Date: Tue, 29 Aug 2006 04:38:58 -0600
From: Michael Fuhr <mike@fuhr.org>
To: Henry Ortega <juandelacruz@gmail.com>
Cc: pgsql-sql@postgresql.org
Subject: Re: Trigger on Insert to Update only newly inserted fields?
Message-ID: <20060829103857.GA20419@winnie.fuhr.org>
References: <2bffcc330608280702y16c25090i816589e20163e7e8@mail.gmail.com>
	<20060828145737.GA5004@winnie.fuhr.org>
	<2bffcc330608280853x50b1034nca246a5dcc935e7@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <2bffcc330608280853x50b1034nca246a5dcc935e7@mail.gmail.com>

On Mon, Aug 28, 2006 at 11:53:36AM -0400, Henry Ortega wrote:
> CREATE FUNCTION updated_end_date() RETURNS trigger AS '
> BEGIN
>    update table set end_date=(select effective-1 from table t2 where
> t2.employee=table.employee and t2.effective>table.effective order by
> t2.effective limit 1);
>    RETURN NEW;
> END;
> ' LANGUAGE 'plpgsql';
> 
> That updates ALL of the records in the table which takes so long.
> Should I be doing things like this? Or is the update query on my trigger
> function so wrong?

You're updating the same table that has the trigger?  Beware of
endless trigger recursion.

You're not restricting the UPDATE with a WHERE clause, which explains
why it updates the entire table.  Maybe you meant this:

  update table set end_date = (...) where employee = new.employee;

The subselect for each row also slows down the update, although you
might not be able to avoid that if requirements demand a potentially
distinct end_date for each row.

-- 
Michael Fuhr