I'm grouping by all columns, and not selecting any aggregation columns (in order to remove duplicate rows). I've selected "Retain Row Order". However, the output appears to renumber the RowIDs in a new, but no particular order.
I'm using KNIME version 2.12.1
the groupby node creates a new data table including new rowids in order to do so the node sorts the input table based on the group columns. To retain the row order a atificial column is added prior sorting and grouping which is used to retain the original order of the rows once the result table has been created. That is why the row id of the result table is not sorted but the rows are in the same order as the input rows. If you need to keep the rowid you will need to extract the original rowid with the RowId node into a separate column prior grouping and then put it back in place again by using the RowId node after the groupby node.