TM1 Forum

I am investigating PAW and data input in a cube that contains a continuous time dimension.
This investigation is actually broader in that it will pertain to any data input in a cube where we cross 2 hierarchies of the same dimension.

The case is as follows: the dimension month contains 36 level 0 elements (2018-2019-2020 and each time their 12 children).
There is a hierarchy for the years (2018 is the sum of 2018_01, 2018_02, 2018_03, ..., 2018_12, and so on).
There is a hierarchy for the months (01 for January is the sum of 2018_01, 2018_02, 2018_03, and so on).

If we cross both hierarchies, input can be done in a grid, even though we only have 1 dimension:
This works fine. Every cell in the grid (36 cells) is each time the result of only 1 valid underlying cell, even though 36 cells are underneath it: 1 combination is valid.

But by choosing this approach, C-level rules are out of the question. We get into a conflict between data input and a rules-calculation:
This means that the user is restricted to a view at level 0, 36 periods in the rows, or in the columns, but not 3*12 or 12*3.

Likewise, prohibiting data spreading is not possible anymore. Once we do that (it's a server-wide setting, the user cannot do data input any more !


I haven't looked at security yet.

I also find it confusing to see that invalid cells are in fact C-level cells:
The cell in green is a valid cell, open for data entry.
The cells in red are invalid, as the N-level elements they reference, are not equal.
But why would they become consolidated and receive a value of 100 ?

Hi Wim,

There is a hierarchy for the months (01 for January is the sum of 2018_01, 2018_02, 2018_03, and so on).

Did you mean "is the sum of 2018_01, 2019_01, 2020_01, and so on"?

I also find it confusing to see that invalid cells are in fact C-level cells:

Interesting bug, I've managed to reproduce it. BUT: if you split the N: and C: rules, it works fine

Elessar wrote: ↑Sat Sep 12, 2020 12:34 pm Did you mean "is the sum of 2018_01, 2019_01, 2020_01, and so on"?

Yes. To simulate input at 1 month but then through consolidations.