• Hi Andy,

    "One of the rules is that partitioning columns have to be a part of any unique index (as I understand it, this helps the optimizer know that it only needs to check a single partition) if you want that index to be on the same partition."

    I've read this over and over, and I still don't get it. By adding another field to my PK or UQ, I change what can be unique in that table. i.e. if previously ContactId was my primary key, and now I add ModifiedDate, then I can end up with an infinite number of duplicates for a given ContactId.

    Am I missing something?

    Thanks,

    David.