TM1 Forum

Hello, I have seen situations before that sound similar. I would advise to have a look at settings in the TM1s.cfg file: http://www-01.ibm.com/support/knowledgecenter/SS9RXT_10.2.2/com.ibm.swg.ba.cognos.tm1_inst.10.2.2.doc/c_usesqlfetchscrollextended_1.html Parameter type: optional, dynamic UseSQLFe...

I like Paul's idea though and will probably use that in future. Any extra overhead of the dummy consolidation and maybe some element security to hide it from users and ensuring leaf subsets for processing on every dimension except month I think comes out positively versus the extra overhead of need...

A) Create a consolidation called Chk for Period Data Make it consolidate the 12 periods with weights from random numbers. Then even if you have a scenario like Jan = 2 and Mar = -2, which would give a Total Year of 0, the random weights should still mean that the Chk for Period Data consolidation g...

This is the reason why I prefer to first have a view of level 0 months. Output in the same or a different cube which combinations should be picked up. A flag value of 1 would suffice. Then a second TI process exports the data using CellGetN and AsciiOutput based on a view containing your temporary f...

Is your data source correct?

How can you have duplicate records?
Is your loop over months incorrect?
Data source cube view ?

st2000 wrote:so also non-admin-users can start processes via a button without an own admin license

Sure. If not, budgetting or forecast processes could not use Action buttons in websheets for example.

gtonkin wrote:I guess it is not really returning the index of the element in the subset that is useful

If the function SubsetElementDelete only takes an index instead of an element name, I would be interested in the index number in the subset.

Hello, What is the logic here ? So you put a value on a leaf level city, but they should not add up? Why do you create a consolidation of cities if you don't want to consolidate ? You could change the weights of certain cities in their total equal to 0, so they are not taken into account. Like the w...

Here is the full coding: - export process (Prolog tab code only, no Data tab) - import process (Prolog tab and Data tab, data source is a subset) vDim = 'Fct_Customer'; and SubsetCreateByMDX( 'ElementsForHierarchy', '{Filter( {TM1FilterByLevel( {TM1SubsetAll( [' | vDim | '] )}, 0 )}, [' | vDim | ']....

talsam wrote:. what should I do ?

Have a look at the settings in the data source tab. Delimiters for fields, the comma separator, thousands separator: all very important when loading text data.

That is a lot faster, thank you !

Hello, Wouldn't it be nice to copy a view name (Ctrl-C, C, C, C) and then go to an application folder, do right-mouse button click and choose "Paste as child" ? Drag-'n-drop a view from the bottom of the list of cubes, all the way to the top of the Server Explorer and you know what I mean ...

Thank you Ty, I will have a look at it tomorrow. In fact I found a solution which is pretty easy. Loop over the dimension and use Elisanc to see if we have an ancestor. Collect all these in a subset. Then do a double loop over the subset elements - see if each pair of elements has a parent-child rel...

Sorry, Wim, I see now what you're wanting. Say, given "Coy", you want to see the hierarchies Coy is a member of represented fully from the top to Coy. Hmmmm, tricky... Yes, the final end result being that one must be able to recreate any given element and all ancestors / unique paths afte...

Wim, [tmp_ty].members seems to return the subset you were looking for originally by the way. Hello Ty, Can you apply Members on a subset ? Because I am after a function or functions that can show the hierarchical representation of ALL ancestors of a given element. Even if the element and certain an...

Since you can't do that in TI, in conjunction with ASCIIOutout, then you'll have to write a bunch of loops inside loops. This is one case where java extensions could provide useful as recursive functions and calls to AsciiOutput are trivial in java. Good point ! It's not part of my knowledge yet so...

Thank you very much Tom. That should indeed work with nested loops but I wanted to avoid hardcoding the number of loops and spelling out the loops as such. I was actually more after a solution without further possibility to run into an issue later on. And I think I can do it using loops in (clever M...

Thank you Tom, I will experiment with that code and in parallel I will write my own MDX/TI routines.

Wim

That is possible yes, although in the customer dimension example, all weights are 1. But the tricky part is that consolidations can occur with the same name in different parallel hierarchies. Their names are not unique. Then asking for the parents "skips" certain instances of the elements/...