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

I'm confused about the `cache_key_fn` (and the default one mentioned). It would help to see an example of a custom one.

Oh, `task_input_hash` meaning, what's the hash (some unique, portable representation of all inputs) for the inputs/parameters of a task?

&gt; You can define a task that is cached based on its inputs by using the Prefect `task_input_hash`. This is a task cache key implementation that hashes all inputs to the task using a JSON or cloudpickle serializer. If the task inputs do not change, the cached results are used rather than running the task until the cache expires.

From <https://docs.prefect.io/latest/concepts/tasks/#caching|Prefect docs for tasks> ^

i can try to find a custom example to show you

That all jives with me. Powerful when you're passing custom classes and various types of input parameters. I can loosely imagine what a custom one might look like. E.g., if you need something that cloudpickle can't deal with directly.