Hi All,
I have seen a few posts similar to this, but none of the suggestions get me anywhere. In my SQL DB, I have a some rows of a fact table that sum to zero at the fact level. The specific column is formatted as decimal(19,5).
In the cube, I have the measure set as a double. When I browse the corresponding fact rows, I initially see a 0.00, but then if you look at the cell properties, it is actually -1.77635683940025E-15
I have reviewed the precision document located here (http://www.microsoft.com/en-ie/download/confirmation.aspx?id=4106) but to me it is stating that this problem can happen with imprecise data types like float and real. In my case, I have a precise data type.
I have switched my measure data type to Currency, and this does solve the problem, but introduces a new problem in that my cube does not reconcile exactly to the warehouse fact tables. This makes sense as Currency used 4 decimals of precision and the warehouse 5. Changing the warehouse to 4 might be an option, but I would like to know if anyone has any further insight to this issue and is there somewhere else that introduces imprecision?
Thanks in advance,
Mark