Page 1 of 1

Feeder for C level calculation

Posted: Fri May 18, 2018 1:26 am
by macsir
Hi, guys

If I am calculating C level average

Code: Select all

['Average XYZ'] =C: ['ABC'] \ ['EFG'];
and user don't need this average at N level and don't care about the Zero Suppression,
I won't use any feeder for this to gain the max cube performance. ( ['ABC'] , ['EFG'] are all real numbers.)
Is that true?

Re: Feeder for C level calculation

Posted: Fri May 18, 2018 5:35 am
by lotsaram
And if you were to write a feeder then what exactly do you think you would feed or need to feed?

Feeders are an exclusively LEAF LEVEL concept. They exist for one purpose only: to enable consolidation of rule calculated leaf cells. Where no consolidation is required ergo no feeding is required. C level rules therefore by definition never require feeding since if any feeding is to be done it is at leaf level.

In your case the best solution is simply to make "Average XYZ" the parent of "ABC and "EFG".

Re: Feeder for C level calculation

Posted: Sat May 19, 2018 12:34 am
by macsir
Sorry, maybe title is confusing, what I mean is feeder for Average XYZ at N level to enable right consolidation. Since it is just doing average at C level and no zero suppression required, I really don't think feeder for Average XYZ at N level is not needed.
I am aware of making "Average XYZ" the parent of "ABC and "EFG" but that will change the measure dimension structure, which I don't want.
Thanks for advice anyway.