| 摘要 |
Performance of constraint checking is enhanced by dynamically determining an optimal process to perform the checking operation. If additional dataspace can be allocated, the parent index values are stored in the dataspace and a binary search is performed between the extracted foreign keys and the dataspace, thereby eliminating the need for a sort operation on extracted foreign keys to perform constraint checking. If all of the parent indexes needed for constraint enforcement cannot be stored in dataspace, then as many parent indexes as possible are stored in dataspace and multiple reader and multiple sort tasks are utilized to process the remaining parent indexes. Depending on the availability of resources and the characteristics of the database table, determined dynamically, complete or partial optimization of the constraint checking operation is accomplished.
|
