For prefect mapping, is it possible to revert back to the old 'breath first' behavior?

Hi Bob, while there isn’t a first-class API for toggling this behavior, you can enforce breadth first by adding a “dummy” task at each level of your mapped pipeline:

level_one = task_one.map(some_list)
one_wait = Task(name="Level 1 Block")(upstream_tasks=[level_one])

level_two = task_two.map(some_other_list, upstream_tasks=[one_wait])
Task(name="Level 2 Block")(upstream_tasks=[level_two])

etc.

cool, thanks Chris. Will be great to expose this as a parameter option in the future

An even simpler approach seems to be working for me. Am I overlooking something?

level_one = task_one.map(some_list)

level_two = task_two.map(some_other_list, upstream_tasks=[level_one])

Ah that’s great! The middle task is unnecessary - you can set the dependency directly, nice