I understand that Direct Query is going to be slow when I’m loading 500 million rows.
I guess I’m looking for a solution that doesn’t require me to maintain multiple versions of the same dataset that are not identical in every way except for the fact that one has RLSS applied and one does not.
The limitations of Quicksight are frustrating, and make it feel half baked when you hit them.
With RLS , you cannot duplicate a SPICE dataset. But you can use the same SPICE dataset as a new dataset. This new dataset only supports Direct query (with RLS implemented ) , so you are essentially querying the same SPICE dataset.