[GitHub] [airflow] jedcunningham commented on a change in pull request #21019: Function to expand mapped tasks in to multiple "real" TIs

[GitBox <gi...@apache.org>]

jedcunningham commented on a change in pull request #21019:
URL: https://github.com/apache/airflow/pull/21019#discussion_r791045742



##########
File path: airflow/models/baseoperator.py
##########
@@ -1800,6 +1801,53 @@ def wait_for_downstream(self) -> bool:
     def depends_on_past(self) -> bool:
         return self.partial_kwargs.get("depends_on_past") or self.wait_for_downstream
 
+    def expand_mapped_task(self, upstream_ti: "TaskInstance", session: "Session" = NEW_SESSION) -> None:
+        """Create the mapped TaskInstances for mapped task."""
+        # TODO: support having multiuple mapped upstreams?
+        from airflow.models.taskmap import TaskMap
+        from airflow.settings import task_instance_mutation_hook
+
+        task_map_info: TaskMap = (
+            session.query(TaskMap)
+            .filter_by(
+                dag_id=upstream_ti.dag_id,
+                task_id=upstream_ti.task_id,
+                run_id=upstream_ti.run_id,
+                map_index=upstream_ti.map_index,
+            )
+            .one()
+        )
+
+        unmapped_ti: Optional[TaskInstance] = upstream_ti.dag_run.get_task_instance(
+            self.task_id, map_index=-1, session=session
+        )
+
+        maps = range(task_map_info.length)
+
+        if unmapped_ti:
+            # The unmapped TaskInstance still exisxts -- this means we haven't

Review comment:
       ```suggestion
               # The unmapped TaskInstance still exists -- this means we haven't
   ```

##########
File path: tests/models/test_dagrun.py
##########
@@ -874,3 +875,20 @@ def test_verify_integrity_task_start_date(Stats_incr, session, run_type, expecte
     assert len(tis) == expected_tis
 
     Stats_incr.assert_called_with('task_instance_created-DummyOperator', expected_tis)
+
+
+@pytest.mark.xfail(reason="TODO: Expand mapped literals at verify_integrity time!")

Review comment:
       I'm not sure we can _only_ have it in the mini scheduler, as that can be turned off. Maybe toggle where it is run based on the mini scheduler being on or off?




-- 
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: commits-unsubscribe@airflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org