Mulitple Fact Tables in a Star Schema

I am working for an institution that has more than one fact table in a star schema

I have never seen anything like this

I know that we have performance problems but I would like some input on why these performance problems exist and any other implications of multiple fact table scenarios

Regards

Kamal

Yes, I know that a Star Schema by definition should only have one Fact Table. However the data warehouse that has been developed at the organization that I am now working for, for whatever reason, has anywhere from 2 to 5 fact tables all linked together.

I know that this is wrong but I am faced with the challenge of explaining not only why it is wrong but the serious implications of this in terms of poor performance and lack of future capability.

Any assistance I can get I would be grateful for.


Regards

Kamal

By definition, a star schema is a single fact table with its related dimensions. A dimensional warehouse is a collection of star schema that share common dimensions.

It is not clear what you are describing. If you mean the star has two fact tables with direct relationships between the two, then yes, that is not a star schema and performance will not be very good. Facts tend to be very large tables and direct joins between two very large tables doesn't perform well. In a dimensional warehouse you do not combine facts using direct joins between the fact tables.