Airflow operators

Operators are one of the building blocks of Airflow DAGs. There are many different types of operators available in Airflow. The PythonOperator can execute any Python function, and is functionally equivalent to using the @task decorator, while other operators contain pre-created logic to perform a specific task, such as executing a Bash script (BashOperator) or running a SQL query in a relational database (SQLExecuteQueryOperator). Operators are used alongside other building blocks, such as decorators and hooks, to create tasks in a DAG written with the task-oriented approach. Operators classes can be imported from Airflow provider packages.

In this guide, you’ll learn the basics of using operators in Airflow.

To view a list of available operators available in different Airflow provider packages, go to the Astronomer Registry.

Assumed knowledge

To get the most out of this guide, you should have an understanding of:

• Basic Airflow concepts. See Introduction to Apache Airflow.
• Basic Python. See the Python Documentation.

Operator basics

Operators are Python classes that encapsulate logic to do a unit of work. They can be viewed as a wrapper around each unit of work that defines the actions that will be completed and abstract the majority of code you would typically need to write. When you create an instance of an operator in a DAG and provide it with its required parameters, it becomes a task.

A base set of operators is contained in the Airflow standard provider package, which is pre-installed when using the Astro CLI. Other operators are contained in specialized provider packages, often centered around a specific technology or service. For example, the Airflow Snowflake Provider package contains operators for interacting with Snowflake, while the Airflow Google provider package contains operators for interacting with Google Cloud services. There are also several packages that contain operators that can be used with a set of services:

• Common SQL
• Common IO
• Common Messaging

Operator examples

Following are some of the most frequently used Airflow operators. Note that only a few of the possible parameters are shown, refer to the Astronomer registry for a full list of parameters for each operator.

• PythonOperator: Executes a Python function. It is functionally equivalent to using the @task decorator. See, Introduction to the TaskFlow API and Airflow decorators.

  1
  from airflow.providers.standard.operators.python import PythonOperator
  2
  3
  def _my_python_function():
  4
      print("Hello world!")
  5
  6
  my_task = PythonOperator(
  7
      task_id="my_task",
  8
      python_callable=_my_python_function,
  9
  )

• BashOperator: Executes a bash script. See also the Using the BashOperator guide.

  1
  from airflow.providers.standard.operators.bash import BashOperator
  2
  3
  my_task = BashOperator(
  4
      task_id="my_task",
  5
      bash_command="echo 'Hello world!'",
  6
  )

• KubernetesPodOperator: Executes a task defined as a Docker image in a Kubernetes Pod. See, Use the KubernetesPodOperator.

  1
  from airflow.providers.cncf.kubernetes.operators.pod import KubernetesPodOperator
  2
  3
  my_task = KubernetesPodOperator(
  4
      task_id="my_task",
  5
      kubernetes_conn_id="<my-kubernetes-connection>",
  6
      name="<my-pod-name>",
  7
      namespace="<my-namespace>",
  8
      image="python:3.12-slim", # Docker image to run
  9
      cmds=["python", "-c"], # Command to run in the container
  10
      arguments=["print('Hello world!')"], # Arguments to the command
  11
  )

• SQLExecuteQueryOperator: Executes a SQL query against a relational database.

  1
  from airflow.providers.common.sql.operators.sql import SQLExecuteQueryOperator
  2
  3
  my_task = SQLExecuteQueryOperator(
  4
      task_id="my_task",
  5
      sql="SELECT * FROM my_table",
  6
      database="<my-database>",
  7
      conn_id="<my-connection>",
  8
  )

• EmptyOperator: A no-op operator that does nothing. This is useful for creating placeholder tasks in a DAG.

  1
  from airflow.providers.standard.operators.empty import EmptyOperator
  2
  3
  my_task = EmptyOperator(task_id="my_task")

All operators inherit from the abstract BaseOperator class, which contains the logic to execute the work of the operator within the context of a DAG.

