Inserting an element in a dimension at the right index

Sorting element names is difficult when the dimension is set on Automatic order and the sorting is done on an alias. Like Project IDs as principal name, Project Name as alias. You would like to sort on the Name most of the time.

When the ID is not in the same alphabetic order as the name, sorting is flawed. We have a Manual sort order but it seems that the InsertionPoint argument of the function DimensionElementInsert is not respected. (Why not - is it a bug ?). Following the documentation it should work though: https://www.ibm.com/docs/en/planning-analytics/2.0.0?topic=functions-dimensionelementinsert. I see that InsertionPoint only works in a dimension without parents. That's very limited I'm afraid.

The InsertionPoint argument is something that I could really use in my current project: I need to add project names in a dimension, but the principal names are project IDs and the project names go to an alias. When we expand a consolidated element, the elements are shown on an incrementing index basis - always. So an MDX based sorted subset is going crazy...

Therefore, we would like to have the sorting already in the dimension itself, not relying on an MDX looking at an alias.

Think of a custodian that needs to look up an element in a long list of elements, when sorting is off. A name alias can be very different from the corresponding ID - sorting-wise.

If InsertionPoint is empty, we can keep the current behaviour. I like it.

But when InsertionPoint is as advertised in the documentation, an existing element, it should go in the right place.

To make it concrete: I have no single option to sort elements, index-wise, based on ID's such that the Name alias is in the order that I want/need. This is a manual job of changing indexes by dragging elements... I thought about setting up a helper dimensions in a TI process (sorted automatically), inserting elements there, then back in the 'real' dimension. No help.

Other use cases: we now have dimensions with principal names as: 1, 10, 100, 11, 12, 13, ..., 19, 2, 20, 21, 22, 23, etc. Not very pretty IMHO. When we pad with leading 0, Excel gives us issues and/or the source systems cannot find back the element without padded 0 (adding another alias is an option but it should not be needed).

Anything related to the above discussion of dimensions also holds true for hierarchies. If IBM could make inserting elements easier for sorting, that would be greatly appreciated !