I currently have a cube that does simple calculations and aggregations. The cube pulls in about 5 years of fact table information (80 million rows) and has multiple dim tables supporting it. The cube contains 28 dim and two of them have about 15 attributes in them. The cube has also been partitioned by year.
As we've added dim, the cube has started to processes slower and slower, but not the dim. We have added some dim that do not contain all the keys in the fact table (by design) but have set the properties on the dim to ignore when a key is not found ( not sure if thats the best way to approach that). Anyone have any experience with this or faced a similar situtation?
