can't
use them in trigger?

Paolo

Browse pgsql-hackers by date

  From Date Subject
Next Message José Soares 1999-08-23 14:02:38 Re: [HACKERS] RE: [SQL] datediff function
Previous Message Alke 1999-08-23 13:37:37 Array-fields and many-to-many relations