Since you cannot directly enter a DD name into the Connect ETL Source File dialog, you will need to use environment variables for
- the source file(s) of any task that reads from the data sets specified in the JCL, and
- the target file of any task that writes to a data set specified in the JCL.
Names of files that are used only to pass data from one task to another do not need to be environment variables. For optimal performance, intermediate files should be changed to “direct data flows” in the Connect ETL Job Editor.
To use an environment variable as a file name simply dereference the variable in the “File name” box in the source file dialog of the Task Editor. You can dereference variables by preceding it with a dollar sign (UNIX syntax) or by enclosing it in percent signs (Windows syntax).
DMXMFJCL currently checks for the following DD names and sets the corresponding environment variable:
| DD Name | Environment Variable |
| SORTIN | SORTIN |
| SORTJNF1 | SORTJNF1 |
| SORTJNF2 | SORTJNF2 |
| SORTINnn | SORTINnn |
| SORTINn | SORTINn |
| SORTOUT | SORTOUT |
| SORTOFxx | SORTOFxx |
| SORTOFx | SORTOFx |
For merge tasks, DD names must have unique numbers. If the JCL DD statements contain duplicates (e.g. SORTIN7 and SORTIN07), DMXMFJCL will use only the SORTINnn version. See the annotated JCL snippet below for an example of how a merge task will use DD names.