You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'd like the file target/run_results.json to be generated with an empty results list even when the selector selected no nodes. The current behavior is to not generate a new run_results.json when the selector returned no nodes.
A test run with no nodes selected is a legitimate occurence when using the state:modified selector.
Describe alternatives you've considered
Parsing dbt test output for the string WARNING: Nothing to do. Try checking your model configs and model specification args. This is a bit hacky and unreliable though.
Additional context
Using dbt test with the state:modified selector in my CI pipeline. I parse run_results,json to get the test outcome, but that file doesn't get generated when there are no model differences between the branch and PROD. However, the absence of run_results.json could also mean the test process somehow failed...
Who will this benefit?
Anyone who depends on run_results.json - for example CI pipeline devs.
Are you interested in contributing this feature?
I'll browse the source code and give it a shot, but no promises.
The text was updated successfully, but these errors were encountered:
@panasenco I buy it! Thanks so much for opening the issue, and for laying out your rationale so clearly. It looks like this was a super simple change, too.
Describe the feature
I'd like the file
target/run_results.json
to be generated with an empty results list even when the selector selected no nodes. The current behavior is to not generate a newrun_results.json
when the selector returned no nodes.A test run with no nodes selected is a legitimate occurence when using the
state:modified
selector.Describe alternatives you've considered
Parsing
dbt test
output for the stringWARNING: Nothing to do. Try checking your model configs and model specification args
. This is a bit hacky and unreliable though.Additional context
Using
dbt test
with thestate:modified
selector in my CI pipeline. I parserun_results,json
to get the test outcome, but that file doesn't get generated when there are no model differences between the branch and PROD. However, the absence ofrun_results.json
could also mean the test process somehow failed...Who will this benefit?
Anyone who depends on
run_results.json
- for example CI pipeline devs.Are you interested in contributing this feature?
I'll browse the source code and give it a shot, but no promises.
The text was updated successfully, but these errors were encountered: