Hi Team,

I'm loading some ratio type measures, from a table that contains discrete facts for the numerator and denominator.

The situation is that the denominator fact has values for all dimension combinations and the numerator fact has values for some combinations.

PMF loads zeros for combinations where there are no numerator values, which is accurate as far as math is concerned (zero divided by something is zero). I end up with measure by dimension breakout that is full of zeros. These are weighted against the 'real' values, for a high level average that is lower than it should be from a requirements perspective. 0, 3, 0 averages to 1, but I'd like to average to 3.

I wonder if there is a way to filter away these zero values instead of loading them, or suppress them somehow in the system?

Cheers,

Joey

This message has been edited. Last edited by: Moogle,

Hey Joey,

If you don't want these at all in the system, you can filter them out at the Measure load. Just click on the Filters tab on the Measure Load Panel and add the appropriate filter. You can ither use the drop downs or hand code the filter. Either way, it would be something like this:

NE 0.0

Eric

Hi Eric,

Thank you for your reply. I don't think I was clear on what I wanted, because filtering for NE 0.0 works at the fact level, not at the measure level where I am seeing the zeros. However, I did find a solution, and it was obvious in retrospect:

I filtered away non-valid dimensions.

So, if my data looked like this:

AB 100
ON 0
MB 50

I put a measure load filter saying WHERE PROVINCE_CODE NE 'ON' and then the report looked like this:

AB 100
MB 50

Works well.

Cheers,

Joey