HELP! GoupBy Configuration with flow variables

KNIME Analytics Platform
1

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

As can be seen in the screenshot below:

2025-02-27_15-30-17
2025-02-27_15-30-17785Ă—667 25.3 KB

GroupBy Columns:

  1. Added Include List.
  2. Exclude List

In Aggregation Section:

  1. I added the column Name List.
  2. aggregationMethod is Maximize for every columns
  3. inclMissingVals is true for every column
  4. 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.

2

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

3

Problem.knwf (128.2 KB)

4

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:

  1. Loop per column
  2. 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.

image
image1029Ă—415 20.4 KB

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

Cheers
Mike

1 Like
5

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.

6

@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.

hlTscRUaBG
hlTscRUaBG729Ă—439 9.21 KB

Yes, programmatic approach is still required.

7

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:

image
image1301Ă—468 29.7 KB

Here is the updated workflow:

8

I have placed the annotation with comment.

XNMGvOj7bM
XNMGvOj7bM1468Ă—890 57.2 KB

This logic needs to be dynamic that is working. @rfeigel solved this one.
The only thing left is the more logic it will have the rows it will make. Because I don’t know columns from Logic file. I cannot put it directly in GroupBy. So I have to filter out which columns are added and dynamically pass in the group by with the use of variable to make multiple rows into one row.

9

I may be dense, but I still don’t understand your issue. Apparently the structure of the logic file is indeterminant. Could you provide several examples of what the logic file might look like?

10

Hi @mwiegand and @rfeigel,

I see the same issue was also reported. Was it solved, or did the post just get closed by default?

Driving GroupBy with flow variables