DEFINE DATA

DEFINE DATA is the clause in IDCAMS DEFINE CLUSTER that lets you specify the data component: its name and, where supported, data-component-level attributes such as control interval size or volumes. Every VSAM cluster has a data component; when you run DEFINE CLUSTER you can omit the DATA clause and let IDCAMS generate a default data component name and use cluster-level parameters, or you can include the DATA clause to control the data component name and attributes explicitly. This page explains what DEFINE DATA is, when to use it, and what you can specify inside it.

What Is the DATA Clause?

In a DEFINE CLUSTER command, the cluster-level parameters (NAME, INDEXED/NONINDEXED, RECORDSIZE, KEYS, FREESPACE, CYLINDERS, etc.) apply to the cluster as a whole. The data component is created automatically; it is the physical storage that holds the records (or, for LDS, the byte stream). The DATA clause is the part of the command where you can define or qualify that data component: give it a specific catalog name and, in some environments, set attributes that apply only to the data component. Syntax-wise, the DATA clause comes after the closing parenthesis of the cluster parameters and is written as DATA ( ... ) with subparameters inside the parentheses.

When to Use the DATA Clause

You can omit the DATA clause entirely. If you do, IDCAMS still creates the data component and assigns it a default name. On many systems the default is the cluster name with a suffix such as .DATA (e.g. MY.FILE.KSDS.DATA). So the cluster is still usable; you just do not control the data component name or data-level options. You use the DATA clause when: (1) you want the data component to have a specific name (e.g. for naming standards or LISTCAT readability), or (2) you need to specify attributes at the data component level, such as CONTROLINTERVALSIZE or VOLUMES, that you want to apply only to the data component (for example when the index component is defined separately with different CI size or volume).

Syntax of the DATA Clause

The DATA clause follows the cluster parameters and is followed by the INDEX clause for KSDS. General form:

1
2
3
4
5
6
7
8
9
  DEFINE CLUSTER ( -
    NAME(cluster-name) -
    ... cluster parameters ...) -
  DATA ( -
    NAME(data-component-name) -
    [ CONTROLINTERVALSIZE(n) ] -
    [ VOLUMES(volser) ]) -
  INDEX ( ... )   /* KSDS only */
 

The only parameter that is commonly required inside DATA when you use the clause is NAME, if you want an explicit data component name. The other parameters are optional and depend on your IDCAMS implementation. CONTROLINTERVALSIZE sets the CI size for the data component in bytes (e.g. 4096 or 8192). VOLUMES specifies the volume serial(s) where the data component is to be allocated.

Parameters in the DATA Clause

NAME is the catalog entry name for the data component. It must be unique within the catalog and is usually derived from the cluster name (e.g. MY.APPL.KSDS.DATA). CONTROLINTERVALSIZE (or CISZ) sets the size of each control interval in the data component; this affects how many records fit in one I/O and how free space is calculated. If you do not specify it at the data level, the default or cluster-level value is used. VOLUMES in the DATA clause places only the data component on the specified volume(s); the index component (for KSDS) can be on different volumes if specified in the INDEX clause.

Example: DEFINE CLUSTER with DATA and INDEX

A full example that names both the data and index components:

1
2
3
4
5
6
7
8
9
10
  DEFINE CLUSTER ( -
    NAME(MY.APPL.KSDS) -
    INDEXED -
    RECORDSIZE(100 100) -
    KEYS(12 0) -
    FREESPACE(10 5) -
    CYLINDERS(20 5)) -
  DATA (NAME(MY.APPL.KSDS.DATA) CONTROLINTERVALSIZE(4096)) -
  INDEX (NAME(MY.APPL.KSDS.INDEX))
 

Here the cluster is MY.APPL.KSDS. The DATA clause gives the data component the name MY.APPL.KSDS.DATA and sets the data CI size to 4096 bytes. The INDEX clause names the index component. Programs and JCL still reference only the cluster name MY.APPL.KSDS; the data and index component names are used internally by the catalog and the access method.

Data Component vs Cluster-Level Parameters

Most attributes that affect the data component (RECORDSIZE, FREESPACE, space allocation) are specified at the cluster level in DEFINE CLUSTER. They apply to the data component because the data component is where the records and free space live. The DATA clause is for the few attributes that are specifically tied to the data component as a separate catalog object: its name and, where supported, its CI size and volume placement. So you do not repeat RECORDSIZE or FREESPACE inside DATA unless your documentation says so; you set them once at the cluster level.

Key Takeaways

• DEFINE DATA is a clause inside DEFINE CLUSTER that defines or qualifies the data component.
• You can omit the DATA clause; IDCAMS will create the data component with a default name.
• Use the DATA clause to give the data component an explicit NAME or to set CONTROLINTERVALSIZE or VOLUMES at the data level.
• The data component is always part of a cluster; there is no standalone DEFINE DATA command that creates a data-only object.
• Programs and JCL reference the cluster name only; they do not use the data component name.

Explain Like I'm Five

When you order a filing cabinet (cluster), the cabinet has a drawer where the papers (data) go. DEFINE DATA is where you give that drawer its own label (name) or say how big each section of the drawer is (CI size). You don't have to—the system can label it for you—but if you want a specific name or size, you use DEFINE DATA. The drawer is always part of the cabinet; you can't order just a drawer by itself.

Test Your Knowledge