DSNTYPE Parameter

Purpose

The DSNTYPE (Dataset Type) parameter specifies the type or structure of a dataset being allocated. It defines the physical and logical organization of data, affecting how the dataset can be used, its performance characteristics, and its maximum size limits.

Where DSNTYPE is Used

• In DD statements for dataset allocation
• In TSO ALLOCATE commands
• In IDCAMS DEFINE commands (indirectly)
• In dynamic allocation services (SVC 99)
• Within SMS data class definitions

Basic Syntax

1
2
3
//ddname DD DSN=dataset.name,
//        DSNTYPE=value,
//        ...other parameters...

Where value is one of the supported dataset type values.

DSNTYPE Values

Additional DSNTYPE Specifications

Beginning with z/OS 1.13, DSNTYPE can be specified with additional attributes using the following format:

1
DSNTYPE=(type,format)

DSNTYPE and Other Parameters Interaction

Usage Examples

Basic Sequential Dataset

1
2
3
4
5
//SEQFILE  DD DSN=USER.BASIC.FILE,
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(TRK,(10,5)),
//            DCB=(RECFM=FB,LRECL=80,BLKSIZE=27920),
//            DSNTYPE=BASIC

This example creates a standard sequential dataset with traditional format.

Large Format Dataset

1
2
3
4
5
//LARGEFILE DD DSN=USER.LARGE.FILE,
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(TRK,(100000,10000)),
//            DCB=(RECFM=FB,LRECL=80,BLKSIZE=27920),
//            DSNTYPE=LARGE

This example creates a large format sequential dataset that can exceed 65,535 tracks per volume.

Program Library (PDSE)

1
2
3
4
5
//LOADLIB  DD DSN=USER.PDSE.LOADLIB,
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(CYL,(10,5)),
//            DCB=(RECFM=U,BLKSIZE=32760),
//            DSNTYPE=(LIBRARY,1)

This example creates a PDSE library for program objects (load modules) with format 1 specification.

Data Library (PDSE)

1
2
3
4
5
//SRCLIB   DD DSN=USER.PDSE.SOURCE,
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(CYL,(5,2)),
//            DCB=(RECFM=FB,LRECL=80,BLKSIZE=27920),
//            DSNTYPE=(LIBRARY,2)

This example creates a PDSE library for data members (like source code) with format 2 specification.

Traditional PDS

1
2
3
4
5
//PDSFILE  DD DSN=USER.PDS.DATA,
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(CYL,(5,2,50)),
//            DCB=(RECFM=FB,LRECL=80,BLKSIZE=27920),
//            DSNTYPE=PDS

This example creates a traditional partitioned dataset with 50 directory blocks.

Extended Format Dataset

1
2
3
4
5
6
//EXTFILE  DD DSN=USER.EXTENDED.DATA,
//            DISP=(NEW,CATLG,DELETE),
//            STORCLAS=EXTENDED,
//            SPACE=(CYL,(100,50)),
//            DCB=(RECFM=FB,LRECL=1024,BLKSIZE=27648),
//            DSNTYPE=EXTREQ

This example creates an extended format dataset, requiring the allocation to fail if extended format cannot be provided.

Key Characteristics by DSNTYPE

Basic vs. Large vs. Extended Formats

PDS vs. LIBRARY (PDSE) Comparison

Technical Considerations

Extended Format Benefits

• Compression: Data compression can reduce storage requirements
• Striping: Data can be spread across multiple volumes for improved I/O performance
• Extended Addressability: Supports datasets larger than 4GB
• System-Managed Buffering: Enhanced buffer management for VSAM datasets
• Extended Format CI Size: Control interval size can be larger than 32K

PDSE Benefits

• No Compression Needed: PDSEs reclaim space automatically when members are deleted
• Flexible Directory: Directory expands dynamically as needed
• Member Generations: Can maintain multiple versions of program objects
• Integrity: Better recovery from system failures
• Sharing: Enhanced member-level sharing for improved concurrency

Common Issues and Troubleshooting

Best Practices

1. Use LIBRARY Instead of PDS: For most new library allocations, prefer PDSE (LIBRARY) over traditional PDS to avoid compression requirements and directory limitations.
2. Format Selection for Program Libraries: Use DSNTYPE=(LIBRARY,1) for load module libraries and DSNTYPE=(LIBRARY,2) for data libraries.
3. Large Datasets: Use LARGE or EXTREQ for datasets that might exceed the 65,535 tracks per volume limit.
4. Fallback Strategy: When extended format is desired but not critical, use EXTPREF instead of EXTREQ to allow fallback to basic format.
5. SMS Integration: Coordinate DSNTYPE with appropriate STORCLAS and DATACLAS settings for optimal performance and functionality.
6. Compression Consideration: Use extended format (EXTREQ) with COMPACTION=YES for datasets with high compressibility to save space.
7. Performance Optimization: Use extended format with striping for high-performance sequential processing of large files.
8. Compatibility Awareness: Be aware of older applications that might not support newer dataset types like PDSE or extended format.

JES2 vs JES3 Considerations

DSNTYPE is primarily a data management parameter rather than a job entry subsystem parameter, so it functions identically in both JES2 and JES3 environments. However, there are a few system-level considerations:

Related Concepts

• DSORG Parameter - Dataset organization, which defines the logical structure
• SPACE Parameter - Space allocation for datasets
• DCB Parameter - Data Control Block parameters for dataset characteristics
• RECFM Parameter - Record format specification
• STORCLAS Parameter - Storage class for SMS-managed datasets
• DATACLAS Parameter - Data class for SMS-managed datasets