Airflow DAG导入失败：'function'对象无'update_relative'属性错误求助

Hey there, let's work through this Airflow issue step by step!

Looking at your error traceback, the problem is clearly happening at this line:

normalise_task.set_downstream(predict_data)

The core issue here is that predict_data is a plain Python function, not an Airflow Task object (like an instance of PythonOperator, BashOperator, etc.). Airflow's set_downstream method only works when you pass it a valid Task instance—you can't directly pass a function here, which is why you're getting the "'function' object has no attribute 'update_relative'" error.

Here's how to fix this:

• Wrap your predict_data function into an Airflow Operator
  First, import the right operator (most commonly PythonOperator for Python functions), then create a Task instance from your function:

  from airflow.operators.python import PythonOperator
  
  # 假设这是你的预测逻辑函数
  def predict_data():
      # 这里写你的预测处理代码
      pass
  
  # 将函数包装成合法的Task实例
  predict_task = PythonOperator(
      task_id="predict_data_task",  # 给任务设置唯一ID
      python_callable=predict_data,  # 指定要执行的目标函数
      dag=dag  # 关联到你定义的DAG对象（确保dag是已初始化的DAG实例）
  )
  

• Set the dependency correctly using the Task instance
  Now that predict_task is a valid Task object, you can set the downstream dependency properly:

  normalise_task.set_downstream(predict_task)
  

• Bonus: Use Airflow's modern dependency syntax
  Airflow now recommends using the >> operator for setting dependencies—it's more readable and less error-prone. You can replace the set_downstream line with:

  normalise_task >> predict_task
  

One quick sanity check: also make sure normalise_task itself is a valid Task instance (not a function). If you made the same mistake there, you'd run into a similar error.

备注：内容来源于stack exchange，提问作者shashank bhardwaj