Hi Richard! Thanks for asking. The answer is: it depends :) I would see some approaches for that:

1. Every developer cleans his mess before you will run the generate-schema command versus DEV and then update to UAT ( so you don't have those tables then)
2. Manually remove those changesets (with those 5 not needed tables) from files generated by generate-schema and then update to UAT
3. You run generate-schema only at the beginning of your journey with SQLcl Liquibase. After that, every developer creates his changesets and commits them to git. Those 5 tables will not be added there, so they will not be deployed to UAT during the update.
4. After running generate-schema on DEV, edit the changesets with those 5 tables and add a tag contexts="DEV_ONLY" ( or any other name ). When deploying to UAT, use liquibase update -changelog-file controller.xml contexts="!DEV_ONLY." Thanks to that, changes with contexts DEV_ONLY will not be deployed.

And a few more approaches, I suppose. Hope it helps. Rafal

Rafal Grzegorczyk Thanks again for the suggestions. I think point #4 would be a great idea to add tags and using it while updating the UAT environment.

Trying to get started on a DEV environment where there are 100s of temp tables created as part of nightly processing every day. :-)