When chaining Ops together in a job, are there any special requirements to being able to use syntax like

my_op_two(my_op_one())

? I was just trying to use the below code but getting this error:

dagster._core.errors.DagsterInvalidDefinitionError: In @job create_database_clones, received too many inputs for invocation create_database_clones_op. Only 0 defined, received 1

The only way I got this to work was by adding

ins={"start": In(Nothing)}

into the decorator on

create_database_clones_op()

So even if you don't explicitly return something in python, None still gets returned from a function call. So When you do

create_database_clones_op(drop_database_clones_op())

drop_database_clones_op

is passing a None into the

create_database_clones_op

. But

create_database_clones_op

doesn't expect any parameters, which is why you get the error with

Only 0 defined, received 1

. So if you don't expect any outputs to be passed from one op into another op, you have to do the

ins={"start": In(Nothing)}

, which is how you indicate to Dagster that no data should be passed

ah perfect — thanks for the simple explanation!