TFM_FILL-R processing: Regular vs. Parametric Scenarios

[janis]

Hi Veda Support Team,
I am encountering an issue where the ~TFM_FILL-R table produces different results depending on whether it putted in a Regular Scenario or a Parametric Scenario.
Specifically, in the Parametric Scenario, some expected scenarios are missing from the output, even though they appear correctly when processed as a Regular Scenario.
Attached are the following for reference:
A screenshot of my ~TFM_FILL-R table definition (row for BASE sc. added as without this row data for BASE sc will not be generated. ??).

The output from a Parametric scenario synchronization.

The output generated during a Regular Scenario synchronization (in Green missing data in Parametric scenario).

System Details: Veda 2.0 Version: 4.3.0.2.
Could you please help me understand why the parametric sc. is failing to capture these values and if there are specific rules, I should be applying for FILL-R in this context.
Thank you!

[rohit]

Hi Janis,

Thank you for reporting this issue. To help us investigate the behavior you are seeing and provide a solution as quickly as possible, could you please provide a minimum reproducible case?

If your current model is large, please try to strip it down to the bare minimum of technologies, commodities, and time periods that still demonstrate the problem.

You can upload the file here or provide a WeTransfer link in case you encounter issues uploading the file.

[Antti-L]

@janis : Are you sure it is not just the normal alphabetic issue?
I.e., TFM_FILL gets values only from scenarios that are alphabetically preceding. So, possibly you happened to name your normal scenario alphabetically later than all those four source data scenarios, while the Scen_Par_*  scenario was only after "Feature*".  In my test I indeed saw the fill results depending on the name, as I expected. But I guess one might also think that perhaps the Parscen scenarios could be processed after all normal scenarios.

[AKanudia]

Since parameteric scenarios are processed after the regular scenarios, they should be able to see all the regular scenarios regardless of the name.

But this is probably not tested enough, as very few users have used this powerful feature so far. I am happy to see that it is becoming more popular

[Antti-L]

Thanks, Amit. I confirm that my tests seemed to demonstrate that the name affects the result, using exactly the same FILL-R table in two Parscen scenarios alphabetically differently positioned with respect to the source data in a regular scenario.

[janis]

@janis : Are you sure it is not just the normal alphabetic issue?
I.e., TFM_FILL gets values only from scenarios that are alphabetically preceding. So, possibly you happened to name your normal scenario alphabetically later than all those four source data scenarios, while the Scen_Par_*  scenario was only after "Feature*".  In my test I indeed saw the fill results depending on the name, as I expected. But I guess one might also think that perhaps the Parscen scenarios could be processed after all normal scenarios.

Hi Antti!!!
Base on your test outcomes,  I agree with this. It looks like a classic alphabetical sequencing issue.
Before that (probablly wrong) I’ve noticed that if a scenario name contains a hyphen ('-'), it can cause the scenario to be ignored or fail to process correctly in this context.

[Antti-L]

Thanks for the reply, confirming that your scenario naming would be consistent with the similar issue I am seeing. I would then think that there is probably no need to provide a test case (@Rohit)?  But after Amit's response, I am no longer sure whether this behaviour is what one should expect or not.

[janis]

Thanks for the reply, confirming that your scenario naming would be consistent with the similar issue I am seeing. I would then think that there is probably no need to provide a test case (@Rohit)?  But after Amit's response, I am no longer sure whether this behaviour is what one should expect or not.

Actually I send him case recently.