Are they benchmarking hash join on hadoop and redshift?