> Jason's original request could be satisfied if the procedure did a FOR

SELECT

> loop to drive the updates with no change to the current architecture, but
> that leaves open the original question: What problems are solved by

knowing

> the number of rows affected?

Yes, I do often use FOR SELECT loops with the DB_KEY and then do the update
using the DB_KEY lookup which is an acceptable way to accomplish the same
thing. Although I must say the AFFECTED clause makes things a little cleaner
and self-evident at a glance.

My ulterior motive was to get the powers that be to recognize that
RowsAffected is precisely the loop count of the operation's rows walked and
not some physical inventory of records tweaked somehow.

Seems I have established my point of their equivalence.

Thanks for taking the bait Ann. <g>

Regards,
Jason Wharton
CPS - Mesa AZ
http://www.ibobjects.com