I have a FACT table with over 60 million rows and over 20 dimensions, all but one dimension (customer-14 million rows) are small. The CUBE processing is taking 45 minutes. I tried processing all the related dimensions first and then processed the cube with not much of a difference. I then created Query based partitions based of the Year by joining the FACT to the Time Dimension. In total i created 10 partitions and there is a slight improvement , 42 mins. I have the following questions.
1) Do i need to create Aggregations?
2) Do i need to create 10 aggregations, one per each partition.
3) I do not want to include certain dimensions in the aggregation (so select None for these dimensions).
Thanks for your time.