Description of status

The process of securing control of data sets for a job is called data set integrity processing. Data set integrity processing avoids conflict between two or more jobs that request use of the same data set. For example, two jobs, one named READ and another named MODIFY, both request data set FILE.

READ wants only to read and copy certain records
MODIFY deletes some records and changes other records

If both jobs use FILE concurrently, READ cannot be certain of the integrity of FILE because MODIFY is changing records in the data set. MODIFY should have exclusive control of the data set.

Indicate the type of control needed by coding the data set's status:

   //ddname  DD  DISP=(NEW,...
   //ddname  DD  DISP=(OLD,...
   //ddname  DD  DISP=(MOD,...
   //ddname  DD  DISP=(SHR,...

For exclusive use of a data set, code:

NEW: the data set is being created in this job step.
OLD: the data set existed before this job step.
MOD: the system first assumes that the data set exists. For an existing sequential data set, MOD causes the read/write mechanism to be positioned after the last record in the data set. The read/write mechanism is positioned after the last record each time the data set is opened for output.
If the system cannot find volume information for the data set on the DD statement, in the catalog, or passed with the data set from a previous step, the system assumes that the data set is being created in this job step. For a new data set, MOD causes the read/write mechanism to be positioned at the beginning of the data set.

Note: For a new generation of a generation data group (GDG) data set (where (+n) is greater than 0), VOLUME=REF or VOLUME=SER can be coded.

For shared use of a data set, code:

SHR: the data set existed before this job step and can be read by other concurrent jobs.

Exclusive control of a data set

When a job has exclusive control of a data set, no other job can use that data set until completion of the last step in the job that refers to the data set except when downgrading ENQs. A job should have exclusive control of a data set in order to modify, add, or delete records.

In some cases, you may not need exclusive control of the entire data set. You can request exclusive control of a block of records by coding the DCB, READ, WRITE, and RELEX macro instructions. See z/OS DFSMS Using Data Sets .

Shared control of a data set

Several jobs can concurrently use a data set on a direct access device if they request shared control of the data set. None of the jobs should change the data set in any way.

If more than one step requests a shared data set, code SHR on every DD statement that requests the data set, if it is to be used by concurrently executing jobs.

Examples

   //DD1   DD  DSNAME=PERMDS,DISP=OLD
   //DD2   DD  DSNAME=&&TEMPDS,DISP=NEW
   //DD3   DD  DSNAME=GENDS(+1),DISP=(NEW,CATLG)