I’m a little confused with the migration from v1 to Orion. Should we continue to use .map? Or write async tasks?

2.0 is Dags free so you can use it like how you would use Python. That means that instead of:

from prefect import Flow, task

@task
def add_ten(x):
    return x + 10

with Flow('simple map') as flow:
    mapped_result = add_ten.map([1, 2, 3])

You can simply write this:

from prefect import flow, task

@task
def add_ten(x):
    return x + 10

@flow 
def my_flow():
    result = [add_ten(num) for num in [1, 2, 3])

subtle difference though in that you would now have to call sumbit() to really make those calls concurrent e.g.

result = add_ten.submit(num) for num in [1, 2, 3]

that is true. It is because a lot of Prefect users are not using concurrency and were confused about why the returned result is not a Python object. Since we prioritize incremental adoption, we set the default return value of a task to be a Python object

We’ve also got the

map

operator still, which will run your tasks concurrently. It doesn’t support DFE yet though.