The data source view is a very important concept in Analysis Services. You can think of it as an abstraction layer between the relational database and the multidimensional cube. You have the ability to define objects that do not exist in the database. The data source view provides a logical view of the database, so to Analysis Services, the objects appear to exist in the database. This includes new tables, columns, and even referential integrity or relationships between objects.
Tip
In many cases you do not define referential integrity down in the database layer when you work with data warehouses. The reason for this is that loading of the data is much easier without them. If you have defined foreign keys and primary keys, you need to ensure that you always load your objects in the correct order.
If your data warehouse does not contain referential integrity, it has to be created in the data source view.
You can also add calculations to existing tables. They...