Prevent using trigger_rule="always" in a dynamic mapped task

Author: shahar1

airflow/utils/task_group.py

@@ -37,6 +37,7 @@
 from airflow.models.taskmixin import DAGNode
 from airflow.serialization.enums import DagAttributeTypes
 from airflow.utils.helpers import validate_group_key, validate_instance_args
+from airflow.utils.trigger_rule import TriggerRule
 
 if TYPE_CHECKING:
     from sqlalchemy.orm import Session
@@ -220,10 +221,15 @@ def parent_group(self) -> TaskGroup | None:
 
     def __iter__(self):
         for child in self.children.values():
-            if isinstance(child, TaskGroup):
-                yield from child
-            else:
-                yield child
+            yield from self._iter_child(child)
+
+    @staticmethod
+    def _iter_child(child):
+        """Iterate over the children of this TaskGroup."""
+        if isinstance(child, TaskGroup):
+            yield from child
+        else:
+            yield child
 
     def add(self, task: DAGNode) -> DAGNode:
         """
@@ -593,6 +599,14 @@ def __init__(self, *, expand_input: ExpandInput, **kwargs: Any) -> None:
         super().__init__(**kwargs)
         self._expand_input = expand_input
 
+    def __iter__(self):
+        from airflow.models.abstractoperator import AbstractOperator
+
+        for child in self.children.values():
+            if isinstance(child, AbstractOperator) and child.trigger_rule == TriggerRule.ALWAYS:
+                raise ValueError("Tasks in a mapped task group cannot have trigger_rule set to 'ALWAYS'")
+            yield from self._iter_child(child)
+
     def iter_mapped_dependencies(self) -> Iterator[Operator]:
         """Upstream dependencies that provide XComs used by this mapped task group."""
         from airflow.models.xcom_arg import XComArg

docs/apache-airflow/authoring-and-scheduling/dynamic-task-mapping.rst

@@ -84,6 +84,11 @@ The grid view also provides visibility into your mapped tasks in the details pan
 
     Although we show a "reduce" task here (``sum_it``) you don't have to have one, the mapped tasks will still be executed even if they have no downstream tasks.
 
+.. warning:: ``TriggerRule.ALWAYS`` cannot be utilized in expanded tasks
+
+    Assigning ``trigger_rule=TriggerRule.ALWAYS`` in expanded tasks is forbidden, as expanded parameters will be undefined with the task's immediate execution.
+    This is enforced at the time of the DAG parsing, and will raise an error if you try to use it.
+
 Task-generated Mapping
 ----------------------
 

tests/decorators/test_task_group.py

@@ -22,10 +22,11 @@
 import pendulum
 import pytest
 
-from airflow.decorators import dag, task_group
+from airflow.decorators import dag, task, task_group
 from airflow.models.expandinput import DictOfListsExpandInput, ListOfDictsExpandInput, MappedArgument
 from airflow.operators.empty import EmptyOperator
 from airflow.utils.task_group import MappedTaskGroup
+from airflow.utils.trigger_rule import TriggerRule
 
 
 def test_task_group_with_overridden_kwargs():
@@ -133,6 +134,28 @@ def tg():
     assert str(ctx.value) == "no arguments to expand against"
 
 
+@pytest.mark.db_test
+def test_expand_fail_trigger_rule_always(dag_maker, session):
+    @dag(schedule=None, start_date=pendulum.datetime(2022, 1, 1))
+    def pipeline():
+        @task
+        def get_param():
+            return ["a", "b", "c"]
+
+        @task(trigger_rule=TriggerRule.ALWAYS)
+        def t1(param):
+            return param
+
+        @task_group()
+        def tg(param):
+            t1(param)
+
+        with pytest.raises(
+            ValueError, match="Tasks in a mapped task group cannot have trigger_rule set to 'ALWAYS'"
+        ):
+            tg.expand(param=get_param())
+
+
 def test_expand_create_mapped():
     saved = {}