@Glime I try to execute the same task multiple times (in parallel) on a list of inputs, so I am using

map_task

. The input for the map task is dynamically obtained (via database query), therefore I encapsulate that map task in a dynamic workflow. Below is a generalized example of what I am trying to do:

@task
def get_dynamic_params(...) -> list[str]:
    # Task implementation
    pass


@task
def my_task(dynamic_param: str, param1: str, param2: str,) -> DataFrame:
    # Task implementation
    pass


@dynamic
def my_dynamic_workflow(
    dynamic_params: list[str],
    param1: str,
    param2: str,
) -> list[DataFrame]:
    partial_my_task = functools.partial(
        my_task,
        param1: str,
        param2: str,
    )
    return map_task(partial_my_task)(dynamic_param=dynamic_params)

@workflow
def my_static_workflow(
    param1: str,
    param2: str,
) -> list[...]:
    dynamic_params = get_dynamic_params()
    my_data = my_dynamic_workflow(
        dynamic_params=dynamic_params,
	param1 = param1,
	param2 = param2,
    )
    # Workflow implementation
    pass

When trying to register the workflow, I am getting following errors:

Error 0: Code: MismatchingTypes, Node Id: end-node, Description: Variable [o0] (type [collection_type:<union_type:<variants:<map_value_type:<structured_dataset_type:<> > structure:<tag:"Typed Dict" > > variants:<simple:NONE structure:<tag:"none" > > > > ]) doesn't match expected type [collection_type:<map_value_type:<structured_dataset_type:<> > > ].
	Error 1: Code: ParameterNotBound, Node Id: end-node, Description: Parameter not bound [o0].

Based on the context provided and the error messages you're encountering, it seems like there are a couple of issues with how the workflow and dynamic workflow are set up, particularly around the use of

map_task

within a dynamic workflow and the handling of its outputs. 1. ​*Mismatching Types Error*​: This error suggests that there's a discrepancy between the expected output type of your dynamic workflow and what it's actually returning. The error message indicates that the system expects a collection of

DataFrame

objects but is getting something else. This could be due to how the

map_task

outputs are being handled or returned in the dynamic workflow. Ensure that the

map_task

within the dynamic workflow is correctly set up to return a list of

DataFrame

objects as expected by the workflow's output. 2. ​*Parameter Not Bound Error*​: This error typically occurs when a workflow or task expects an input parameter that hasn't been provided or is incorrectly referenced. In the context of your dynamic workflow, it's possible that the way outputs are being returned from the

map_task

or how the dynamic workflow's outputs are defined might be causing this issue. Ensure that all expected outputs of the dynamic workflow are correctly defined and that the `map_task`'s outputs are being correctly captured and returned. Given the complexity of using

map_task

within a dynamic workflow, especially with dynamic parameters, here are a few suggestions: • ​*Review

map_task

Usage*​: Ensure that the

map_task

is correctly set up to handle a list of inputs and that its usage within a dynamic workflow aligns with Flyte's expectations. The

map_task

should be capable of processing each item in the list of dynamic parameters independently and in parallel. • ​*Check Dynamic Workflow Outputs*​: Verify that the dynamic workflow's outputs are correctly defined and match what the

map_task

returns. The dynamic workflow should return a list of

DataFrame

objects if that's what the subsequent steps or the calling workflow expects. • ​*Partial Function with `map_task`*​: When using

functools.partial

with

map_task

, ensure that the partial function is correctly set up and that all non-dynamic parameters are correctly passed. This setup can sometimes lead to confusion or errors if not done correctly. Without the ability to directly modify your code or see the full implementation, these suggestions are based on common issues and best practices. If these suggestions do not resolve your issue, consider providing more detailed code snippets or error logs for further analysis. --- This is an AI-generated response, take its recommendations with caution. Please help us improve the model accuracy by giving this response a thumbs up/thumbs down reaction. 👍 button 👎 button