Hi, is there a way how I can make to subgraphs dependent on each other? In the job i want to run graph1 before graph2. These are effectively the same graphs, just with different configurations.

You could use the run_status_sensor (https://docs.dagster.io/concepts/partitions-schedules-sensors/sensors#run-status-sensors) to launch a new run via the GraphQL API (https://docs.dagster.io/concepts/dagit/graphql). This might be a slightly hacky workaround, but it should work

I thiiiiink you can have a dummy input to graph 2 that is supplied by Graph 1, like so:

@graph
def g1():
   result = op1()
   return result

@graph
def g2(depends_on=None):
   op2()

@job
def two_graphs():
   result = g1()
   g2(result)

ok, that might work. Currently, I want to reuse the same graph, just with different configs. AFAIK, there are not optional inputs for a graph. The first graph would not have an input at the start.

i think the only way to do order-based stuff is by passing in a dependency, so you might have to add the

depends_on

input to the op itself

Hehe, thanks. I will keep an eye on that.

Okay so I solved the problem like this:

@op
def do_nothing(value):
    return

@graph(ins={'deps': In(Nothing)})
def my_graph(deps):
    do_nothing(deps)
    ...