For eg, If we take 3 facts alone for analysis, the data in the Facts are at following levels.
Fact 1 : has data at date_key and Product_Key
Fact 2: has data at date_key, Entity_Key
Fact 3 : has data date Key, Entity_Key, Selling_Entity_Key
When I start creating the link table, I am getting "data inconsistency type D" which indicates RAM being used to the maximum.
I also observed that if I just use the Fact 1 and Fact 2, with a outer join populating a Key / Link table, the number of records are a Cartesian products. This is logical as we have 2 different set of level in the Fact 1 and Fact 2. I did this test for a data set of 10 rows.
But I understand from this, I wont be able to create the Key table using this approach as it would be a Cartesian of all the table rows.
Can you please guide me in using the approach to handle this kind of different Facts with just 1 -2 related fields and other unrelated key fields among the fact tables.