Base operator for all operators.

Module Contents



Metaclass of BaseOperator.


Abstract base class for all operators.


Abstract base class that defines how we get an operator link.



coerce_timedelta(value, *, key)


get_merged_defaults(dag, task_group, task_params, ...)

partial(operator_class, *, task_id[, dag, task_group, ...])


Given a number of tasks, builds a dependency chain.

cross_downstream(from_tasks, to_tasks)

Set downstream dependencies for all tasks in from_tasks to all tasks in to_tasks.


Simplify task dependency definition.









airflow.models.baseoperator.coerce_timedelta(value, *, key)[source]
airflow.models.baseoperator.get_merged_defaults(dag, task_group, task_params, task_default_args)[source]
airflow.models.baseoperator.partial(operator_class, *, task_id, dag=None, task_group=None, start_date=NOTSET, end_date=NOTSET, owner=NOTSET, email=NOTSET, params=None, resources=NOTSET, trigger_rule=NOTSET, depends_on_past=NOTSET, ignore_first_depends_on_past=NOTSET, wait_for_past_depends_before_skipping=NOTSET, wait_for_downstream=NOTSET, retries=NOTSET, queue=NOTSET, pool=NOTSET, pool_slots=NOTSET, execution_timeout=NOTSET, max_retry_delay=NOTSET, retry_delay=NOTSET, retry_exponential_backoff=NOTSET, priority_weight=NOTSET, weight_rule=NOTSET, sla=NOTSET, max_active_tis_per_dag=NOTSET, max_active_tis_per_dagrun=NOTSET, on_execute_callback=NOTSET, on_failure_callback=NOTSET, on_success_callback=NOTSET, on_retry_callback=NOTSET, run_as_user=NOTSET, executor_config=NOTSET, inlets=NOTSET, outlets=NOTSET, doc=NOTSET, doc_md=NOTSET, doc_json=NOTSET, doc_yaml=NOTSET, doc_rst=NOTSET, **kwargs)[source]
class airflow.models.baseoperator.BaseOperatorMeta[source]

Bases: abc.ABCMeta

Metaclass of BaseOperator.

class airflow.models.baseoperator.BaseOperator(task_id, owner=DEFAULT_OWNER, email=None, email_on_retry=conf.getboolean('email', 'default_email_on_retry', fallback=True), email_on_failure=conf.getboolean('email', 'default_email_on_failure', fallback=True), retries=DEFAULT_RETRIES, retry_delay=DEFAULT_RETRY_DELAY, retry_exponential_backoff=False, max_retry_delay=None, start_date=None, end_date=None, depends_on_past=False, ignore_first_depends_on_past=DEFAULT_IGNORE_FIRST_DEPENDS_ON_PAST, wait_for_past_depends_before_skipping=DEFAULT_WAIT_FOR_PAST_DEPENDS_BEFORE_SKIPPING, wait_for_downstream=False, dag=None, params=None, default_args=None, priority_weight=DEFAULT_PRIORITY_WEIGHT, weight_rule=DEFAULT_WEIGHT_RULE, queue=DEFAULT_QUEUE, pool=None, pool_slots=DEFAULT_POOL_SLOTS, sla=None, execution_timeout=DEFAULT_TASK_EXECUTION_TIMEOUT, on_execute_callback=None, on_failure_callback=None, on_success_callback=None, on_retry_callback=None, pre_execute=None, post_execute=None, trigger_rule=DEFAULT_TRIGGER_RULE, resources=None, run_as_user=None, task_concurrency=None, max_active_tis_per_dag=None, max_active_tis_per_dagrun=None, executor_config=None, do_xcom_push=True, inlets=None, outlets=None, task_group=None, doc=None, doc_md=None, doc_json=None, doc_yaml=None, doc_rst=None, **kwargs)[source]

Bases: airflow.models.abstractoperator.AbstractOperator

Abstract base class for all operators.

Since operators create objects that become nodes in the DAG, BaseOperator contains many recursive methods for DAG crawling behavior. To derive from this class, you are expected to override the constructor and the ‘execute’ method.

Operators derived from this class should perform or trigger certain tasks synchronously (wait for completion). Example of operators could be an operator that runs a Pig job (PigOperator), a sensor operator that waits for a partition to land in Hive (HiveSensorOperator), or one that moves data from Hive to MySQL (Hive2MySqlOperator). Instances of these operators (tasks) target specific operations, running specific scripts, functions or data transfers.

