Dev-iL commented on code in PR #61491:
URL: https://github.com/apache/airflow/pull/61491#discussion_r2769955273
##########
airflow-core/src/airflow/migrations/versions/0033_3_0_0_add_tables_for_backfill.py:
##########
@@ -49,7 +48,7 @@ def upgrade():
sa.Column("dag_id", sa.String(length=250), nullable=False),
sa.Column("from_date",
airflow.utils.sqlalchemy.UtcDateTime(timezone=True), nullable=False),
sa.Column("to_date",
airflow.utils.sqlalchemy.UtcDateTime(timezone=True), nullable=False),
- sa.Column("dag_run_conf", sqlalchemy_jsonfield.jsonfield.JSONField(),
nullable=False),
Review Comment:
> You still modify the data type in the DB right?
I did not intend to modify the dtype in the DB - if I did, it's a mistake.
> If we don't modify this, how can we eliminate the import?
To clarify - when we `airflow db migrate` it runs all migrations from the
first to the last. What happens if one of the migrations requires a dependency
that doesn't exist? My impression is DB upgrades as well as downgrades are done
on the newer Airflow version (otherwise some migrations would not exist). Then,
- In the newer version it will work because the migration doesn't use the
now-missing extra dep.
- In the old airflow version it will work because the migration is an older
version, using the old type hint, but it's fine because the extra dependency is
available.
@potiuk @jscheffl thoughts?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
