Alan Hodgson wrote:
> On Wednesday 18 July 2007 14:29, "Roderick A. Anderson" <raanders@acm.org>
> wrote:
>> In the mean time if the script gets triggered again and the first
>> instance isn't finished the second needs to not be able to select those
>> records already being handled.
>
> select for update won't do that. It will sit waiting for locks on the same
> rows the first process is handling.
Let's see how do I put it ... Da[r|m]n! But I now remember reading
something to this effect in one of my PG books or the on line docs.
Some other method is now in order.
Thanks to all for the thoughts and ideas.
I'll post my solution ... when I figure one out!
Rod
--