Fix NoneType error when updating serialized DAG

[anshuksi282-ksolves]

Issue: #56306
FIX: Prevent AttributeError when updating SerializedDagModel for dynamic DAGs

This PR fixes a stability issue in SerializedDagModel.write_dag by adding a null check before updating an existing record for dynamic DAGs.

Context and Problem

In the section of write_dag dedicated to updating dynamic DAGs (i.e., when dag_version exists but has no task instances), the code retrieves the latest SerializedDagModel instance using cls.get(dag.dag_id, session=session).

If cls.get(dag.dag_id, session=session) returns None (for example, due to a race condition where the record was just deleted or not found in the current session), the subsequent code attempts to modify attributes:

latest_ser_dag._data = new_serialized_dag._data

This would raise an AttributeError because latest_ser_dag is None.

Solution

• Added a null check to ensure that latest_ser_dag exists before updating its attributes.
• This prevents AttributeError and improves the stability of DAG serialization for dynamic DAGs.

---

^ 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.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

[boring-cyborg]

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

• Pay attention to the quality of your code (ruff, mypy and type annotations). Our prek-hooks will help you with that.
• In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
• Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
• Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
• Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
• Be sure to read the Airflow Coding style.
• Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
  Apache Airflow is a community-driven project and together we are making it better 🚀.
  In case of doubts contact the developers at:
  Mailing List: dev@airflow.apache.org
  Slack: https://s.apache.org/airflow-slack

[anshuksi282-ksolves]

Hi ,I’ve fixed a stability issue in SerializedDagModel.write_dag for dynamic DAGs (#56306).
Added a null check before updating an existing record to prevent AttributeError.
Tested locally, and dynamic DAG serialization now works reliably.
Appreciate your review and feedback. Thanks!

[ephraimbuddy]

Looks like bad rebase, please update and have only the query change

[anshuksi282-ksolves]

Hi @ephraimbuddy,

Thanks for reviewing! Sorry about the earlier confusion — the bad rebase included extra changes by mistake.
I’ve now cleaned up the branch so it contains only the query change in SerializedDagModel.

Appreciate your patience and review!

[anshuksi282-ksolves]

Hi @ephraimbuddy,

Thanks for the approval!

I see the CI tests are failing with the error: AttributeError: 'ScalarResult' object has no attribute 'limit'.
I also noticed the where clause in the query might be too strict (filtering by bundle_version and bundle_name), which could cause issues like the assert 6 == 4 failure I saw in other test logs.

I believe I can fix both issues at once. The plan is:

• Fix the AttributeError by moving .limit(1) inside the select() and adding .first().
• Fix the logic error by updating the where() clause to only filter by dag_id.

dag_version = session.scalars(
            select(DagVersion)
            .where(
                DagVersion.dag_id == dag.dag_id,
            )
            .options(joinedload(DagVersion.task_instances))
            .options(joinedload(DagVersion.serialized_dag))
            .order_by(DagVersion.created_at.desc())
            .limit(1)
        ).first() 

Should I go ahead and push this combined fix?

Thanks!

[ephraimbuddy]

Maybe:

dag_version = session.scalar(
            select(DagVersion)
            .where(
                DagVersion.dag_id == dag.dag_id,
                DagVersion.bundle_name == bundle_name
            )
            .options(joinedload(DagVersion.task_instances))
            .options(joinedload(DagVersion.serialized_dag))
            .order_by(DagVersion.created_at.desc())
            .limit(1)
        )

In the above, I used scalar instead of scalars and removed first(), there's no need since limit is 1. I removed bundle_version in query. I think this will work

[anshuksi282-ksolves]

Hi @ephraimbuddy, thanks! I've updated the code to use session.scalar() as you suggested
Do I still need to add theif latest_ser_dag:null check to fix the original NoneType error, or is this query change sufficient on its own?

[ephraimbuddy]

Hi @ephraimbuddy, thanks! I've updated the code to use session.scalar() as you suggested Do I still need to add theif latest_ser_dag:null check to fix the original NoneType error, or is this query change sufficient on its own?

You can do that. Then use return statement when the latest_ser_dag is null. You can return False in null case and get rid of assert check

[anshuksi282-ksolves]

Hi @ephraimbuddy, I've rebased and pushed the changes.

I used session.scalar() as you suggested, thanks! I noticed you included bundle_name in the query, but that was causing the assert 6 == 4 test to fail (it was creating new rows instead of updating).

So, I've kept only dag_id in the where clause, which I think will fix it.

I also added the if not latest_ser_dag: return False check for the original NoneType error, as you confirmed.

This should hopefully fix all the CI failures now. Thanks!