This class is abstract and shouldn’t be instantiated. Instantiating a class derived from this one results in the creation of a task object, which ultimately becomes a node in DAG objects. Task dependencies should be set by using the set_upstream and/or set_downstream methods.

  • task_id (str) – a unique, meaningful id for the task

  • owner (str) – the owner of the task. Using a meaningful description (e.g. user/person/team/role name) to clarify ownership is recommended.

  • email (str | Iterable[str] | None) – the ‘to’ email address(es) used in email alerts. This can be a single email or multiple ones. Multiple addresses can be specified as a comma or semicolon separated string or by passing a list of strings.

  • email_on_retry (bool) – Indicates whether email alerts should be sent when a task is retried

  • email_on_failure (bool) – Indicates whether email alerts should be sent when a task failed

  • retries (int | None) – the number of retries that should be performed before failing the task

  • retry_delay (datetime.timedelta | float) – delay between retries, can be set as timedelta or float seconds, which will be converted into timedelta, the default is timedelta(seconds=300).

  • retry_exponential_backoff (bool) – allow progressively longer waits between retries by using exponential backoff algorithm on retry delay (delay will be converted into seconds)

  • max_retry_delay (datetime.timedelta | float | None) – maximum delay interval between retries, can be set as timedelta or float seconds, which will be converted into timedelta.

  • start_date (datetime.datetime | None) – The start_date for the task, determines the execution_date for the first task instance. The best practice is to have the start_date rounded to your DAG’s schedule_interval. Daily jobs have their start_date some day at 00:00:00, hourly jobs have their start_date at 00:00 of a specific hour. Note that Airflow simply looks at the latest execution_date and adds the schedule_interval to determine the next execution_date. It is also very important to note that different tasks’ dependencies need to line up in time. If task A depends on task B and their start_date are offset in a way that their execution_date don’t line up, A’s dependencies will never be met. If you are looking to delay a task, for example running a daily task at 2AM, look into the TimeSensor and TimeDeltaSensor. We advise against using dynamic start_date and recommend using fixed ones. Read the FAQ entry about start_date for more information.

  • end_date (datetime.datetime | None) – if specified, the scheduler won’t go beyond this date

  • depends_on_past (bool) – when set to true, task instances will run sequentially and only if the previous instance has succeeded or has been skipped. The task instance for the start_date is allowed to run.

  • wait_for_past_depends_before_skipping (bool) – when set to true, if the task instance should be marked as skipped, and depends_on_past is true, the ti will stay on None state waiting the task of the previous run

  • wait_for_downstream (bool) – when set to true, an instance of task X will wait for tasks immediately downstream of the previous instance of task X to finish successfully or be skipped before it runs. This is useful if the different instances of a task X alter the same asset, and this asset is used by tasks downstream of task X. Note that depends_on_past is forced to True wherever wait_for_downstream is used. Also note that only tasks immediately downstream of the previous task instance are waited for; the statuses of any tasks further downstream are ignored.

  • dag (airflow.models.dag.DAG | None) – a reference to the dag the task is attached to (if any)

  • priority_weight (int) – priority weight of this task against other task. This allows the executor to trigger higher priority tasks before others when things get backed up. Set priority_weight as a higher number for more important tasks.

  • weight_rule (str) – weighting method used for the effective total priority weight of the task. Options are: { downstream | upstream | absolute } default is downstream When set to downstream the effective weight of the task is the aggregate sum of all downstream descendants. As a result, upstream tasks will have higher weight and will be scheduled more aggressively when using positive weight values. This is useful when you have multiple dag run instances and desire to have all upstream tasks to complete for all runs before each dag can continue processing downstream tasks. When set to upstream the effective weight is the aggregate sum of all upstream ancestors. This is the opposite where downstream tasks have higher weight and will be scheduled more aggressively when using positive weight values. This is useful when you have multiple dag run instances and prefer to have each dag complete before starting upstream tasks of other dags. When set to absolute, the effective weight is the exact priority_weight specified without additional weighting. You may want to do this when you know exactly what priority weight each task should have. Additionally, when set to absolute, there is bonus effect of significantly speeding up the task creation process as for very large DAGs. Options can be set as string or using the constants defined in the static class airflow.utils.WeightRule

  • queue (str) – which queue to target when running this job. Not all executors implement queue management, the CeleryExecutor does support targeting specific queues.

  • pool (str | None) – the slot pool this task should run in, slot pools are a way to limit concurrency for certain tasks

  • pool_slots (int) – the number of pool slots this task should use (>= 1) Values less than 1 are not allowed.

