Handling Transactional-Level Data - Expert Cube Development with SSAS Multidimensional Models - page 131

At this point, we create a new measure group named Bridge Sales Reasons ,
containing a single, hidden measure named Bridge Sales Reasons Count
that will count the number of rows of the BridgeOrdersSalesReasons table;
this measure might never be used in queries, but every measure group has to
contain at least one measure. This will be our intermediate measure group in the
many-to-many dimension relationship. In the Dimension Usage tab, we add
the Sales Reason dimension, which will have a regular relationship with the
Bridge Sales Reasons measure group. What is less intuitive is how to create the
relationship between the Sales Order dimension and the new measure group.
The Sales Order dimension is a degenerate dimension and its content is stored
in the fact table. Moreover, the granularity of the bridge table is the order number,
while the granularity of the Sales Order dimension is the order line number.
However, we can define a regular relationship with a non-key granularity attribute,
such as Order Number using the SalesOrderNumber column, as we can see in the
following screenshot: