Hi Team,

I've got a requirement to do distinct counts by dimension level for a measure. I read on this forum that this is a feature released with 5.2.x, and indeed, I see an option under settings | Summarization called Distinct Count Aggregation. I read the manual and this is what it says about that:

Select whether to allow aggregate measure values with distinct Dimension
counts for higher-level aggregation.

Sounds perfect, that is what I want.

My question is about how I am supposed to structure the data to support this.

Currently we have counts at the lowest level of all dimensions. Another way of saying this is that we have distinct counts where we've grouped by all associated dimensions at the lowest level.

I can't imagine this would provide the needed information, because how would PMF be able to recalculate the distinct counts for higher level dimensions from already aggregated data?

So, what should my measure source data look like to support this new feature? Do I need to have pre-calculated values at every possible dimension level, and PMF displays those instead of aggregating lower level?

I'm also open to reading a doc about this, but I didn't find anything beyond what I pasted above or saw in this forum.

Cheers,

Joey

You need to set up and load a dimension that will contain the data to break out the distinct count; check out page 290 of the PMF 5.2 documentation, please.

thanks