Hi All, I am running a kedro pipeline along with databricks notebooks. using the below code in a databricks.

reload_kedro("../../../", env="base", extra_params=extra_params)

 with KedroSession.create(project_path="../", env="base", extra_params=extra_params) as session:
        session.run(pipeline_name="kedro_pipeline_name")

It runs a kedro session with the input parameters. Is there a way the kedro pipeline can return params/value back to the databricks notebooks? something similar to a

return

statement.

Hey Adith, what values do you want the session to return? The

session.run()

returns a

dict

of datasets that are “free outputs”, i.e. datasets that are

MemoryDataSet

by default

I am looking to return a dictionary of values (which can be the output of a kedro node). So if I am not wrong, if I define the output of the kedro node as a MemoryDataSet, I can access the output as a return value of session.run()?

Yes if the outputs are not defined in the catalog, they’re treated as

MemoryDataSet

and should be returned as the result of the

session.run()