Partitions and Parallel Operations

Both partition and parallel operations enhance data reorganization operations and improve database performance. For example, partitioning can be used to support large historical tables and their indexes by dividing historical data into time-related partitions that can be managed independently and added or dropped conveniently. This reduces down time due to scheduled maintenance or media failure since activities on one partition do not impact other partitions. Furthermore, data can be laid out across multiple partitions to reduce I/O bottle neck and enhance performance.

Use of the parallel option during data reorganization also improves performance by performing operations concurrently. This is especially useful when multiple partitions are involved, for example, rebuilding indexes in multiple partitions. For example, a single large index rebuild might take ten hours. If the same index is partitioned into ten parts, each partitioned index can be rebuilt in about an hour. And since each individual partitioned index can be rebuilt independently and in parallel, a large partitioned index can be rebuilt in minutes.