TM1 Forum

Hello,

problem #1:

Rule: ['Viso dienų skaičius per mėnesį' ] = N: ATTRN('Laikotarpiai', !Laikotarpiai, 'Dienų sk.'); (rules takes attribute "Dienų sk." (number of days on month) from time dimension)
Feeder: ['2010.01'] =>['Viso dienų skaičius per mėnesį' ];
Problem: for some reason, feeder is not working for some elements (when I do check feeders on "Katilinė Bukčių g. 58" one of child elements is shown as "not fed"). Restarting a server does not help. Any ideas why some elements are not fed? Dimension hierarchy is OK.

Problem 2#:

Rule (it is in cube "Faktas"):
['Dienų skaičius priode(Energijos)''' ] = N: DAYno(DB('Faktas', !AG, !Laikotarpiai, 'Periodo pabaigos data (Energijos)'));

Feeder #1: ['Periodo pabaigos data (Energijos)'] => ['Dienų skaičius priode(Energijos)'''];
Feeder #2: ['Periodo pabaigos data (Energijos)'] => DB('Faktas', !AG, !Laikotarpiai, 'Dienų skaičius priode(Energijos)''');

Should I use feeder 1 or feeder 2? The accumulation result seems to be the same; will it be the same in all cases? Which feeder would you recommend to use?

Re problem #1:

Let me start by saying that what you are doing with this looks very weird to me. Can you tell us why you need to populate so many nodes in this cube with what is effectively a set of constants?
Secondly, why do you need a feeder in the first place? Consolidating numbers of days seems pretty pointless anyway.

The feeder itself will feed only if there is any value on any element of the dimension containing the element 'Viso dienų skaičius per mėnesį' for '2010.01'. Apparently, there isn't.

Re problem 2#:

Both feeder statements mean exactly the same and there is no practical or theoretical difference between them.
However you can't always use the "[]" syntax; sometimes you are forced to use the DB() syntax, namely when referencing other cubes, nesting functions or applying S rules.

qml wrote:Re problem #1:
Let me start by saying that what you are doing with this looks very weird to me. Can you tell us why you need to populate so many nodes in this cube with what is effectively a set of constants?
Secondly, why do you need a feeder in the first place? Consolidating numbers of days seems pretty pointless anyway.

The feeder itself will feed only if there is any value on any element of the dimension containing the element 'Viso dienų skaičius per mėnesį' for '2010.01'. Apparently, there isn't.

This TM1 project is a direct migration from Cognos Planning, and this measure was used in calculations. I need to consolidate it so it can be a part of non N: level rules.

By saying "any value on any element" you mean entered value, not calculated one, right?

kudzis wrote:By saying "any value on any element" you mean entered value, not calculated one, right?

No, not really. This can be a calculated value as well, as long as it's fed itself. Of course, at the beginning of such a feeder chain there has to be a normal input value somewhere.