Remove execution_date and logical_date from arguments in api_connexion #43678
Conversation
|
Are we going to remove it first and then move the API ? I was under the impression we wanted to make it new "while" moving it to the new fast_api framework? But maybe there is a good reason to do it first here? |
As long as there's no conflict, I think it's ok to do it first 🤔 so that whoever's working on moving the API |
| @@ -285,7 +285,7 @@ def stats_tags(self) -> dict[str, str]: | |||
| return prune_dict({"dag_id": self.dag_id, "run_type": self.run_type}) | |||
|
|
|||
| @property | |||
| def logical_date(self) -> datetime: | |||
| def logical_date(self): | |||
| return self.execution_date | |||
There was a problem hiding this comment.
I think this is going to be removed in another PR?
| ret_data = {} | ||
| data["execution_date"] = data["logical_date"] | ||
| data["logical_date"] = data["logical_date"] |
There was a problem hiding this comment.
| data["logical_date"] = data["logical_date"] |
I think we probably don't need this line?
| @@ -212,8 +212,8 @@ class SetTaskInstanceStateFormSchema(Schema): | |||
| @validates_schema | |||
| def validate_form(self, data, **kwargs): | |||
| """Validate set task instance state form.""" | |||
| if not exactly_one(data.get("execution_date"), data.get("dag_run_id")): | |||
| raise ValidationError("Exactly one of execution_date or dag_run_id must be provided") | |||
| if not exactly_one(data.get("logical_date"), data.get("dag_run_id")): | |||
There was a problem hiding this comment.
Is this still needed after the removal in airflow/api_connexion/openapi/v1.yaml
| @@ -173,12 +173,12 @@ def _fetch_dag_runs( | |||
| query = query.where(DagRun.updated_at <= updated_at_lte) | |||
|
|
|||
| total_entries = get_query_count(query, session=session) | |||
| to_replace = {"dag_run_id": "run_id", "execution_date": "logical_date"} | |||
| to_replace = {"dag_run_id": "run_id", "logical_date": "logical_date"} | |||
|
For these pull requests, we will probably need the |
This PR removes the
execution_dateandlogical_datearguments fromapi_connexionthat are used to retrieve DAG runs, aligning with the broader changes introduced in Airflow 2.2 and preparing for Airflow 3.0. The functions now userun_idas the sole identifier for DAG runs, simplifying the process and eliminating deprecated behaviour.Motivation:
In Airflow,
execution_datehas historically been used to distinguish different DAG run instances. However, the introduction ofrun_idand the DAG run concept in Airflow 2.2 shifts away from usingexecution_dateas an identifier. Continuing to rely onexecution_dateintroduces limitations, such as the inability to handle multiple DAG runs at the same logical time, especially in cases likeTriggerDagRunOperatorwhen dynamic runs are generated.By removing
execution_datein favor ofrun_id, this PR eliminates these limitations. This also removes the unique constraint onexecution_dateat the database level, paving the way for a cleaner and more flexible scheduling system in Airflow 3.0.^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named
{pr_number}.significant.rstor{issue_number}.significant.rst, in newsfragments.