Generation Data Group

GDG

Enhanced Definition

A Generation Data Group (GDG) is a collection of chronologically related non-VSAM datasets on z/OS that share a common base name. Its primary purpose is to maintain multiple versions of a dataset, typically output from recurring batch jobs, allowing for easy access to current and historical data while automating the management of older versions.

Key Characteristics

- Chronological Ordering: Datasets within a GDG are ordered by their creation time, identified by a unique generation number and version number (e.g., G0001V00).
- Fixed Limit: A LIMIT parameter, specified during GDG base definition, dictates the maximum number of generations that z/OS will retain in the catalog.
- Automatic Management: When a new generation is created and the LIMIT is reached, z/OS automatically uncatalogues (and optionally scratches) the oldest generation, ensuring controlled data retention.
- Relative Naming: Generations can be referenced using relative numbers in JCL DD statements: (+1) for the next new generation, (0) for the current/most recent, (-1) for the previous, and so on.
- GDG Base Entry: Requires a GDG base entry in the z/OS catalog, which is created using the IDCAMS DEFINE GDG command. This base entry stores the LIMIT and other GDG attributes.
- Non-VSAM Datasets: GDGs are specifically designed to manage sequential (PS) and partitioned (PO) datasets, not VSAM datasets.