Message-ID: <1CAD483B723BD611B0C10090274FF0685547CC@NXCDMAIL>
From: Peter Atkins <peter.atkins@NXCD.com>
To: 'Ken Corey' <ken.corey@atomic-interactive.com>
Cc: "'pgsql-sql@postgresql.org'" <pgsql-sql@postgresql.org>
Subject: Re: Returning PK of first insert for second insert use.
Date: Mon, 29 Jul 2002 13:20:01 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"

Thank you for explaining that in detail it makes sense now. I'll give it a
try.
Thanks again!
-p

-----Original Message-----
From: Ken Corey [mailto:ken.corey@atomic-interactive.com]
Sent: Monday, July 29, 2002 1:05 PM
To: Peter Atkins
Cc: 'pgsql-sql@postgresql.org'
Subject: RE: Returning PK of first insert for second insert use.


On Mon, 2002-07-29 at 20:52, Peter Atkins wrote:
> Is there a possibility of another application accessing the DB and using
the
> id before my function has completed the transaction? I'm concerned with
the
> possibility of cross-over of ID's if the insert hangs.
> 
> There's no way to return the id of that insert inherently, and then use it
> for the second insert? I think SQL uses something like ADD_ID, not sure.

That's the beauty of the nextval statement. The database internally
sequences requests to it so that you're kept out of harm's way.

Say process A called the function,and nextval returns 16.  The function
now continues on its way, but is not finished when process B then calls
the function (before A is done), and nextval returns 17.

So, then function called by process A returns 16, and the function
called by process B returns 17.

That means that unless the results of process B depend in some way upon
the results of process A, there's no problem.

-Ken

-- 
Ken Corey  CTO  http://www.atomic-interactive.com  07720 440 731