Libraries

IBM Z® Table Accelerator tables are processed in memory, but normally they are loaded into memory from disk-resident libraries, to which they may be stored after changes have been applied to the memory-resident version. A IBM Z Table Accelerator library is a specialized dataset initialized in a special format by IBM Z Table Accelerator. Each IBM Z Table Accelerator library contains a directory of all the tables in the library, and the contents and definition for each table.

A company can utilize any number of libraries. To distinguish the libraries, each is given a name by which it may be referenced in an application program. The name used is the DDNAME which allocates the library dataset in the JCL for the application (or, in the case of CICS®, the name of the CICS File definition that defines the library file to CICS). The delivered default library is MAINLIB. Additional libraries are defined through a call to IBM Z Table Accelerator or by using the DKJTEXEC utility.

Libraries are allocated as VSAM, BDAM, or physical sequential (which will be used as if BDAM). Libraries are managed by the batch utility DKJTEXEC.

Generations

A IBM Z Table Accelerator library will hold up to nine generations of each table, as specified in the table definition. For more information see GENERATION (fullword binary).