Overlapping typed datasets in .NET2: best practices?

[henk gijsbert]

Assume I have one adressTable that is used for as well companies as
persons and I have need to retrieve companies with adresses as well as
persons with adresses. I could generate typed datasets, one with
companytable+addresstable and one with personTable+addresstable,
resulting in as well identical addresstables generated in both datasets
as identical AddressTableAdapters.

This doesn't feel right. What is a good practice in such a case:
-Do it anyway as described
-Generate 3 typed datasets with 1 table each and use untyped datasets
to combine the typed tables (this is what I am thinking of)
-Use one typed dataset and one untyped
-something else?

Regards,
Henk

 

[Miha Markic [MVP C#]]

I use a dataset per process. Thus your option 1.

 

[henk gijsbert]

Thanks Miha. Is there anyone with an opposing opinion?

 

[Miha Markic [MVP C#]]

btw, I use template based code generator (CodeSmith) to automatically
produce all required datasets...