VSAM data sets

You create a VSAM data set by running the Access Methods Services (AMS) utility program IDCAMS in a batch job, or by using the TSO DEFINE command in a TSO session. The DEFINE command specifies to VSAM and z/OS® the VSAM attributes and characteristics of your data set. You can also use it to identify the catalog in which your data set is to be defined.

If required, you can load the data set with data, again using IDCAMS. You use the AMS REPRO command to copy data from an existing data set into the newly created one.

You can also load an empty VSAM data set from a CICS® transaction. You do this by defining the data set to CICS (by allocating the data set to a CICS file), and then writing data to the data set, regardless of its empty state. See Loading empty VSAM data sets.

When you create a data set, you can define a data set name of up to 44 characters. If you choose not to define a name, VSAM assigns the name for you. This name, known as the data set name (or DSNAME), uniquely identifies the data set to your z/OS system.

You can define VSAM data sets accessed by user files under CICS file control as eligible to be backed up while CICS is currently updating these data sets. For more information about backing up VSAM files open for update, see Defining backup while open (BWO) for VSAM files.

VSAM bases and paths

You store data in data sets, and retrieve data from data sets, using application programs that reference the data at the record level.

Depending on the type of data set, you can identify a record for retrieval by its key (a unique value in a predefined field in the record), by its relative byte address, or by its relative record number.

Access to records through these methods of primary identification is known as access through the base.

Sometimes you may need to identify and access your records by a secondary or alternate key. With VSAM, you can build one or more alternate indexes over a single base data set, so that you do not need to keep multiple copies of the same information organized in different ways for different applications. Using this method, you create an alternate index path (or paths), that links the alternate index (or indexes) with the base. You can then use the alternate key to access the records by specifying the path as the data set to be accessed, that is by allocating the path data set to a CICS file.

When you create a path you give it a name of up to 44 characters, in the same way as a base data set. A CICS application program does not need to know whether it is accessing your data through a path or a base; except that it may be necessary to allow for duplicate keys if the alternate index was specified to have non-unique keys.

Reuse of data sets

If you define a data set with the AMS REUSE attribute, it can also be emptied of data during a CICS run. This allows it to be used as a work file. When the status of a file referencing the data set is CLOSED and DISABLED (or UNENABLED), you can use the SET EMPTY command, either from an application program using the EXEC CICS command-level interface, or from a main terminal using the main terminal CEMT command. This command sets an indicator in the installed file definition so that when the file is next opened, the VSAM high-used relative byte address (RBA) is set to zero, and the contents of the data set are effectively cleared.

If you define a data set to VSAM with the average and maximum record lengths equal, and define a file to CICS with fixed length records to reference that data set, the size of the records written to the data set must be of the defined size. For example, if a record in a data set has been read for update, you get errors when rewriting the record if, for example, you:
  • Defined the record sizes to VSAM as 250 bytes, with the parameter RECORDSIZE(250 250)
  • Defined the file to CICS with the parameter RECFORM=FIXED
  • Loaded the data set with records that are only 200 bytes long