DSXchange

You should rarely need Force Compile. When you change a job design but make no changes to Transformer stage(s) in that job, when you compile the job the step of generating, compiling and linking C++ code from the Transformer stage is skipped (since nothing changed). Force Compile forces this step to...

There are no assignments. For example

The parentheses are not necessary; I included them only for clarity.

Welcome aboard. Sounds like you need to get some DataStage training.

The dsenv script is on the server, in the DataStage engine directory.

Can you compile other Transformer stages? Can you compile this one without the routine reference?

This error is sometimes seen if the compiler license is expired or exceeded.

So you learn that it DOES do implicit conversion in this case.

Two outputs from Transformer stage, one for GL and one for LCL. Bring them together subsequently with an Aggregator stage.
I think you can figure out the column derivations.
Add a "key" column (derived as @INROWNUM) that you can use to re-unite the original rows.

Bring environment variable APT_DISABLE_COMBINATION into your job as a job parameter and, when you run the job, set it to True.

Please post the exact and entire error message.

I believe that you can only compile a Build stage from within the stage properties dialog.

Get the line terminator character defined correctly.
Or add FILLER PIC X. to the file definition.

Welcome aboard. The generator operator can generate a proportion of null or invalid values, so you could try setting that to 100% for the column in question. Otherwise interpose a Transformer stage.

No, but if you convert all of those to "" and check the length of the remainder you have a solution.
Your constraint expression would have the form

Delete the Data Set from Data Set Management utility then re-run the job. Do the same results obtain?