Hi all, I'm wanting to reuse an op to load multiple datasets from different datastores. My code looks something like

@op(required_resources={"file_system"})
def read_files(context):
    ...

first_dataset = AssetsDefinition.from_op(read_files)

second_dataset = AssetsDefinition.from_op(read_files)

where I want to be able to define the

"file_system"

resource for each of the two assets. I've tried

config_mapping

with and without

jobs

,

with_resource

but it seems only the resource defined in

Definition

is used. Maybe my implementaiton is wrong or maybe I've got entirely the wrong approach. Any suggestions would be appreciated.

hi Jeremy - have you considered an "asset factory"?

def make_dataset_asset(name, file_system_resource_key):
    @asset(name=name, required_resource_keys={file_system_resource_key})
    def _asset(context):
        ...

    return _asset


first_dataset = make_dataset_asset("asset1", "fs1")
second_dataset = make_dataset_asset("asset2", "fs2")

Thanks for the reply sandy. I think I had copy and pasted an asset previously when I was short on time, and a factory would be a little more elegant. But, it does mean rewriting any assets I want to reuse or knowing I want to reuse them a priori, which seems like a big blow to re-usability.