Dynamic Eliminations
I’ve been trying to do some dynamic eliminations using sets of sets. i.e creating two crossjoined sets (from 2 different dimensions) then making a set of the resulting sets. Something that I managed to do to great effect in BI Office. (I actually got the idea from Ian McDonald on the community website:https://community.pyramidanalytics.com/t/k9mz3h/dynamic-eliminations). I can create the lists ok using the formula section (although I had to write the MDX because I couldn’t find the functionality to drag and drop); however when I try and bring a cross-joined set into a report, it will only go into the ‘background’ section, there doesn’t seem to be the ability to bring it into the report. Is this a bug or has the ability to use dynamic eliminations been removed in this version? Or am I going about this the wrong way?
Cheers
Rob
3 replies
-
Hi Rob,
TBH, I haven't tried it in Pyramid. I'll take a look and let you know. Watch this space.
Ian
-
Hi Rob,
I have tried recreating the solution in Pyramid that my BI Office Blog described. However, I have found that creating a crossjoin set, whereby the set contains elements from a hierarchy or attribute that is not the "home" of the set is not supported in Pyramid.
From our Dev team I received the following when I queried the errors I was getting when trying to execute the crossjoin:
In Pyramid, we do not support a custom set whose output does not match that of its parent hierarchy. This means:
- We wont support a set defined on the “Year” hierarchy that is made up of elements from the “Country” hierarchy.
- It also means that we don’t support a set that returns more than a single column in the result.
Given this, the crossjoin function cannot be used at the root of a custom set. It can only ever be used as a means to derive a set.
So it seems we cannot recreate the dynamic eliminations solution you used in BI Office in Pyramid. This limitation is Architectural in nature and will not be changed in the foreseeable future.
Regards,
Ian
-
Ian Macdonald no worries Ian. Cheers for looking into this and getting back to me. I can get the desired result via manually eliminating the unnecessary set combinations, it was just a tad laborious to do so. Shame it’s not part of the scope going forward because I found it fantastically useful during my time at Hays.
Cheers
Rob