Im currently learning SSAS using the a simplified order schema. The order fact contains multiple dates i.e create_date_key, order_date_key, shipped_date_key each of which is linked back to the calender dimension via a foreign key relationship.
When I process the cube, I end up with a single calendar dimension and multiple dimensions within the cube named after each date_key i.e. CREATE DATE, ORDER DATE, SHIPPED DATE.
Now I thought this was all fine and dandy, then I started reading about role-playing dimensions and did some testing on my cube. It seems OK. The data correlates on a daily basis to what I expected. If I drag in for example both order and invoice date, they seem to group fine.
I'm still concerned I'm missing something with the role playing dimensions. Should I be explicitly creating these? What is the implication if I dont?
"There are only two lasting bequests we can hope to give our children; one of these is roots, the other wings" - Hodding Carter
When I process the cube, I end up with a single calendar dimension and multiple dimensions within the cube named after each date_key i.e. CREATE DATE, ORDER DATE, SHIPPED DATE.
Now I thought this was all fine and dandy, then I started reading about role-playing dimensions and did some testing on my cube. It seems OK. The data correlates on a daily basis to what I expected. If I drag in for example both order and invoice date, they seem to group fine.
I'm still concerned I'm missing something with the role playing dimensions. Should I be explicitly creating these? What is the implication if I dont?
"There are only two lasting bequests we can hope to give our children; one of these is roots, the other wings" - Hodding Carter