PDSE
The modern version of a partitioned dataset, which reclaims space automatically and removes the old size limits.
Also written library extended
PDSE stands for partitioned dataset extended. It does the same job as a classic PDS, a library of named members, while fixing the problems the original design accumulated over the decades.
The most visible improvement is space. In an old PDS, replacing a member leaves the old copy occupying space until someone runs a compress, and running out of directory entries means rebuilding the dataset entirely. A PDSE reclaims space as it goes and grows its directory automatically, so neither problem arises.
It also allows more than one person to update different members at the same time, which matters on a shared development library, and it can store program objects, the modern form of executable code.
Most new libraries at most sites are PDSEs. Plenty of older ones are still classic PDSs, and both are used the same way from ISPF and JCL, so day to day you rarely need to care which you are looking at. The difference surfaces when a library fills up or when a compress appears in a job.
Related terms
- Partitioned datasetA dataset that works like a library or folder, holding many separately named members inside one dataset.
- MemberOne named item inside a partitioned dataset, such as a single program, job or copybook within a library.
- DatasetWhat a mainframe calls a file: a named collection of records with a defined structure, rather than a loose stream of bytes.
- Load moduleThe runnable form of a program, stored in a library, which JCL executes by name.