Please note that I just read couple of articles before working on this development, I still have some questions, that's the reason for this post.
I am building a shipments InfoCube that has ship date, order creation date, requested delivery date. I have set Sales Order, Shipment, Material as line item dimensions since they are pretty unique.
I created a new 'Date' dimension with high cardinality option in the InfoCube for these four fields together. After the full load I noticed that the high cardinality date dimension has 350,000 records. The total possible records in this InfoCube is about 1.5 million.
Questions -
- Does it make sense to have separate dimensions for these 4 date fields or does one high cardinality dimension for the 4 fields make more sense in this case. If I create 4 separate date dimensions for these 4 date fields, the maximum number of records per dimension would be 3000.
- Do high number of dimensions affect the performance of the InfoCube? In this case I have reached the highest possible limit of 16 once I separated the dates into unique dimensions.