Thanks, I came up with the same solution has the one you describe.

>For batch updates (you would send the whole recordset back), one way to do it is to work with a blank recordset obtained previously from the original table and persisted on the client. Do a bogus select statement that will return an empty recordset and persist it with the Save method. Then when you are ready to populate open it from disk and start working with it as if you had opened it from the back end.
>
>One caveat of the persisted rs is that you need to devise some way of updating it for changes on the table structure.
>
>A problem you may have already come across with rs built from scratch is that the back end does not recognize it as a valid table when you try to do the batch update.