Dimension Consolidation Not Working With Rules

[Kevin Show]

I have a cube that uses Rules.
Deal Number N1153 is the consolidation of two n-elements N130952 and N130177.
The rule pushes the number 976 into both n-elements, so from normal TM1 dimension consolidation the Deal Number N1153 should be the sum of the two n-elements, 976 + 976 = 1953.
However, the consolidation of the dimension does not occur and the consolidation element N1153 is set to one of the n-elements.

N1153________976
___N130952___976
___N130177___976

However, if I do the following to the rule based cube.
1. turn the rules off
2. poke any number into the n-elements
3. then zero the n-elements
4. turn the rules back on
The dimension consolidation works.

N1153_________1,953
___N130952______976
___N130177______976

Has anyone experience this rule problem and is there a fix for it?
I'm using version 9.4 MR1 (Build: 9.4.10000.50077)

Thanks in-advance for any help

[tomok]

Can't help you unless you post your rule code.

[Martin Ryan]

I'd guess one of your N elements (N130952 or N130177) is not being correctly fed.

Either that or you're missing a "N:" as part of your rule, which you've magicked in during the second step.

Martin

[rkaif]

I also think that its a Feeding problem. Post your Rules/Feeders here to get the exact answer

[Kevin Show]

Hi Guys,

Thanks for looking at the problem.
The guys here at work are puzzled and suggested to post the issue to the forum.

A bit of background is that we are re-writing the cube to replace the rules with TIs.
The issue arise when we began to compare the results between the Rule based cube and the TI based cube.

So to simplify the issue we cleared out all the rules and changed the rule to only the following

['NA','OB',!Deal_Number, 'NA Sale', '0', 'RoundingAdjustment'] = N: 1;

No Feeders.
Rounding Adjustment is a measure.

The results is still
N1153 ___________1
___N1390952_____1
___N130177______1

The weird thing it only occurrs at a few cells, such as the above.

When we remove the rule and force a number into the cell, then the dimension consolidation works correctly.
Then we re-instate the rule, the dimension consolidation works correctly.

We have re-boot the server.

[lotsaram]

Kevin,

What you are describing is classic underfeeding. Assuming the cube is not unloaded, once a cell is fed it STAYS FED for the remainder of the server session. A cell can be fed either from a feeder or from underlying data. That is why removing the rule entirely then entering data then replacing the rule (even if the data entry is cleared out) has the effect of feeding the calculation.

If you unload then reload the cube or bounce the server then you will find that your consolidated node in question will have a value of zero not 1.

To make sure rule calculated cells consolidate correctly they must be fed.

[kpk]

No Feeders.

Ok, no feeders, but what about SKIPCHECK? Do you turn on the skipcheck in the first line of the rule?

[Kevin Show]

Thanks for the replies and help,

Issue was corrected by feeding the cells

Cheers
Kevin