Hi. Is it correct statement that Prefect 2 map function does not guarantee the order? Example:

l1 = task1.map([1, 2, 3])
l2 = task2.map(l1)

task3.map(l1=l1, l2=l2)

Prefect does not guarantee that in

task3

list

l1

elements will be passed with corresponding

l2

elements?

it does; if you see any issue with ordering here, it would be great if you open an issue with MRE

For ConcurrentTaskRunner as well?

concurrent will try to run tasks concurrently but the submission is in the same order you provided and the results are ordered so e.g. numbers as results from extract will be passed in the right order to transform

from prefect import flow
from prefect_dask import DaskTaskRunner
from flows.parallel.tasks import extract, transform, load


@flow(task_runner=DaskTaskRunner())
def dask_flow_map():
    numbers = extract.submit()
    transformed_numbers = transform.map(numbers)
    load.submit(numbers=transformed_numbers)

regardless of which task was submitted first etc

Hmm, interesting, can't create MRE(not reproducible), but see issues in prod flow

gotcha, hard to fix if we can't have MRE, if you could dig deeper and file a bug report if you find something more specific, we will def look into it 🙌