https://dagster.io/ logo
#ask-community
Title
# ask-community
j

Jayme Edwards

09/13/2022, 9:26 PM
Hey y'all I have a job that needs to call 4 ops that run parallel, then run dbt_run_op at the end. It runs dbt_run_op before finishing all the others. I have seen an article that shows how to make ops depend on each other, but it doesn't seem to work with dbt_run_op since it expects a context parameter, and context doesn't seem to get passed in "nested" ops: https://docs.dagster.io/tutorial/ops-jobs/connecting-ops#lets-get-serial
🤖 1
s

sean

09/13/2022, 10:05 PM
Hi Jayme, I may need to hand this off to someone on our team more knowledgable about our dbt integration, but first-- if you view your job in dagit, does it show the graph structure you would expect (i.e. reflects the dependencies you’re aiming for)?
o

owen

09/13/2022, 10:18 PM
hi! just at a high level, the structure of your job should look something like:
Copy code
@job
def my_job():
    dbt_run_op(start_after=[op_a(), op_b(), op_c(), op_d()])
the dbt ops provide a start_after parameter, which is a nothing-type input: https://docs.dagster.io/concepts/ops-jobs-graphs/graphs#defining-nothing-dependencies
blob cheer 1
j

Jayme Edwards

09/13/2022, 10:27 PM
@owen thanks, I've only been using jobs that call individual ops. It looks like this graph is required to create dependencies between ops when there is no input into one I guess?
@owen nevermind I tried your code snippet. Works perfectly. Thank you!!!
🎉 1
o

owen

09/13/2022, 10:31 PM
no problem 🙂
❤️ 1