Python 3.10 despite the tuple typing annotation there is still

TypeError: Task is not iterable

exception.

@task(timeout=timedelta(seconds=30))
def fetch_symbols(...) -> tuple[MyCustomType, ...]:
  ...

This is Orion right? Or Prefect 1.0?

What are you trying to do downstream? Are you using a for loop or accessing the tuple by index?

Yes, loopin in batches to map other task

You can’t loop inside the Flow because that list is not defined when you are constructing the DAG. It only exists during flow run time. So in order to construct the DAG, you need to use mapping which is similar to looping. For loops are executed during build time which is why they fail. Mapping is executed during runtime

So no logic can be implemented in the flow without extra hoops or wrapping into another tasks?

You can in some extreme cases, but normally no. I would say it’s preferred to put the logic inside tasks but you can’t do stuff like using the native Python

if

,

while

,

for

inside the Flow block to loop over task results

Can I submit tasks dynamically?

Definitely not in Prefect 1.0 because that violates the DAG concept. The only dynamicism available is mapping and task looping, but in Prefect 2.0 we removed the DAG requirement so you can do dynamic task submission

I mean after I get a list of items, I would like to submit them in parallel for other tasks

Yeah that would be the use case of mapping

I am using local run, do I need a special configuration to see actual parallelism?

set

flow.executor = LocalDaskExecutor()

. You can see this