Skip to content

can not use output of task decorator as input for external_task_ids of ExternalTaskSensor #27952

Description

@MaiHoangViet1809

Apache Airflow version

2.4.3, 2.5.0

What happened

when use output from task decorator as parameter (external_task_ids) in ExternalTaskSensor, it show up this log:

Broken DAG: [+++++/airflow/dags/TEST_NEW_PIPELINE.py] Traceback (most recent call last):
  File "+++++/env3.10.5/lib/python3.10/site-packages/airflow/models/baseoperator.py", line 408, in apply_defaults
    result = func(self, **kwargs, default_args=default_args)
  File "+++++/env3.10.5/lib/python3.10/site-packages/airflow/sensors/external_task.py", line 164, in __init__
    if external_task_ids and len(external_task_ids) > len(set(external_task_ids)):
TypeError: object of type 'PlainXComArg' has no len()

note: +++++ is just a mask for irrelevant information.

What you think should happen instead

this document https://airflow.apache.org/docs/apache-airflow/stable/tutorial/taskflow.html
show that we can use it without any warning, note about it.

found relative problem in https://github.com/apache/airflow/issues/27328
should move all the check in __init__ into poke method.

How to reproduce

from airflow.decorators import dag, task
from airflow.operators.python import get_current_context

from airflow.sensors.external_task import ExternalTaskSensor
from datetime import datetime

configure = {"dag_id": "test_new_skeleton",
                    "schedule": None,
                    "start_date": datetime(2022,1,1),
}

@task
def preprocess_dependency() -> list:
    return ["random-task-name"]

@dag(**configure)
def pipeline(): 
    t_preprocess = preprocess_dependency()
    task_dependency = ExternalTaskSensor(task_id=f"Check_Dependency",
                                         external_dag_id='random-dag-name-that-exist',
                                         external_task_ids=t_preprocess ,
                                         poke_interval=60,
                                         mode="reschedule",
                                         timeout=172800,
                                         allowed_states=['success'],
                                         failed_states=['failed', 'skipped'],
                                         check_existence=True,)


dag = pipeline()

Operating System

REHL 7

Versions of Apache Airflow Providers

No response

Deployment

Virtualenv installation

Deployment details

No response

Anything else

No response

Are you willing to submit PR?

  • Yes I am willing to submit a PR!

Code of Conduct

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions