Values from execution plan:
thread 1 - 78 million rows
thread 2 - 756,000 rows
thread 3 - 836,000
thread 4 - 731,000
thread 5 - 182,000
thread 6 - 158,000
thread 7 - 193,000
thread 8 - 301,000
Is there a trick to evenly distributing the rows across the threads? I feel this might decrease the query execution time.