Remove unnecessary `start_date` from `schedules` in `meltano.yml`
This came up in a discussion with @umutykaya:
When a new schedule is created using meltano schedule
, the new entry under schedules
in meltano.yml
gets a start_date
value that defaults to the value of the extractor's start_date
setting, which then also shows up in the meltano schedule list --format=json
output so that the default Airflow DAG generator can use it.
Airflow DAGs that have a schedule_interval
need a start_date
, but as we don't need catchup, the only requirement is that it is in the past. The actual value does not matter, but having it in meltano.yml
gives users the mistaken impression that it does.
I think we can remove the key from meltano.yml
and hard-code a value like 1970-01-01
in the DAG generator (for new users) and in the meltano schedule list --format=json
output (for existing DAG generators that expect the start_date
key).