Arguments of the BaseOperator class can be passed to all operators. The most common arguments are:

• task_id: A unique identifier for the task. This is required for all operators.
• retries: The number of times to retry the task if it fails. This is optional and defaults to 0. See Rerun Airflow DAGs and tasks.
• pool: The name of the pool to use for the task. This is optional and defaults to None. See Airflow pools.
• execution_timeout: The maximum time to wait for the task to complete. This is optional and defaults to None. It is a good practice to set this value to prevent tasks from running indefinitely.

You can set these arguments and other BaseOperator arguments (other than task_id which needs to be unique per operator) at the DAG level for all tasks in a DAG. By using the default_args dictionary. You can override these values for individual tasks by setting the same arguments in the task definition.

1
import hashlib
2
import json
3
4
from airflow.exceptions import AirflowException
5
from airflow.decorators import dag, task
6
from airflow.models import Variable
7
from airflow.models.baseoperator import chain
8
from airflow.operators.empty import EmptyOperator
9
from airflow.utils.dates import datetime
10
from airflow.providers.amazon.aws.hooks.s3 import S3Hook
11
from airflow.providers.amazon.aws.transfers.local_to_s3 import (
12
    LocalFilesystemToS3Operator,
13
)
14
from airflow.providers.amazon.aws.transfers.s3_to_redshift import S3ToRedshiftOperator
15
from airflow.providers.postgres.operators.postgres import PostgresOperator
16
from airflow.operators.sql import SQLCheckOperator
17
from airflow.utils.task_group import TaskGroup
18
19
20
# The file(s) to upload shouldn't be hardcoded in a production setting,
21
# this is just for demo purposes.
22
CSV_FILE_NAME = "forestfires.csv"
23
CSV_FILE_PATH = f"include/sample_data/forestfire_data/{CSV_FILE_NAME}"
24
25
26
@dag(
27
    "simple_redshift_3",
28
    start_date=datetime(2021, 7, 7),
29
    description="""A sample Airflow DAG to load data from csv files to S3
30
                 and then Redshift, with data integrity and quality checks.""",
31
    schedule=None,
32
    template_searchpath="/usr/local/airflow/include/sql/redshift_examples/",
33
    catchup=False,
34
)
35
def simple_redshift_3():
36
    """
37
    Before running the DAG, set the following in an Airflow
38
    or Environment Variable:
39
    - key: aws_configs
40
    - value: { "s3_bucket": [bucket_name], "s3_key_prefix": [key_prefix],
41
             "redshift_table": [table_name]}
42
    Fully replacing [bucket_name], [key_prefix], and [table_name].
43
    """
44
45
    upload_file = LocalFilesystemToS3Operator(
46
        task_id="upload_to_s3",
47
        filename=CSV_FILE_PATH,
48
        dest_key="{{ var.json.aws_configs.s3_key_prefix }}/" + CSV_FILE_PATH,
49
        dest_bucket="{{ var.json.aws_configs.s3_bucket }}",
50
        aws_conn_id="aws_default",
51
        replace=True,
52
    )
53
54
    @task
55
    def validate_etag():
56
        """
57
        #### Validation task
58
        Check the destination ETag against the local MD5 hash to ensure
59
        the file was uploaded without errors.
60
        """
61
        s3 = S3Hook()
62
        aws_configs = Variable.get("aws_configs", deserialize_json=True)
63
        obj = s3.get_key(
64
            key=f"{aws_configs.get('s3_key_prefix')}/{CSV_FILE_PATH}",
65
            bucket_name=aws_configs.get("s3_bucket"),
66
        )
67
        obj_etag = obj.e_tag.strip('"')
68
        # Change `CSV_FILE_PATH` to `CSV_CORRUPT_FILE_PATH` for the "sad path".
69
        file_hash = hashlib.md5(open(CSV_FILE_PATH).read().encode("utf-8")).hexdigest()
70
        if obj_etag != file_hash:
71
            raise AirflowException(
72
                """Upload Error: Object ETag in S3 did not match
73
                hash of local file."""
74
            )
75
76
    # Tasks that were created using decorators have to be called to be used
77
    validate_file = validate_etag()
78
79
    # --- Create Redshift Table --- #
80
    create_redshift_table = PostgresOperator(
81
        task_id="create_table",
82
        sql="create_redshift_forestfire_table.sql",
83
        postgres_conn_id="redshift_default",
84
    )
85
86
    # --- Second load task --- #
87
    load_to_redshift = S3ToRedshiftOperator(
88
        task_id="load_to_redshift",
89
        s3_bucket="{{ var.json.aws_configs.s3_bucket }}",
90
        s3_key="{{ var.json.aws_configs.s3_key_prefix }}" + f"/{CSV_FILE_PATH}",
91
        schema="PUBLIC",
92
        table="{{ var.json.aws_configs.redshift_table }}",
93
        copy_options=["csv"],
94
    )
95
96
    # --- Redshift row validation task --- #
97
    validate_redshift = SQLCheckOperator(
98
        task_id="validate_redshift",
99
        conn_id="redshift_default",
100
        sql="validate_redshift_forestfire_load.sql",
101
        params={"filename": CSV_FILE_NAME},
102
    )
103
104
    # --- Row-level data quality check --- #
105
    with open("include/validation/forestfire_validation.json") as ffv:
106
        with TaskGroup(group_id="row_quality_checks") as quality_check_group:
107
            ffv_json = json.load(ffv)
108
            for id, values in ffv_json.items():
109
                values["id"] = id
110
                SQLCheckOperator(
111
                    task_id=f"forestfire_row_quality_check_{id}",
112
                    conn_id="redshift_default",
113
                    sql="row_quality_redshift_forestfire_check.sql",
114
                    params=values,
115
                )
116
117
    # --- Drop Redshift table --- #
118
    drop_redshift_table = PostgresOperator(
119
        task_id="drop_table",
120
        sql="drop_redshift_forestfire_table.sql",
121
        postgres_conn_id="redshift_default",
122
    )
123
124
    begin = EmptyOperator(task_id="begin")
125
    end = EmptyOperator(task_id="end")
126
127
    # --- Define task dependencies --- #
128
    chain(
129
        begin,
130
        upload_file,
131
        validate_file,
132
        create_redshift_table,
133
        load_to_redshift,
134
        validate_redshift,
135
        quality_check_group,
136
        drop_redshift_table,
137
        end,
138
    )
139
140
141
simple_redshift_3()

