Use PARALLEL and DIRECT PATH options for data loads. When loading large datasets with SQL*Loader, always use DIRECT=TRUE and PARALLEL=TRUE for maximum throughput. Validate after every load. Run row ...