Page 1 of 1

Duplicate element in dimension

Posted: Tue Jun 26, 2012 8:14 pm
by Michel Zijlema
Hi,

I had my share of TM1 fun today. I managed to insert a number of elements 'more or less' twice in the same dimension today (TM1 9.5.2 - see screenshot). :shock:
ambiguous_element.jpg
ambiguous_element.jpg (37.2 KiB) Viewed 2551 times
In the example of the screenshot the principal name of the element originally in is 'IT1:03001000', with '03001000' as the alias value (dimension index# 2). After this the element '03001000' was added to the dimension (dimension index# 576) by a TI process. The strange thing was that the process was testing for existence of the element (alias) using DimIx on '03001000', which resulted in 0 and the DimensionElementPrincipalName function returned '03001000' instead of 'IT1:03001000'. Because of this the 2nd '03001000' label was added as a new element... (the element is an example, this happened for multipe similar combinations).
I didn't have the time to dive into this (as I already lost time because of this strange behaviour), but one thing that could be related is that the developer named the alias 'ALIAS'. Maybe this is causing the strange behaviour. :?:

Michel

Re: Duplicate element in dimension

Posted: Tue Jun 26, 2012 8:37 pm
by declanr
That certainly is a good one!

A couple of months ago I found that I had managed to create "duplicate" elements when I had a rule generating the alias in the ElementAttributes cube...
annoying part was that the only way I found out what was happening was that when a websheet (using the principle names) starting to swap values in between 2 columns... the duplicated elements were in the rows not the columns!

Needless to say due to the fact I was focussing on the dimension creating the columns it took me a significant amount of time to find the fault.
Point being that you should be on the look out for unusual behaviour and not in the normal "ripping your hair out" TM1 sense but in the monumental "press that big red , detonate button" sense.


And to add to your specific issue; I know this will seem overkill and a lot of extra processing to some but I tend to always convert everything back to Principle name and do my calculations against that, I can't quite remember why I started this but I think it was something along the lines of TM1 being a bit "buggy" (the usual rip your hair out sense) and in a similar way to yours didn't pick up the dimix when using the alias.