HELP! GoupBy Configuration with flow variables

prashant7526 · February 27, 2025, 7:35pm

I am trying to configure the GroupBy node with flow variables.

As can be seen in the screenshot below:

GroupBy Columns:

Added Include List.
Exclude List

In Aggregation Section:

I added the column Name List.
aggregationMethod is Maximize for every columns
inclMissingVals is true for every column
ColumnTypes: I have no idea what I should send.

I am getting the following message:
2025-02-27_15-30-54

I need help; what am I doing wrong, or what data or kind of data does it require?
It would be helpful if anyone could demo it in the workflow with any data.

rfeigel · February 28, 2025, 2:19am

Can you share your current workflow with anonymized data if necessary? That would make it a lot easier for someone to help.

prashant7526 · February 28, 2025, 12:07pm

Problem.knwf (128.2 KB)

mwiegand · February 28, 2025, 12:38pm

Hi @prashant7526,

I had a look on what you want to achieve, either JOIN or PAD. The GroupBy cannot PAD and it is not suited to concatenate two or more columns.

In simple terms, what you want to accomplish is:

Loop per column
Conditionally, based on the column name, either JOIN or PAD

Though, your test data is extremely simple and processing the data via the “good old fashioned way” is easier.

Here is a possible solution for you.

In case a programmatic approach is still required, let me know.

Cheers
Mike

prashant7526 · February 28, 2025, 12:40pm

The below GroupBy is working fine and doing its job. I want the same result but with the use of variables.

I just need to know how to configure upper GroupBy with the use of variables and get the same result. I know it can be done. Just don’t know what the specifics of GroupBy when working with the flow variable tab.

JOIN and PAD have already been achieved till Loop End. I am checking the workflow @mwiegand you’ve provide this might help.

prashant7526 · February 28, 2025, 12:51pm

@mwiegand I checked the flow I want that GroupBy that is working (in the flow you have shared). But I want to configure it with variables and should work in that way.

The whole point here is the below column I cannot hard code; these might be there or will not depending on how the user configures the Logic File. It might be the case there will be new columns, and that is why I want to have it in variables.

Yes, programmatic approach is still required.

mwiegand · February 28, 2025, 1:06pm

Then your challenge is not the processing of the data but ensuring the ingress data is consistent.

Just create a reference table to add the missing columns like so:

Here is the updated workflow: