An operator is the executable implementation behind a workflow task. It lets a DAG declare what each task does while the orchestration system schedules, dispatches, runs, and monitors that work. section 9.3.1 section 9.2.2
Choose an operator according to the task and the amount of customization required:
- Use predefined or community operators such as MySqlOperator, SimpleHttpOperator, Docker operator, and PostgresOperator to implement common tasks without writing task logic. section 9.3.1
- Use PythonOperator to run a customized Python function. section 9.3.1
- Use a default shell or Python operator, or build a custom operator when the orchestration system lacks the required task behavior. section 9.2.2
Operators in a data workflow
A daily Airflow DAG can use FileSensor to wait for data.csv, PythonOperator to transform data, and PostgresOperator to create a database table. Dependencies determine their execution sequence. section 9.3.1
