To benefit from this kind of partitioning, you can specify a different partition type, such as round robin, at a repartition point near the source. Alternatively, to perform some concurrent reading of the source data, you can specify disjoint SQL overrides for each partition.