Issuing FTP Commands on MVS and zOS Operating Systems

Like most FTP clients, the browsers in RA FTP Agent Jobs do not support MVS and z/OS file systems. However, you can use the RA FTP Agent Jobs to issue FTP commands on those operating systems, with some limitations as described below. For more information on issuing MVS and z/OS FTP commands, see your operating system documentation.

Partitioned Datasets Represented as Directories

With MVS and z/OS operating systems, partitioned datasets are represented as directories within the graphical user interface. However, a double-click does not display the contents of the partitioned datasets the way it would for directories. You can work around this by typing in fully pathed file names.

You can use Free From Command commands to pass allowable FTP commands to Jobs that run on MVS and z/OS operating systems.

Using Fully Qualified/Partially Qualified Dataset Names

A typical MVS or z/OS set-up defines a prefix for each dataset name. It is normally the user ID of the logged on user. Each partially defined dataset name within an RA FTP Agent Job will be automatically prefixed by the mainframe system with the defined prefix. This can be worked around by fully qualifying the dataset name and enclosing it in single quotation marks.

For example, if a dataset is fully qualified named EXAMPLES.SOURCES.COBOL, then a file name like examples.sources.cobol will be extended by the mainframe system to USERID.EXAMPLES.SOURCES.COBOL. Because won't exist, you will get an error. So the above mentioned dataset must be specified fully qualified in RA FTP Agent Jobs enclosed by single quotation marks. 'EXAMPLES.SOURCES.COBOL' will work in this case.

Accessing a Member of a Partitioned Dataset

Partitioned datasets look like directories in the graphical user interface, but they are not directories. On an MVS or z/OS file system there are no directories. A partitioned dataset contains a certain amount of members that can be accessed like individual files, but a partitioned dataset can never have another partitioned dataset as member. To access a member of a partitioned dataset member the member's name follows the dataset name and must be enclosed in ( ) brackets.

For example, a partitioned dataset EXAMPLES.SOURCES.COBOL might contain members such as COB1, COB2, and COB3. To access the member cob1, the file name should be EXAMPLES.SOURCES.COBOL(COB1).

You can list the members of a partitioned datasets within the graphical user interface by specifying a * character as member name. For example. EXAMPLES.SOURCES.COBOL(*) entered in the path name of the file browser will list all members. All rules about fully and partially names also apply also to partitioned datasets.

Special Behavior

The mkdir command will create a partitioned dataset. You can issue a cd command to go into it and store members in it. Note that the maximum member name size is eight characters. You can't execute a mkdir command in a partitioned dataset. It will raise an error, as it is not allowed to create a partitioned dataset within a partitioned dataset.

You can use mget datasetname(<pattern>) on a partitioned dataset to get all members where pattern can contain a * character like cob*. It will get all members starting with cob in that case. You cannot use other wildcard patters besides * characters such as ? characters.