Output data set groups

An output group can point to a data set group that defines data sets that have identical characteristics and contain the output of a job, a started task, or a time-sharing user. Each data set group is a processing entity with a set of processing characteristics.

JES2 groups output for jobs processed at the local node use the following processing characteristics:

Burst specification (BURST)
Output class (CLASS)
Destination (DEST)
Forms control buffer name (FCB)
Forms overlay frame (FLASH)
Forms (FORM)
Group name (GROUPID)
Output disposition (OUTDISP)
Process mode (PRMODE)
Print train (UCS)
Installation writer name (WRITER)

You can use the OUTPUT JCL statement to describe these characteristics. Note that the SYSOUT DD statement and the /* OUTPUT JECL statement support all these characteristics except GROUPID and OUTDISP.

JES2 also groups output data sets on the following criteria which a user can specify on the OUTPUT JCL statement:

Address
Building
Department (DEPT)
Name
Print complete notification message (NOTIFY)
Room
Title

These fields can also be modified through SDSF while output resides on the output or hold queue.

When JES2 groups on the information on these fields, it attempts to match the character string exactly. Therefore, it is important to ensure the user specifies the exact string in each of the above so that JES2 can group the output properly. For example, JES2 will not group NAME='JOHN' and NAME='JOHN ' together.

You can change the SYSOUT characteristics that affect grouping by using JES2 Exit 40. See z/OS JES2 Installation Exits for more information about Exit 40.

The following tables illustrate how one output group can represent one of several SYSOUT data sets:

Table 1. Relationship of SYSOUT specification to number of job output elements: Seven output groups built
SEVEN OUTPUT GROUPS BUILT
1	-	SYSOUT=(A,,4PLY), UCS=PN
2	-	SYSOUT=C,UCS=PN
3	-	SYSOUT=A
4	-	SYSOUT=B
5	-	SYSOUT=A,FLASH=FL01
5	-	SYSOUT=A,FLASH=FL01,MODIFY=CM03
6	-	SYSOUT=C,FCB=STD3,UCS=GF10
6	-	SYSOUT=C,CHARS=(GF10,FM10),FCB=STD3
7	-	SYSOUT=C,CHARS=(GF10,FM10),FCB=STD3,UCS=PN
3	-	SYSOUT=A

Table 2. Relationship of SYSOUT specification to number of job output elements: Three output groups built
THREE OUTPUT GROUPS BUILT
1	-	SYSOUT=A
1	-	SYSOUT=A
...	...	...
2	-	SYSOUT=B
1	-	SYSOUT=A
2 ⋮	-	SYSOUT=B
3 ⋮	-	SYSOUT=(A,,2PLY),FCB=8LPI
1	-	SYSOUT=A

JES2 duplicates output groups built for job-related output depending on the number of job copies requested on the /*JOBPARM statement (up to the maximum allowed by the COPIES= parameter on the OUTDEF statement). This allows the devices available for output to govern the number of copies processed for a particular job. JES2 uses several factors to group SYSOUT data sets. The groups also conform to a set of rules, discussed below. Table 3 summarizes these rules.

Note: If the output groups are transmitted through an NJE network, do not use the /*JOBPARM JECL statement to produce multiple copies. Instead, use the COPIES= parameter at the DD level. If COPIES= is used at the DD level, JES2 does not duplicate output groups and the output groups are handled appropriately.

JES2 uses the following factors to group output:

The data set output class matching the job's message class
The USERSET specification on the OUTDEF statement.
The user assigned an output group name using the GROUPID= keyword on the JCL output statement.
The DMNDSET parameter specification on the OUTDEF statement.

JES2 forms output groups whose data sets always agree on the following characteristics:

address
building
department
destination
external writer
name
output class
output disposition
output group name
print complete notification message
process mode
room
title

If DMNDSET=YES on the OUTDEF initialization statement and the data set matches the job's message class or if USERSET=YES on the OUTDEF initialization statement and the user assigned the output group a name, then JES2 scans the setup characteristics of the output group's data sets to determine which, if any, of the setup characteristics vary. Otherwise, the output groups will also agree on setup characteristics.

Table 3. Interrelationship Between the Four Factors Affecting Output Grouping
DMNDSET= *	DOES MSGCLASS AND OUTPUT CLASS MATCH?	USERSET=YES		USERSET=NO
		OUTPUT GROUP NAME ASSIGNED		OUTPUT GROUP NAME ASSIGNED
		YES	NO	YES	NO
YES	YES	1	1	1	1
YES	NO	1	2	2	2
NO	YES	1	2	2	2
NO	NO	1	2	2	2
* USERSET and DMNDSET are parameters on the OUTDEF initialization statement. Note: Data sets within groups formed agree on class, external writer, destination, process mode, output disposition, and output group name. However, setup characteristics (those characteristics that determine the data set groups eligible for processing a given device) can vary. Data sets within groups formed agree on class, external writer, destination, process mode, output disposition, output group name and setup characteristics.

To delay printing data sets, an installation can specify:

Reserved classes
HOLD=YES on a JCL DD statement
OUTDISP=HOLD on the JCL Output statement
OUTDisp=HOLD or LEAVE on one of the following initialization statements:
- JOBCLASS
- OFF(n).SR
- OFF(n).ST
- OUTCLASS

Output disposition for SYSOUT data sets discusses the OUTDISP parameter. z/OS MVS JCL Reference explains HOLD=YES on a JCL DD statement. Note that the use of the GROUPID= keyword on the JCL OUTPUT statement influences data set gathering. If your installation allows user control of output groups by specifying USERSET=YES on the OUTDEF statement, you can form output groups with dissimilar output characteristics.

JES2 does not group spin data sets; each data set constitutes a “group” of one data set.