  • sla (datetime.timedelta | None) – time by which the job is expected to succeed. Note that this represents the timedelta after the period is closed. For example if you set an SLA of 1 hour, the scheduler would send an email soon after 1:00AM on the 2016-01-02 if the 2016-01-01 instance has not succeeded yet. The scheduler pays special attention for jobs with an SLA and sends alert emails for SLA misses. SLA misses are also recorded in the database for future reference. All tasks that share the same SLA time get bundled in a single email, sent soon after that time. SLA notification are sent once and only once for each task instance.

  • execution_timeout (datetime.timedelta | None) – max time allowed for the execution of this task instance, if it goes beyond it will raise and fail.

  • on_failure_callback (None | airflow.models.abstractoperator.TaskStateChangeCallback | list[airflow.models.abstractoperator.TaskStateChangeCallback]) – a function or list of functions to be called when a task instance of this task fails. a context dictionary is passed as a single parameter to this function. Context contains references to related objects to the task instance and is documented under the macros section of the API.

  • on_execute_callback (None | airflow.models.abstractoperator.TaskStateChangeCallback | list[airflow.models.abstractoperator.TaskStateChangeCallback]) – much like the on_failure_callback except that it is executed right before the task is executed.

  • on_retry_callback (None | airflow.models.abstractoperator.TaskStateChangeCallback | list[airflow.models.abstractoperator.TaskStateChangeCallback]) – much like the on_failure_callback except that it is executed when retries occur.

  • on_success_callback (None | airflow.models.abstractoperator.TaskStateChangeCallback | list[airflow.models.abstractoperator.TaskStateChangeCallback]) – much like the on_failure_callback except that it is executed when the task succeeds.

  • pre_execute (TaskPreExecuteHook | None) –

    a function to be called immediately before task execution, receiving a context dictionary; raising an exception will prevent the task from being executed.

    This is an experimental feature.

  • post_execute (TaskPostExecuteHook | None) –

    a function to be called immediately after task execution, receiving a context dictionary and task result; raising an exception will prevent the task from succeeding.

    This is an experimental feature.

  • trigger_rule (str) – defines the rule by which dependencies are applied for the task to get triggered. Options are: { all_success | all_failed | all_done | all_skipped | one_success | one_done | one_failed | none_failed | none_failed_min_one_success | none_skipped | always} default is all_success. Options can be set as string or using the constants defined in the static class airflow.utils.TriggerRule

  • resources (dict[str, Any] | None) – A map of resource parameter names (the argument names of the Resources constructor) to their values.

  • run_as_user (str | None) – unix username to impersonate while running the task

  • max_active_tis_per_dag (int | None) – When set, a task will be able to limit the concurrent runs across execution_dates.

  • max_active_tis_per_dagrun (int | None) – When set, a task will be able to limit the concurrent task instances per DAG run.

  • executor_config (dict | None) –

    Additional task-level configuration parameters that are interpreted by a specific executor. Parameters are namespaced by the name of executor.

    Example: to run this task in a specific docker container through the KubernetesExecutor

                {"image": "myCustomDockerImage"}

  • do_xcom_push (bool) – if True, an XCom is pushed containing the Operator’s result

  • task_group (airflow.utils.task_group.TaskGroup | None) – The TaskGroup to which the task should belong. This is typically provided when not using a TaskGroup as a context manager.

  • doc (str | None) – Add documentation or notes to your Task objects that is visible in Task Instance details View in the Webserver

  • doc_md (str | None) – Add documentation (in Markdown format) or notes to your Task objects that is visible in Task Instance details View in the Webserver

  • doc_rst (str | None) – Add documentation (in RST format) or notes to your Task objects that is visible in Task Instance details View in the Webserver

  • doc_json (str | None) – Add documentation (in JSON format) or notes to your Task objects that is visible in Task Instance details View in the Webserver

  • doc_yaml (str | None) – Add documentation (in YAML format) or notes to your Task objects that is visible in Task Instance details View in the Webserver

property dag: airflow.models.dag.DAG[source]

Returns the Operator’s DAG if set, otherwise raises an error.

property operator_class: type[BaseOperator][source]
property task_type: str[source]

@property: type of the task.

property operator_name: str[source]

@property: use a more friendly display name for the operator, if set.

property roots: list[BaseOperator][source]

Required by DAGNode.

property leaves: list[BaseOperator][source]

