CubeProcessFeeders doesnt seem to process

[BigG]

Hi, anyone have issues with CubeProcessFeeders(CubeName); Ti function.

I have a rule in a cube for Forecast Amount to retreive from Forecast consolidated element for the budget version

Code: Select all

skipcheck;
['measure':'Forecast Amount'] =n: IF(ATTRS(Version,'Budget',Active) = Y, DB('cube',!year,!Branch,!GL,ELCOMP(Version,'Forecast',1),!Month,'Forecast Amount'), STET);

Feeders;

['Version': 'Forecast', 'measure':'Forecast Amount'] => ['measure':'Forecast Amount']

This works until I change the active budget version, then the feeders will only work if I change the feeder to

Code: Select all

['Version': 'Forecast', 'measure':'Forecast Amount'] => ['measure':'Forecast Amount', 'specific new active budget']

And then change back to

Code: Select all

['Version': 'Forecast', 'measure':'Forecast Amount'] => ['measure':'Forecast Amount']

and it continues to work... the CubeProcessFeeders(CubeName); Ti function tested in proilog, metadata, data, epilog does not seem to initate the feeder. Does anyone know a possible reason why this is the case?

Thanks in advance

[rozef]

It is because TM1 never liberate memory until TM1 service been restart.
So your feeders are still executed even after have been deleted.

The CubeProcessFeeders(CubeName) just refresh the current feeders in case where you have condition on them which point on variable data.
It won't delete your old feeders. You have to restart TM1.

[David Usherwood]

There is also CubeUnload, which clears feeders. The memory doesn't go back to Windows, but does go back to the elegantly named garbage pool.
Word of warning - I have attempted to write TIs which process sections of a very large cube using CubeUnload. On CX90 = TM1 904, if you executed CubeUnload and then refer to the cube, in the same TI (process or chore) the server crashes. Support accepted that as an issue but I haven't seen any fix. My workround was to restart the server via a .CMD script run by TI. This has worked out well.

[BigG]

thanks for replies. Seems then my feeders are not correct.

Heres is an example of my version dim with Consoldiated elements --> leaf elements:

Forecast --> Active leaf forecast
Budget Current Year ---> Approved Budget current year
Budget Next Year ---> Active Budget in progress
Actual
then I have a whole heap of archived 'versions' sitting with no heirarchy in a single list within the same dim. this list increases and decreases as we continue to archive and purge.

Problem is I have a measure called Forecast, Actual and Budget and I need to retreive data across versions real-time (through rules)

Take one example of feed issue in cube rules. If I feed Actual version into all other versions I need to be specific to every version possible for the 'measure' actual to continue to update.

I know I can overfeed each of the consolidated elements, but this only covers the key active versions. But/So then I would need to create a 'dummy' consolidated element for all other 'archived' versions. Prefer not to do this unless it is the only way to continually feed rules...

Hope this makes sense - any further suggestions would be helpful,