Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

code:
```@task(name = 'add')
def func_1(param_1, param_2):
    return param_1 + param_2

@task(name = 'minus')
def func_2(param_1, param_2):
    return param_2 - param_1

with Flow('flow', executor=LocalDaskExecutor(scheduler="processes", num_workers=2)) as flow:
    param_1 = Parameter("param_1")
    param_2 = Parameter("param_2")
    task_1 = func_1.map(
        param_1 = param_1, 
        param_2 = unmapped(param_2)
        )
    task_2 = func_2.map(
        param_1 = param_1, 
        param_2 = unmapped(param_2),
        task_args=unmapped(dict(trigger=all_finished))
        )

    task_2.set_upstream(task_1, mapped=True)
   

if __name__ == "__main__":
    flow.run(parameters = dict(param_1 = [1,2,3,4], param_2 = 5 ))```
And I have this error:
```  File "/Users/acsjhba/opt/anaconda3/envs/py39/lib/python3.9/site-packages/prefect/core/task.py", line 810, in map
    task_args = task_args.copy() if task_args else {}
AttributeError: 'unmapped' object has no attribute 'copy'```

this behaviour stops me from setting *trigger* between upstream and downstream mapped tasks….

How can it be solved, thanks!

I know for single task, I can set `@task(name = 'minus', trigger = all_finished)` in the task decorator, but how can I do that in flow of flows, specifically for *create_flow_run*

is it solved now? you deleted the original message

nope, it is not solved..
but I have an idea, instead of specifying the task_args at create_flow_run level, we specify it in its child tasks level (in that task decorator), thus in create_flow_run, do you think that will work?

And also, another problem is upstream_tasks = [xxx] cannot be wrapped in unmapped function.. So I may need to set something like `downtstream_flow_id.set_upstream(upstream_flow, mapped=True)` , correct?

can you perhaps try solving it in 2.0? given 2.0 is GA now, not sure it makes sense to dive as deep now

it may be easier to implement in 2.0, we added mapping in the release yesterday, too

I’m so sorry, currently our company still using 1.0… shortly we don’t see any plan to switch

does it mean Prefect 1.0 will no longer be supported as well as questions :sob:??
We purchase the 1.0 cloud

we always try to help, but just not sure if it's worth digging deep into building new solutions with 1.0. Both Prefect 1.0 and Cloud 1.0 are now in a maintenance mode and there will be no new features, so I'd encourage you to reconsider it with the migration

your use case is way easier to tackle in 2.0 - perhaps you can sign up for a free version and build only this project in 2.0 at first? this way you can avoid risky big bang migration and can slowly start migrating over starting with this project