In PB we’re noticing behaviour that if the logic in a pipeline isn’t first applied by building from the pipeline then the pipeline won’t run with the new logic when triggered from outside the PB UI, e.g. in data lineage or through schedules. It seems the “deploy” button in PB acts like a commit of the new logic and until this is done the previous logic is used, with the save button acting more like a saved state of the work. Is this a correct assumption?
Would it be possible to have a commit button. In some cases we don’t want to immediately build the pipeline but do want to commit the changes so that the next time the schedule runs it uses the new logic.
Hi!
The save button indeed only saves the code, when deploy deploys it to the pipeline running (so it gets picked up by any schedule running those datasets).
So save ~= commit, deploy ~= deploy to the branch’s prod.
I would recommend to use branches if you do not want to immediately build the pipeline but want to be able to test and build without impacting what’s running on your main branch (and protect it if you want others using it to respect that practice).
