We hold all our subset information in a subset cube, so the cube has }dimension and a measure.
The measures are Subset name, subset MDX, whether it expands above, whether the subset is static or dynamic, included in chore etc.
We then have a process that creates all the subsets from this cube information.
This makes PAW report development much easier, as you just drag on the already created subset.
Doesn't solve your leafs issue but I would recommend going down that route, makes life a lot easier in my opinion.
Original Message:
Sent: Mon March 04, 2024 10:02 AM
From: Wim Gielis
Subject: Leaf level selection
Hi Stuart,
Thanks for confirming. I will come up with a generic process to create such a subset in all hierarchies of all dimensions.
Something like _L0, everywhere the same name for consistency. Or "hierarchy name_L0" to avoid confusion.
------------------------------
Wim Gielis
Original Message:
Sent: Mon March 04, 2024 09:31 AM
From: STUART KING
Subject: Leaf level selection
Hi Wim,
The context menu in the cube viewer cannot modify the current set of members for the hierarchies on the row and column axes. The context menu options can only change with members from the current set are being displayed. Most of the options in the context menu act as a filter against the current set of members. If you need to add or remove members to the current set you need to go into the set editor.
The level options in particular are a bit odd. level000 is the top level. If you select level001 it will show all members from the top most level until that level.
We are planning on revising this context menu at some point in the future to add additional options and make the current options a bit more obvious.
I think the current best option is to make a 'leaves' subset and then users can switch to that subset as needed from the selector tile in the overview area of the cube viewer.
------------------------------
Stuart King
IBM Planning Analytics Offering Manager
Original Message:
Sent: Sun March 03, 2024 06:08 AM
From: Wim Gielis
Subject: Leaf level selection
Hi George,
Thanks for replying !
I created a very simple dimension with 1 total and 3 children (level 0). It gives me in the context menu level000 and level001.
level000 then returns the total (consolidated element) and level001 gives me both the total and 3 leaves.
Not particularly helpful when we want to go to the leaves.
Using the chevron on the dimension tile I see level000, level001, Leaf level.
Leaf level does what it name indicates so that's good. But why not having this option in the context menu and why the difference in interface options ?
Agree 100% with your other comments.
------------------------------
Wim Gielis
Original Message:
Sent: Sun March 03, 2024 03:50 AM
From: George Tonkin
Subject: Leaf level selection
HI Wim,
level005 is probably your leaf level. My gut feel is that due to ragged rollups being possible, a Leaf member could really exist at any level.
What may make more sense is popping up the options like in the set editor when right-clicking the header:

and when right-clicking a member, maybe the same options as we get on the right-hand pane in the set editor:

These could replace Drill Up and Drill Down imo.
------------------------------
George Tonkin
Business Partner
MCI Consultants
Johannesburg
Original Message:
Sent: Sat March 02, 2024 09:35 AM
From: Wim Gielis
Subject: Leaf level selection
Hello all,
When we are browsing cube data, to me an important selection is the leaf level. In the end, cube data is loaded against these elements.
What would be the quickest/most efficient way to select the level 0 elements in a dimension/hierarchy ? Assuming a dimension in the rows for the examples below.
Option 1: The level selection is a useful one. When we use the chevron for the dimension 'selector', we need to click 4 times and scroll a bit to get our result.
Option 2: Also for selections, when we right-click the grey are above the first element in the rows, we can go to "Level options", and surprise, we see levels but not the leaf level. That seems weird to me. Not only the difference in the UI, but also given that it is very useful.
Option 3: Use a dynamic subset. To have such a subset, for every dimension/hierarchy in a TM1 model, we need to write and run a TI process. The name of the subset, ideally, is alphabetically the first in the list. But it appears that the Default subset is listed first (not alphabetically) since a level 0 subset called '_L0' is not listed first. That's okay. Still the need to create subsets for every hierarchy. Granted, with TI it's not terribly difficult.
We can also use the ellipsis next to the Set selection, and enter a named set like: _L0, we need to type it ourselves but that's maybe an alternative too.
What do others use (if efficiency and speed of navigating the data is important to you - for standard tasks I don't have much time to spare) ? Entering the Set Editor, for every dimension where we want level 0, seems too many clicks and mouse gestures. Snap commands are not an option imo.
Why is there no leaf level selection in the context menu of a set ?
Thanks,
Wim
------------------------------
Wim Gielis
------------------------------