I have a relational data warehouse, and a SSAS cube built on the DW.
In the relational DW, the fact table can be
1. either one table with partition,
2. or multiple fact tables with the same structure.
In the cube, I can create multiple partions.
In case 1, the cube partions are bound to query to the same table but different partion conditions.
In case 2, the cube partions are bound to different fact tables.
My questions are --
a. Are the above understanding right?
b. In both case 1 and 2, what fact tables are in the DSV? Especially in case 2, do I add all fact partitioned tables in DSV and create relationship with all dimensions?
c. What is the ETL of the relational DW for both case 1 and 2?
d. How do I process cube on changes? One thought is when doing ETL, record which partions are changed and processfull of that cube partition.
Many thanks in advanced.