I was wondering if it is at all possible to filter out a calculated field in Nprinting?
In my case, the calculation is as follows:
=if(count(distinct([Order Unit Price]))=1, 'L', if(count(distinct([Order Unit Price])) <4, 'M', 'H' ))
This is basically giving a score for the number of different prices being offered, by product and customer (Low, medium, high). The score is on a straight table, which is pasted into Excel in Nprinting.
I would like to filter out all of the 'L's (Lows) from Nprinting, as they are not relevant to the user. I tried creating a variable that aggregates the expression over product and customer and then filtering out the 'L's in that variable, however the filter only selects the relevant products, not customer/products.
There is a workaround that I could use, but it would involve setting a conditional colour of white to hide rows that are 'Low'. For the sake of performance and processing time (the Nprinting takes 16 hours normally), is there another way to do it?
Calculated dimension should still work ok - just give it a try and see how it goes.
on the other hand I would do it in script....
In Load script it would be ideal, and better for performance, but I guess such composite key may be difficult to achieve as it may be coming from different dimensional tables. Then you would have to create such key probably in Fact Table. At the same time if you decide to create such concatenated key in data model then also you can calculate your aggregate (count(distinct([Order Unit Price]))) in script and flag records greater than 1... and than just use a flag as a filter in your report.
but your quick win is calculated dimension - I just tested it and it worked.
cheers Lech, When applicable please mark the correct/appropriate replies as "solution" (you can mark up to 3 "solutions". Please LIKE threads if the provided solution is helpful to the problem.