I'm creating a multi fact package for my group. I'm joining our invoice fact data to match our sales fact data which is at the order line level. Some of our invoices do not have order lines as they are interest or adjustments without having an established order line.
What I'm wondering is it acceptable to add a 'junk' fact key so that I have results returned regardless of the invoice being associated to an order line or not? The data would have all zeros. I think this would be fine the only problem I can think of is if we do a count of ALL order lines, but that doesn't make sense anyway. My concern is that Report Authors may miss data if I do not have a matching sales fact for the interest/adjustments with zero values.
I have lots of dimensions that I do this for, but I'm not sure this is the right thing to do with fact tables?
Any reasons why I should not do this?
fact: invoice
invoice
transaction id
dim: invoice
fact: factless fact for the join
invoice
order line number
dim:
order line number
fact: sales
order line number