Batch input to the binder consists of the primary input data set
and additional data sets. You define the primary input data set using
job control statements. You can include more modules by specifying
additional control statements and by directing the binder to use call
libraries.
Input data sets can contain control statements, object modules
of any type, load modules and program objects. The following table
shows the data set types in which data can reside.
|
Sequential data set |
PDS member |
PDSE member |
z/OS UNIX file |
Control Statements |
X |
X |
X |
X |
Object Modules (all types) |
X |
X |
X |
X |
Load Modules |
|
X |
|
|
Program Objects |
|
|
X |
X |
A single library member can contain only one program object or
load module, but any number of control statements and object modules
in combination.
z/OS UNIX files
can contain binder input of all types except load modules. You
specify z/OS® UNIX either by coding the PATH parameter on
your JCL or by providing the path name on the INCLUDE or AUTOCALL
control statements. See Binder control statement reference for more information. Where
sequential processing or archive file access is required, you must
include the full file name on the PATH parameter; otherwise, code
only the directory name for PATH, omitting the last level of qualification
(file name). The file name will be supplied by the binder, either
from the INCLUDE statement or from the unresolved reference during
autocall.
In addition to the data set type, you must consider how the binder
will access the data set. Sequential access requires that a physical
sequential data set be specified or that a member name be specified
with the library dsname. Partitioned access requires that a partitioned
data set, PDSE,
z/OS UNIX archive
file, or
z/OS UNIX directory
be specified without an associated member or file name. Access requirements
depend on the time that the input is required:
- Primary input is accessed sequentially. Any library in the concatenation
must include a member name with the dsname or path.
- Secondary (included) input can be either sequential or partitioned.
If partitioned, the member name(s) must be specified on the INCLUDE
control statement.
- Autocalled input must be partitioned.
The binder supports mixed concatenations of the above, with the
following exceptions:
- You must not mix data set types in a single concatenation. All
concatenated data sets must be either partitioned or sequential, not
both. A PDS or PDSE member is treated as a sequential data set
- The binder does not support z/OS UNIX files concatenated
with other z/OS UNIX files
or data sets of any type.
Note: This topic refers to binder processing and input. These concepts
apply equally to linkage editor and batch loader processing unless
noted otherwise in
Using the linkage editor and batch loader. The linkage editor
and batch loader cannot process program objects, extended object modules,
GOFF modules, s or z/OS UNIX files.