Why does the "Previous Year" time function still return current year data?

govindakarisma · 2025-07-01T11:19:09+00:00

There was an error rendering this rich post.

I would like to understand why, when I enable the Previous Year time function on the cube, the result still shows the current year's data. For example, I only have data for 2024. However, when I activate the function, the values returned are still from 2024 — instead of returning nothing (since 2023 data doesn't exist) or possibly returning 0.

As additional context, the time range in my setup is configured from 2024 to 2045

Find more posts tagged with

Capsules

Data Modeling

Time range

Accepted answers

All comments

Helmut Heimann

Hi @govindakarisma ,

first of all, you've set the block's data to previous year and previous period value, which would make it to be previous previous year (the year before last year).

If you don't see any values, it's probably because you've switched "show all" off.

The next thing is (and this often leads to confustion) the somewhat odd behaviour that applied time-functions are not directly referred to in the layout's header. So, if you have data for 2024 only, apply the PY function to that, you'll still see "2024" in the header (supplemented by any additional header text), which is a bit misleading.

Best,
Helmut

govindakarisma

I'm currently confused by the behavior of the Previous Year time function.

In my setup:

My actual data only exists for the year 2024
I've enabled the Previous Year function on cube “a” (please refer to the screenshots).
I've also enabled “Show all” for both rows and columns.

However, instead of showing data for 2023 in the "Prev Year" column (which should be empty or zero because there's no data for 2023), it’s showing values aligned with 2025.
As you can see in the preview, the data appears under Jan.25 and Feb.25.

This seems to indicate that the time function is shifting values forward rather than backward — which goes against my expectation of what "Previous Year" should do.

For reference, I've attached screenshots of:

The Axes tab (Month as the row)
Data Blocks setup
Time function configuration for cube "a"

Why is the Previous Year function shifting data forward (to 2025), instead of pulling data from the previous year (2023)? Is this a misconfiguration or expected behavior due to how the layout processes time?

Helmut Heimann

Hi @govindakarisma ,

are cubes a and b the same cubes? If not, how is cube a populated?

How does it look if you display block a only, in a second dataview, without the PY function applied?

What is the screen selection, is anything else than the time dimensions selected?

Best,
Helmut

govindakarisma

Yes, cube A and cube B are the same.
I included both in the layout to compare the results:
- Cube A has the Previous Year time function applied
- Cube B shows the same cube without any time function

Here is what it looks like when displaying block A only, without applying the Previous Year (PY) function.
- On the left, the first data view compares block A (with PY function) and block B (without function).
- On the right, the second data view shows block A only, without the PY function applied.

Nothing else is selected on the screen.
The only screen selection applied is for an Active Version in ENT_VERSION_CATEGORY.

Helmut Heimann

Hi @govindakarisma ,

yeah, I see what you mean—but, that's because of the previously mentioned "odd behaviour" regarding the display of the time dimensions when time functions are applied.

The year 2024 is the previous year to 2025, so the values you see in the left column (PY) of the left dataview are "correctly" displayed because the previous year month to Jan.25 actually is Jan.24.

You would see that better, if you set the month in the x-axis, then there would be a column for each month in the year 2025 and that column would be divided into a CY and a PY block

That would look like this

The only way to display the PY values in the previous year month would be to set up a dedicated cube for that previous year where the values of 2025 are pushed to 2024 by a dataflow. But, that would only work for 2025, so that's quite some effort when you want the screen selection to be dynamic—you would have to have separate cubes for each "PY" or the dataflow has to run each time you change the screen's selection. I'd say that's not feasible.

Best,
Helmut