13.3.6 Notes on table row partitioning

  1. When a table is row-partitioned, its indexes must also be row-partitioned. If a table is partitioned and stored in multiple RDAREAs while an index is stored in a single user RDAREA, the level of concurrent execution may be reduced due to locking on the index. For details about index row partitioning, see 14.3 Index row partitioning.
  2. You should use the hash facility for hash row partitioning in the following cases:
    • Hash row partitioning is to be used with the table.
    • The amount of data is expected to increase.
    If you add RDAREAs to handle an increase of data for a hash row partitioning table (if you increase the number of table row partitions), data may become uneven among the existing RDAREAs and newly added RDAREAs. The rebalancing facility for hash row partitioning can correct such unevenness of data when you increase the number of table row partitions. For details about the rebalancing facility for hash row partitioning, see the manual HiRDB Version 8 System Operation Guide.