>>Then why use two different strongly-typed datasets in the first place ?

Ah.. you ask those probing questions, Viv.
The source tables here are VFP. They are reference tables that aren't updated by this app, and have record counts less than 1K.


The nature of this app is such that as many as 10 queries of those tables might be required by a transaction before the final result for that transaction can be determined.

In order to avoid going back to the VFP sources repeatedly, I load one dataset with them when the app loads and plan to query from that DS to another while executing the transaction at hand.
My assumption is that a query against a low record count DT will be a lot faster than a round trip to a VFP table.
I can get them into a datarow OK but I'd like to use intellisense here because a lot of fields are used in the calculations, so I conjured up the second instance of the dataset as a solution.

There are not two .xsd's - just two instances of the same .xsd.