Hi - I am trying to understand if there is (1) a reason to configure a result in each task as opposed to just the flow . (2) And is setting the target enough if setting the result in the flow suffices ?

Hey itay! Most use cases work perfectly well with a single result configured on the Flow + individual task targets. To give you an example of where more nuanced configuration is desirable, I might have a setup that looks like:

@task
def return_config_object(param):
    return some_dictionary

@task
def process_big_data(config):
    return pandas_dataframe

In the above case, I might want to see the config in the UI in which case I’ll configure it to use a

PrefectResult

type. The pandas dataframe task, on the other hand, I probably want stored in cloud storage (e.g.,

S3Result

)

👏 Thank you!