Page 1 of 1

More about consolidation and feeders

Posted: Thu Jul 03, 2008 7:23 pm
by rodrigo
Hi guys, i know is a well known question but

The TM1 Developer course manual in the rules chapter says, "dont feed consolidated elements, just leaf".
so my depreciation_period dimension is formed by this tree:

N:NA Projection
C:All Projected Periods
-1
-2
-3
up to 24

My rule is
----------------
SKIPCHECK;
(something)
['Projected Depreciation']=N:IF(NUMBR(!depreciation_period)<=['Months to Project','NA projection'] & ['Flag','NA Projection'] = 1,
['NA Projection','Monthly Depreciation Value'],
continue);
FEEDERS;
(something)
['Months to Project','NA projection']=>['Projected Depreciation','All Projected Periods'];
['Flag','NA projection']=>['Projected Depreciation','All Projected Periods'];
['Monthly Depreciation Value'','NA projection']=>['Projected Depreciation','All Projected Periods'];
----------------
This feeds my rule, but the course text is explicit with C:, should I do a DB with ELISANC? is better? why it is commented the C: feeding advises are always "temptating", C: is feeding the childs? In some cases is valid? sea of questions.. :(

cheers,

Re: More about consolidation and feeders

Posted: Thu Jul 03, 2008 9:19 pm
by David Usherwood
I haven't seen the Developers Course Manual (we wrote our own course) but from what you describe it's leading you a bit astray....

Feeders only _happen_ at N level
A feeder from a C level is _shorthand_ for a feeder from all its N level descendants
A feeder _to_ a C level is _shorthand_ for a feeder to all its N level descendants

So there isn't any real problem about feeding from C level, but the outcome may be to generate (far) too many feeders and blow out memory.The worst (of course) is to feed from a high up C to another high up C as you get a cartesian product of all possible N level pairs. Boom. (Quite often.)

HTH

Re: More about consolidation and feeders

Posted: Fri Jul 04, 2008 8:28 am
by Martin Ryan
Just to expand on David's point as I always found the concept a little confusing till I saw an example.

Say you have a 13th month for balance sheet or some other reason and it pulls values from any of the months. You could write it out long hand like so
['Jan'] => ['13th Month'];
['Feb'] => ['13th Month'];
[...] => ['13th Month'];
['Dec'] => ['13th Month'];

or you can make use of the shorthand David talks about and set up a consolidation 'All Months', which has Jan thru Dec as children and simply write
['All Months'] => ['13th Month'];

As David points out the feeders only happen at the N level, so the effect is identical to the longhand way above. I.e. that one line feeder in fact fires 12 feeders. This is the explosion David talks about.

Similarly the opposite is true you can say
['13th Month'] => ['All Months'];

and this is the equivalent of

['13th Month'] => ['Jan'];
['13th Month'] => ['Feb'];
['13th Month'] => [...];
['13th Month'] => ['Dec'];

Martin

Re: More about consolidation and feeders

Posted: Fri Jul 04, 2008 3:50 pm
by Eric
Just to add my 2 cents not everyone is aware you can setup rules and feeder a few elements in a consolidation.

Code: Select all

[{'Jan','Feb','Dec'}]=>['13 Month'];
Grant it month is a poor example to use for only feeding a pieces of a consolidation, but it does get the point across.

Re: More about consolidation and feeders

Posted: Tue Jul 08, 2008 9:48 am
by jim wood
As far as I am aware feeding from a consolidation is not an issue but feeding to a consolidation can cause performance problems. Thats what I have always been told anyway.

Re: More about consolidation and feeders

Posted: Tue Jul 08, 2008 10:22 am
by David Usherwood
My understanding is that they are both shorthand for feeding all the N level descendants, so there would be no difference - but either could be quite large. Combining both is very likely to give a memory explosion, as each N level source will feed each N level target.

Re: More about consolidation and feeders

Posted: Tue Jul 08, 2008 1:51 pm
by rodrigo
Thanks guys, I am using this with care.