Dimensions not matching Selections

[LReeve]

Hi All,

I am having an issue with an app we are currently working on.

The idea of the measure is to give the unique volume of customers who are new to our business in the past year.

• If you look at the measure with no selections then the intended result is a new customer for the entire company regardless of Type.
• If you filter by Type we expect to see the new customers or that Type ignoring if they were already a customer of a different Type.

The results I am seeing is that if you select one of the types it gives the correct result - but if the measure is a chart of any type it functions similar to my first bullet, but limiting the it to new customers in type 1, compared to customer of any type.

The formula is:

Count({<[Customer Number] = E({<PERIODO = {'201809'}>} [Customer Number])>*<PERIODO = {'201709'}>}  Distinct [Customer Number])

The E() seems to react to Selections but the Dimensions in the bar.

Sample app attached - the correct result is the numbers provided when a selection is made. NOT what is in the chart to begin with.

Would appreciate any help to see if I'm going wrong or if it not behaving as expected.

[sunny_talwar · Accepted solution]

I believe that this is something you won't be able to perform using set analysis because set analysis is performed once per chart and not per dimension. May be use Aggr() to do this

Count({<PERIODO = {'201709'}>} Aggr(If(Count(DISTINCT {<PERIODO = {'201809'}>} [Customer Number]) = 0, [Customer Number]), [Customer Number], Type))

[LReeve]

Hi Sunny,

That makes complete sense and feel a fool for missing it!

Thanks for the answer -