From: jboes@nexcerpt.com (Jeff Boes)
Subject: How to recognize trigger-inserted rows?
Date: 26 Oct 2004 10:00:01 -0700
Organization: http://groups.google.com
Lines: 18
Message-ID: <d40a65a1.0410260900.695a1ac6@posting.google.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
To: pgsql-sql@postgresql.org

I have a table with an INSERT/UPDATE/DELETE statement trigger. The
trigger's action is to insert "jobs" into a queue noting that the
table has changed.

A number of other tables have FK relationships with this table, and
they have their own statement triggers that fire on DELETE.

When I delete a number of rows from the first table, the cascading
deletes into the other tables generate a rather large number of
trigger-fires, so I end up with way too many rows in the queue-table.

What I would like to do is, within the transaction doing the top-level
delete, examine the queue-table for duplicate rows and remove those,
since they are extraneous.

Ideally I would look for rows that have the same transaction ID, but
I'm having trouble determining what the current ID is. I'm using Pg
7.4.