The unpivoting node generates an output table in which the 'ColumnValues' column is of the type 'Non-native [interface org.knime.core.data.DataValue]'. The 'ColumnName' column is a regular StringCell, from the data I would expect the same of the ColumnValues.
What's wrong? Wouldn't complain if other node wouldn't have a problem with the non-native data type.
Nothing is wrong. Columns in KNIME are typed, e.g. int, double, string, date/time, etc. As soon as you have two column that are not compatible or that can't be translated into another without loose in precision, the column aggregation will result in a non-native data type. This for example happens also when you transpose mixed columns, those column are marked with an '?' in the column header. A workaround is to change the column types before or after the node under consideration to string (Rename node).
Thanks, works nicely. Just wasn't too obvious that the rename node can also be used to change the type. Is it worth having a separate node just for obviousness' sake?