Hi all,
I’m currently working on a MVP to build a Financial Model in KNIME. This includes some interactivity in form of a data app that amongst other things allows the user to tweak aggregation levels of the data that was generated.
This means I’m configuring column selections for pivot / unpivot nodes via flow variables.
I came across the following weird behavior of the unpivot node:
Here is some example Input data:
Previously User Selected “Date” as Dimension so now I need to unpivot this table. This user selection is stored in this flow variable:
In Unpivot Node Flow Variable Config dialogue I have set this variable for included_names in retained_columns:
and as the opposite for value_columns
If I then inspect the config dialogue the columns seem to be in the right places:
However when executing the node “Date” is visible in the ColumnNames column:
As a comparison this is Unpivot Node with manual selection:
And the table upon execution, which does not have “Date” as value in ColumnNames:
For me this is quite confusion especially as the config dialogue of both version looks the same…
I have produced an example WF that reproduces this behavior (Instead of “Date” Column this works with “Year-Q” Column:
UnpivotPotentialBug.knwf (98.1 KB)