TM1 Forum

Hi All,

I have a cube where the rule looks something like below. TM1 version is 10.2
a = b-c; where c is consol elment of d & e
e=a*z;

feeder;
b=>a;
c=>a;
a=>e;

This is not working, showing incorrect values, c ignores e while consolidation, while tracing it says e is not fed.
Can someone pls guide me in writing the feeders? Let me know for more details. Thanks in advance.

Hi Deepa. can I suggest you read the FAQ on asking for help? The critical point here would be to post your real code.

As your query is written you would not do a = b-c in a rule you would do it in a consolidation but I can't tell if that would work for you since I don't know how many dimensions this cuts across.

Cheers,

Steve

Yes, please no pseudocode. This post is a perfect example why (since it's impossible):

a = b - c
c = d + e
e = a * z

a = b - (d + e)
a = b - (d + a * z) <--- nope (while algebraically solvable, not TM1 solvable)

Following the Request for Assistance Guidelines will allow us to spend time looking at your problem instead of chasing dead ends in nonsense pseudocode (or transcription typos, etc).

Hi

You have a circular reference in your example.

Regards

Paul Simon

To remove the circular reference, you can try the following formula for a:

a = ( b - d ) \ ( 1 + z )

Sorry about short description of the issue. I should have asked my doubt directly.
Yes, the formula has the circular ref....Is there any solution for circular ref?
It worked before with skipcheck off and started giving n/a error recently. Not sure how it was working with skipcheck off.
I even tried handling the subtraction in consol instead of rules, that also ended up in n/a error.
Now I am Is there any way like conditional feeders to address circular ref errors?

Circular references are just the formulas/rules have dependency on other variables that also depended on the result of current formula. Confusing, I know. But essentially, it is an infinite calculation loop.

From your example:
To calculate a, you need to know b & c
To calculate c, you need to know d + e
And for e, you need to know a & z

a <- c <- e <- a ........

It shouldn't have anything to do with feeders.

I would suggest adjust your rules and consolidations to get rid any circular references.

Thanks for the reply. I am trying to rewrite the rules now.
one more question, this n.a error shows up only after service restart, otherwise it takes the generic formula. But on trace calculation, it doesn't show the rule which is actually being used. any idea why