I'm in the process of learning SAP BW at work. The SAP class material states that in InfoCubes, dimensions are denormalized. I understand the benefit of normalized tables, but don't understand why dimensions would be denormalized.
In a data warehouse, you're trying to get a faster response time.
The data is loaded one time, and read many times. Therefore, normalization of data is not an advantage in a data warehouse.
Denormalization is a bit of a misnomer. In a star schema, the data is not normalized. I do understand that some people are referring to the data loading process as denormalization. But, in general, designing a star schema is not a denormalization
By not normalizing the dimension data, you're improving the response time by increasing the load time and storage space for the data warehouse.