This seems to have changed in FME 2019: as far as I can tell feature caching is now unavailable inside custom transformers.

That's unfortunate. Makes it a bit harder to diagnose custom transformers while developing.

I hope that is reversed.

Consider looking at the existing ideas to see if there's already one that covers this issue, if not consider creating one and let us vote for it :-)

Perhaps this one would be relevant? https://knowledge.safe.com/content/idea/76399/feature-caching-and-custom-transformers.html

I've found that if you want the caches for inside a custom transformer, you need to have that tab open when you run the workspace. Then it works fine. Otherwise, as you noticed, only the main tab is cached.

Edit: But this does appear to be a bug, and not a deliberate change. It's filed with the developers as FMEDESKTOP-8999 and I hope that we'll have a fix soon.

Hi all,

This issue has been in fixed in FME 2020.0. You should now be able to access feature caches in custom transformers again.

You can find the latest release to download here: safe.com/downloads

- Andrea

How about the ability to do partial runs from inside a custom transformer, as suggested in david_r' s link above?