Duplicate Alias - any way around it?

[nmcdermaid]

Hi,

Is there any way to get around not being able to have duplicate aliases? I can't use the actual element name because its a hideous concatenation of all the parents to make it unique.

Is there any way around not also being able to make aliases unique?

I hear that you can make text item non unique but is there a way to make them visible to Report Studio/Analysis Studio? From my understanding you can only make aliases visible to Analysis/Report studio.

[dusherwo]

Short answer: No.
Longer answer:
In TM1 (as with every OLAP tool I have ever worked with), dimension elements are unique references to the data they address. Aliases are alternative references - this is why they must be unique. If they weren't, how would you know what cell you were reading or writing?
You haven't described the content and purpose of your dimension. This might help.

[nmcdermaid]

Here's an example.

We need to track employees over different work groups.

To uniquely identify where the employee worked in a work group we have a composite key made up of the employee key and the workgroup key. That is what is fed for the element.

The consolidation level above this will be just the work group key.

At the employee level I just want to show the employee name. But if this employee worked in two work groups (which they do over time) they will appear twice. The key will be unique (work group +employee) but the employee name is not.


I'll be honest and say that I'm coming from a transformer background where this kind of thing is the norm. I understand that things are probably done differently in the world of TM1, so I'm willing to keep an open mind.


I think I have misunderstood the purpose of an alias - its not meant to be the 'business description' against a unique key - its just another set of unique keys.

If I can plug a non unique business description against a unique key in a text attribute, and that attribute can be seen in C8 tools then thats fine by me. If I understand correctly this can't be seen in Analysis studio so it may be limited for me.


To put it another way I'm used to building dimensions with a unique key at the leaf, then applying a fact with that unique key to it to load the values.

I hope that makes sense.

I think when I've worked with TM1 a bit more I'll get it

[OLAPBPMguy]

You can just hold a simple text attribute rather than an alias to identify the employee only.  This would work fine for reporting in a Websheet but you might not be able to get the desired layout in C8.

Another alternative to this problem is to have a separate employee and workgroup dimension in the cube where the intersection in the cube where the data is identifies the valid workgroup/employee combinations.  This is often preferable to the situation you have at the moment where you need to construct multiple hierarchies with fully qualified leaf elements to identify the employee in the correct context.  This would be the more standard design in a TM1 cube. Remember in TM1 sparsity doesn't really hurt your model.

[nmcdermaid]

Thanks for your reply, I'm think I'm starting to understand the TM1 viewpoint on things.