Here's some background:
We have a cube that has 1 measure group based on a view of a fact table.
We want to partition this cube as it is getting rather large. We have decided to partition that cube based on year. The desired split means we will have 4 partitions: 3 Years and Older, 2 Years ago, 1 Year Ago and Current year. For ease of changing at the begninning of our fiscal year, I made 4 views that are identical to the original fact view with the exception of having a WHERE clause that limited the data to the year(s) mentioned earlier. I then added those 4 views to the DSV and made the appropriate relationships for each view.
I created the partitions based on the related views and deleted the original partition that was all the records. I processed everything and it processed fine. When I went to go browse the data though, I didn't have any. That brought me to checking the measure sources in the measure group. They pointed to the fact table that was getting the whole chunk of data (no WHERE clause).
Question 1: If my partitions are based on different fact tables, where do I point to as the "source" for a particular measure?
Question 2: Is this the recommended approach to doing partitions? Is it normally done via query that limits from 1 fact table?
Eat.Sleep.Code.