1.) Obviously you'll need a level-based measure which represents the total. Do you have that?
2.) Do you want one new measure called "%" which dynamically reacts to a dimensional attribute X/Y/Z added to the analysis as an axis or are you expecting 3 measures for X, Y and Z respectively and the attribute won't be used at all?
1. I have splitted the column with different types of invoices into two in a pivot table. In the first column I have only X-type invoices, in the second one - all others. I have counted number of rows in grand totals for the pivot table, so basically I can calculate % = Count of X-type invoices / Count of (X-type + other) manually, but I do not know how to do it automatically.
Could you please explain how to make a level-based measure which represents the total?
2. I want separate measures, in particular only % of X-type invoices in the total amount of invoices, I do not need a dimensional attribute.
Every fact table should be built with the Count column having a 1 default value ... use of information drives physical design. Having that count column makes your life so much easier.
I agree, Thomas
It's DW 101 stuff ... so have your team get it done!