When ParallelTableScan is used, individual TableScan operators work together under an Exchange operator to do a sequential table scan in parallel. As each TableScan operator requires rows, it takes the next unprocessed table page and returns rows from the page, one at a time. In this way, pages are divided between the TableScan operators to achieve parallel processing. Regardless of how the pages are distributed among the parallel TableScan operators, all the rows in the table are visited.