Best practices

Operators typically only require a few parameters. Keep the following considerations in mind when using Airflow operators:

• The Astronomer Registry is the best resource for learning what operators are available and how they are used.
• The Airflow standard provider package includes basic operators such as the PythonOperator and BashOperator. These operators are automatically available in your Airflow environment if you are using the Astro CLI. All other operators are part of provider packages, some which you must install separately, depending on what type of Airflow distribution you are using.
• You can combine operators and decorators freely in the same DAG. Many users choose to use the @task decorator for most of their tasks, and add operators for tasks where a specialized operator exists for their use case. The example above shows a DAG with one operator (BashOperator) and one @task decorated task.
• If an operator exists for your specific use case, you should use it instead of your own Python functions or hooks. This makes your DAGs easier to read and maintain.
• If an operator doesn’t exist for your use case, you can either use custom Python code in an @task decorated task or PythonOperator or extend an operator to meet your needs. For more information about customizing operators, see Custom hooks and operators.
• Sensors are a type of operator that waits for something to happen. They can be used to detect events in systems outside of Airflow.
• Deferrable Operators are a type of operator that releases their worker slot while waiting for their work to be completed. This can result in cost savings and greater scalability. Astronomer recommends using deferrable operators whenever one exists for your use case and your task takes longer than a minute. A lot of operators that potentially need to wait for something have a deferrable mode which you can enable by setting their deferrable parameter to True.
• Any operator that interacts with a service external to Airflow typically requires a connection so that Airflow can authenticate to that external system. For more information about setting up connections, see Managing your connections in Apache Airflow or in the examples to follow.