I have read both ways, but my main cube has 10 measure groups, and each measure group has a couple partitions.
Routinely, I need to process just one of the 10 measure groups hourly.....
Do I need to process index on the whole cube (all 10 measure groups) afterwards? or just process index on the 1 measure group..
I read that the cube's index goes away after the process data.. so you would need to process it for the whole cube.. but i've also read the contrary..
anyone? thanks!