I have two fact tables with 0:n cardinality. This is for a hosptal operating theatre. Consider: blocks of time are timetabled for operations to be carried out for a list of patients. These blocks of time are called sessions,
and the operations that are booked to be done in those blocks of time 'belong' to those sessions. So, we have a 1:n cardinality relationship between sessions and operations. However, not all operations fall into this 'pre-booked' category.
There are emergencies. These simply get done right there and then, and don't 'belong' to any session. There's a 0:n cardinality relationship between sessions and emergencies. Booked operations and emergencies are interspersed throughout
One piece of information I want is a count of operations per session. That implies a referenced relationship between my operations measure group and my sessions dimension, via the operations dimension as intermediate dimension (I should add
that in my cube, the sessions fact has an associated 1:1 sessions dimension, and the operations measure group likewise has a 1:1 fact dimension). Of course, any attempt to process a cube with such a relationship fails, as some of the operations
side of the fact:fact relationship don't have&nbs
View Complete Post