Partitioning and Table Compression

You can compress several partitions or a complete partitioned heap-organized table. You do this by either defining a complete partitioned table as being compressed, or by defining it on a per-partition level. Partitions without a specific declaration inherit the attribute from the table definition or, if nothing is specified on table level, from the tablespace definition.

The decision whether a partition should be compressed or uncompressed adheres to the same rules as a nonpartitioned table. However, due to partitioning to separate data logically into distinct partitions, such a partitioned table is an ideal candidate for compressing parts of the data (partitions). For example, it is beneficial in all rolling window operations as a intermediate stage before aging out old data. With data segment compression, you can keep more old data online, minimizing the burden of additional storage consumption.

You can also change any existing uncompressed table partition later on, add new compressed and uncompressed partitions, or change the compression attribute as part of any partition maintenance operation that requires data movement, such as MERGE PARTITION, SPLIT PARTITION, or MOVE PARTITION. The partitions can contain data or can be empty.

The access and maintenance of a partially or fully compressed partitioned table are the same as for a fully uncompressed partitioned table. Everything that applies to fully uncompressed partitioned tables is also valid for partially or fully compressed partitioned tables.

This section contains the following topics:

See Also: