On Thu, 25 Sep 2003, Rory Campbell-Lange wrote:
> Thanks for the reply, Stephan. Sorry about the late reply.
>
> Your dissection of the standard (and thanks for the translation, by
> the way!) is a clear reason to use unique row ids. Unfortunately the
> project I am reporting bugs on does not do this at present.
Well, if you're looking for a PostgreSQL only solution (to throw in the
switch), I think someone already sent a query using oid and a subselect.
It's not going to perform super well probably and will only work on
tables with oids, but it should only update one row.