Weird Consolidation Behavior

[Feldmani]

I am having a bit of a brain fart at the moment, so please don't judge, but I have a cube with four dimensions (to simplify), months, department, region, measure, and need some help as I am experiencing some odd behavior.

I have a rule that says [% of Amount] = DB('CubeName',!Dept,!Months,'All Regions','Amount')/DB('CubeName','All Depts',!Months,'All Regions','Amount');

I get the correct percentage in the regions, but then when I look at all regions, instead of calculating the calculation again, it aggregates it...I don't have an N in front of the equal sign, and I have even tried to do a separate N:C: rule, but still it seems to ignore it and aggregate...

Any ideas? I am wondering if this is a version bug at this point, or my brain does. The version I am on is 10.2.20200.55.

[declanr]

Is allow separate n an c rules set in the config?

[tomok]

Where is this rule statement in relation to the others? Perhaps you have a rule earlier in the file that applies to the same area of the cube and that is superceding this one. Have you done a trace calc? Have you tried removing all the other rules to see what effect it has?

[Feldmani]

Allow separate N and C isn't on, nor should it matter in this case, I believe. If it applies the rule of dividing subdivision by total at consolidation, the percentage should be the same, not a roll-up of the below percentages.

I have not tried removing all rules, but I did try to get rid of the rule entirely and just set it to 1, which does effect everything...

[Duncan P]

The way the rule is written it seems that you are calculating the ratio for "all regions" and applying it to every region. Is this what you had intended?

Have you done a trace as suggested by tomok to check that this is the actual statement that is being used for the cell in question?

In order to know whether what you are seeing is inexplicable or not it would be necessary to see the actual values. Is it possible to reproduce this in a form small enough to publish in full on the forum?

[tomok]

Got any STETs anywhere in this rule file? Trying to debug a single rule statement is fruitless without considering the entire rule file given the fact that overlapping rules don't give you any warning message, just grabbing the first rule statement that applies.