Dataflow Gen2 Publish vs Run

This is a huge regression from gen 1 dataflows. We should be able to save and publish them without triggering a run.

I find it frustrating that I can't save changes without publishing/running the task. Most of my data processes are scheduled, and if I've already run today's task, running it again causes data duplication. I need the task to run tomorrow, but without saving the changes now, they won't be there for tomorrow.

Completely agree. This also applies to Power Apps Gen2 Dataflows. I can move dataflows to downstream environments via solutions. Generally I will use Power Automate to orchestrate dataflows for the initial data migration into my downstream environments. The issue?

I have to publish each dataflow individually in my target environment before I can run my orchestration Power Automate flow (which essentially just calls each dataflow in the order I want them to run, sometimes with other Power Automate flows running in between for data transformation). If I don't publish (and consequently, run) each dataflow then the orchestration flow fails because it can't run an unpublished dataflow. Publishing also runs each dataflow, so I have to go back through and bulk delete all the data created when I publish the flows. Then I can run my orchestration flow.

Allowing us to publish gen2 dataflows without running them would be a massive improvement, in my opinion.

Also, finding a fix to having to re-establish dataflow connections every time a solution is imported would be huge as well!