uranusjr commented on code in PR #39912:
URL: https://github.com/apache/airflow/pull/39912#discussion_r1650531353
##########
airflow/models/dagrun.py:
##########
@@ -1538,11 +1538,21 @@ def schedule_tis(
and not ti.task.outlets
):
dummy_ti_ids.append((ti.task_id, ti.map_index))
- elif ti.task.start_from_trigger is True and
ti.task.start_trigger_args is not None:
- ti.start_date = timezone.utcnow()
- if ti.state != TaskInstanceState.UP_FOR_RESCHEDULE:
- ti.try_number += 1
- ti.defer_task(exception=None, session=session)
+ # check whether the operator supports start execution from
triggerer
+ elif ti.task.start_trigger_args is not None:
+ from airflow.models.mappedoperator import MappedOperator
+
+ if isinstance(ti.task, MappedOperator):
+ context = ti.get_template_context()
+ ti.task._expand_start_trigger(context=context,
session=session)
Review Comment:
The thing I’m not confortable here is we’re modifying the MappedOperator
in-place. I wonder if we can do something like this instead:
```python
if isinstance(ti.task, MappedOperator):
start_from_trigger = ti.task.expand_start_from_trigger(context, session)
else:
start_from_trigger = ti.task.start_from_trigger
if start_from_trigger:
...
```
and we can do the rest of the expansion in `defer_task` instead, also in a
similar fashion that does not modify the MappedOperator instance, but only the
unmapped BaseOperator.
--
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]
