Update statistics in parallel on very large databases

If you have an extremely large database and indexes are fragmented, UPDATE STATISTICS LOW can automatically run statements in parallel.

To enable statements to automatically run in parallel, you must run UPDATE STATISTICS LOW with the PDQ priority set to a value that is between 1 and 10. If the PDQ priority is set to 1, 10 percent of the index fragments are analyzed at one time for the current table. If the PDQ priority is set to 5, 50 percent of the index fragments are analyzed at one time for the current table. If the PDQ priority is set to 10, all fragments are analyzed at one time for the current table. (If the PDQ priority is set to a value that is higher than 10, HCL OneDB™ operates as if the PDQ priority is set to 10, analyzing all fragments at one time for the current table.)

If you run UPDATE STATISTICS MEDIUM or HIGH, you can set the PDQ priority to a value that is higher than 10. Because the UPDATE STATISTICS MEDIUM and HIGH statements perform a large amount of sorting operations, increasing the PDQ priority to a value that is higher than 10 provides additional memory than can improve the speed of the sorting operations.