I have an FME workbench that writes to multiple Excel files (all contained within a single zip file) with the intent that the generated files will not really be stored, but merely returned as the output of an FME web app and discarded soon after. Having to write these files on-disk first means I have to consider "hygiene" between runs so that there's no risk of file name collision/overwriting, while also saving them in a location that is available for the FME Flow server to expose to the user via HTTP.
None of the documentation I could find really addresses where to save files meant for use with the "data download service", nor how to ensure that file paths cannot overlap between runs. Is there an FME Flow parameter for this purpose, like a unique run identifier I could add to my file names, or a private directory created only for the output of this run? I'm thinking of something like what TempPathnameCreator outputs, but not as a transformer, since it means I can't use it to parametrize the output paths of my writers. Even if it could, the directory is immediately deleted when the translation is complete, so the data download service couldn't use it.
Is this a feature that FME already covers and that I have yet to find, or am I better off using a Python scripted parameter and making some dynamic UUID path pointing to some shared data directory I would need to clean up manually every month?
