Next
Previous
Contents
Since so much of JCL is dedicated to specifying the exact format of
files (called "allocating datasets" in MVS jargon), much of it can be
ignored when converting JCL jobs to something that can run on Unix.
In this project, I found that only the following information needed to
be kept/translated:
- SAS code written "in-line" with the JCL code.
- EXECute statements that start applications with their command-line
options (usually just one per script file).
- filenames and file "nicknames" from DD statements.
- instances of implied concatenation of files. In JCL this occurs
when one assigns several successive files to the same "nickname".
Next
Previous
Contents