I have a cube that has 14 million facts.
A client has a complex Excel model with 33 pivot tables running off this cube. It takes a long time to refresh the data. Using SQL Profiler I see that it is issuing the CREATE SESSION CUBE command 17 times.
Total time for the full refresh to run is 6.5 minutes. If I run the refresh again it takes 15.5 seconds. The 17 CREATE SESSION CUBE commands take 5 minutes to run.
I have several users using this model.
How can I reduce the time taken for the CREATE SESSION CUBE? Is it better use of partitions or aggregations?
Thanks