union: Don't eliminate duplicate rows based on matches between map_0's output column and map_1's input column, because union is now being used for self-joins and it is legitimate for a term to appear as both an input and an output
union: Don't eliminate duplicate rows based on matches between map_0's output column and map_1's input column, because union is now being used for self-joins and it is legitimate for a term to appear as both an input and an output