I've got a fact table with two fields:

CODE
MINUTES

I've got a dimension table that matches to CODE:

CODE
DESCRIPTION

In the cube editor, I move CODE to be a dimension.  Process the cube - everything fine.  Reviewing the data, I use the CODE dimension to choose a single code value and bingo, I get the right information.


Now attempt #2:

I setup what I belive is a star-schema dimension where code on the fact table is linked to code on the dimension table.

This is fine except that I notice that my total record count reflected by the cube drops.  A little analysis and I discover that not all the codes in the fact table are defined in the dimension table.  In some cases the code in the fact table is NULL, in other cases it simply is not found in the dimension table.

So it looks to me as if the cube eliminates those rows from the fact table (so to speak).  Is this the only thing that can be done or is there a way to retain that information?