Table Denormalization Study in PowerPivot
I recently talked about memory usage in PowerPivot and previously I wrote about difference modeling options using table normalization or not. Thus, I wondered whether table normalization is a good practice for performance and memory or not. The short answer is that yes, it might be better but not so much as you might suppose (and there are difference considerations about processing and query performance). Rule of thumb: a star schema is always a good thing! Have a good reading.
PowerPivot allows definition of relationships between tables loaded in a PowerPivot model. However, whenever these tables come from the same relational database, the same data could be loaded into a single table, by using data denormalization in the SQL query itself.