Hi,
in the manual aggregation I recall that missing columns were marked as such. It appears that if the column name is identical but the type (double vs. int) changes, the columns are not marked as missing. Only when the column name changes they are marked missing.
Here is the test workflow:
PS: It gets even stranger as the GroubBy gets executed despite having an invalig configuration. I Merely copied the GroupBy configured with the int column, reconnected it to the Random Number Assigner and blindlßy executed it.
Best
Mike