>>We took a similar approach (Temp Table used in a join) for a large nightly update. Our original approach was row-by-row because we thought there would be fewer contention issues. We did not understand, at that time how powerful SQL Server is using a set based approach. Our job went from hours to minutes to complete with no contention issues. For other ideas, google "RBAR SQL"
>
>Thanks, how did you negotiate with unique updates syntax, as far as variations in the fields to update, in a batch approach?

Our scenario was different in that we were doing mostly inserts and deletes. If your job is truly unique updates, I'm not sure how to implement that as a batch process. Could you create datatables or Temp tables of your modified data, then use MERGE to update the target tables?