We have a scenario where we need to provide the As on the given day how many products available in the inventory across all possible dimensions. If any property changes we need to show the new value from that date onwards. Instead of creating the snapshots for each day with all the dimensions we created the Type 2 on the table and using the distinct count on the ProductID as a measure.
The issue was when we build the cube we used LinkMembers to get the Active products on a given day(using SCDStart and SCDEnd dates). This is scanning the entire cube for each user query.
Do we have any other design for optimal performance or how to improve LinkeMember performance?