To speed up processing of a sort merge join we attempt to eliminate extra comparisons by detecting if the equi-join column values are distinct. In plans where the sorted values are distinct beyond the equi-join values such that the equi-join column values are non-distinct, then incorrect results can be returned.