One solution that comes to mind:
Uee from the original table two column filter nodes Column Filter — NodePit to create two tables which contain the relevant columns.
In your case
Filter 1: Company, Location, Order Date Short, Qty Short
Filter 2: Company, Location, Order DateLong, Qty Long
Rename the columns in one table 2 to the same names like in table 1 using the
After that make a join using the Concatenate Node https://nodepit.com/node/org.knime.base.node.preproc.append.row.AppendedRowsNodeFactory?numWorkflows=999999 that’s joining on Company and Location.
If required you can filter out “empty” entries on Qty
Thanks for the suggestion. It gave me an idea in which direction to go.
I ended up joining the 4 columns with ColumnExpressions into 2: “order qty short | order date short” and “order qty long | order date long”, so that I would be left with only two columns to deal with. Then with a “Column List Loop Start”, “Loop End” and “Column Combiner” combined them into one column. At the end I split the values with a “Cell Splitter” into two new columns.
