Compression and storage enhancements
Db2 11.5.4 contains additional compression and storage enhancements for column-organized tables that continue to make Db2 an industrial-strength database solution that is suitable for any size of organization..
- Db2 on-premises
- The single container deployments of Db2 Warehouse and IBM Integrated Analytics System (IIAS)
- The container micro-service deployment of Db2 on Red Hat OpenShift
Enhancement | Description |
---|---|
Improved compression for string data types |
For string data types, Db2 now has the ability to use page-based compression algorithms in addition to the existing column dictionary-based compression algorithms. Especially during bulk insert or update operations involving columns that contain many unique values, the compression may be greatly improved for the following data types:
There are two page-based compression algorithms and Db2 will automatically choose the algorithm that provides the best compression. The first algorithm targets data that has longer repeated patterns, such as text and URLs. The second algorithm targets data that has 16 or fewer distinct bytes and only short repeating
patterns. Such data typically includes the following when stored as strings:
This feature is disabled by default because it is fallback incompatible, and it must be enabled explicitly. For details, see the DB2_COL_STRING_COMPRESSION registry variable. |
Reduced Synopsis Table Storage for Small Tables |
Synopsis tables for small column-organized tables can have excessive overhead due to Db2's partitioning and storage allocation. The unused allocated storage for synopsis tables can be excessive in comparison to the base tables. The excessive storage consumption by synopsis tables can be avoided without performance penalty by deferring the creation of synopsis tuples until storage consumption overhead can be minimized. This feature is disabled by default because it is fallback incompatible, and it must be enabled explicitly. For details, see the DB2_COL_SYNOPSIS_SETTINGS registry variable. |
Improved fallback compatibility support |
Fallback-incompatible features are the features that, once enabled, will cause the database to be incompatible with any prior release. Because of the fallback impact, those features are now disabled by default to give users more control. Users must make a conscious decision that they will not fallback and explicitly turn on the new features using external registry variables. After upgrading to the latest release, falling back to a prior release remains a supported operation until the explicit enabling of any fallback-incompatible features. Once any fallback-incompatible feature is enabled, fallback compatibility is no longer guaranteed. Falling back to an earlier release will likely require restoring the database from a backup from that earlier release. |