Hey hi people, I have a challenging question for you guys...
The goal is to create a BI logical schema which allows cross referencing a dynamic number of dimensions.
More precisely, imagine a medical diagnosis composed of many multiple choice symptoms questions; the idea is to cross reference any one of those symptom (list of dynamic possible answers) with a second symptom in a grid chart, having as an expression a simple count of patients.
The challenge here is: the numbers of distinct symptoms is dynamic (and the actual answers are dynamic too). Basically, at the source, the symptoms are values, not fields.
So each symptoms cannot be loaded in a simple separated field. So I though, grouping them into one field would solve that problem... but then as soon as you select the first symptom, it filters out all of the others, which prevents cross referencing.
On top of that, this schema has to allow filtering by a third (and a fourth, etc) symptom answer. Creating a cartesian extrapolation of the data is out of the question here, of course. Heh, we have QV 🙂
In a broad sense, how would approach that?
Thanks a bunch!
Cheers,
Nicolas D.