Required by DAGNode.

property output: airflow.models.xcom_arg.XComArg[source]

Returns reference to XCom pushed by current operator.

property inherits_from_empty_operator[source]

Used to determine if an Operator is inherited from EmptyOperator.

template_fields: Sequence[str] = ()[source]
template_ext: Sequence[str] = ()[source]
template_fields_renderers: dict[str, str][source]
ui_color: str = '#fff'[source]
ui_fgcolor: str = '#000'[source]
pool: str = ''[source]
shallow_copy_attrs: Sequence[str] = ()[source]
partial: Callable[Ellipsis, airflow.models.mappedoperator.OperatorPartial][source]
supports_lineage = False[source]
task_group: airflow.utils.task_group.TaskGroup | None[source]
subdag: airflow.models.dag.DAG | None[source]
start_date: pendulum.DateTime | None[source]
end_date: pendulum.DateTime | None[source]
deps: frozenset[airflow.ti_deps.deps.base_ti_dep.BaseTIDep][source]

Returns the set of dependencies for the operator. These differ from execution context dependencies in that they are specific to tasks and can be extended/overridden by subclasses.


Return self==value.


Return self!=value.


Return hash(self).


Return [This Operator] | [Operator].

The inlets of other will be set to pick up the outlets from this operator. Other will be set as a downstream task of this operator.


Return [Operator] > [Outlet].

If other is an attr annotated object it is set as an outlet of this Operator.


Return [Inlet] > [Operator] or [Operator] < [Inlet].

If other is an attr annotated object it is set as an inlet to this operator.

__setattr__(key, value)[source]

Implement setattr(self, name, value).


Set inlets to this operator.


Define the outlets of this operator.


Return True if the Operator has been assigned to a DAG.


Lock task for execution to disable custom action in __setattr__ and return a copy.


Resolve upstream dependencies of a task.

In this way passing an XComArg as value for a template field will result in creating upstream relation between two tasks.


with DAG(...):
    generate_content = GenerateContentOperator(task_id="generate_content")
    send_email = EmailOperator(..., html_content=generate_content.output)

# This is equivalent to
with DAG(...):
    generate_content = GenerateContentOperator(task_id="generate_content")
    send_email = EmailOperator(
        ..., html_content="{{ task_instance.xcom_pull('generate_content') }}"
    generate_content >> send_email

Execute right before self.execute() is called.

abstract execute(context)[source]

Derive when creating an operator.

Context is the same dictionary used as when rendering jinja templates.

Refer to get_template_context for more context.

post_execute(context, result=None)[source]

Execute right after self.execute() is called.

It is passed the execution context and any results returned by the operator.


Override this method to clean up subprocesses when a task instance gets killed.

Any use of the threading, subprocess or multiprocessing module within an operator needs to be cleaned up, or it will leave ghost processes behind.

render_template_fields(context, jinja_env=None)[source]

Template all attributes listed in self.template_fields.

This mutates the attributes in-place and is irreversible.

  • context (airflow.utils.context.Context) – Context dict with values to apply on content.

  • jinja_env (jinja2.Environment | None) – Jinja’s environment to use for rendering.

clear(start_date=None, end_date=None, upstream=False, downstream=False, session=NEW_SESSION)[source]

Clear the state of task instances associated with the task, following the parameters specified.

get_task_instances(start_date=None, end_date=None, session=NEW_SESSION)[source]

Get task instances related to this task for a specific date range.

run(start_date=None, end_date=None, ignore_first_depends_on_past=True, wait_for_past_depends_before_skipping=False, ignore_ti_state=False, mark_success=False, test_mode=False, session=NEW_SESSION)[source]

Run a set of task instances for a date range.


Perform dry run for the operator - just render template fields.


Get list of the direct relatives to the current task, upstream or downstream.


Return repr(self).

static xcom_push(context, key, value, execution_date=None)[source]

Make an XCom available for tasks to pull.

  • context (Any) – Execution Context Dictionary

  • key (str) – A key for the XCom

  • value (Any) – A value for the XCom. The value is pickled and stored in the database.

  • execution_date (datetime.datetime | None) – if provided, the XCom will not be visible until this date. This can be used, for example, to send a message to a task on a future date without it being immediately visible.

static xcom_pull(context, task_ids=None, dag_id=None, key=XCOM_RETURN_KEY, include_prior_dates=None, session=NEW_SESSION)[source]

Pull XComs that optionally meet certain criteria.

The default value for key limits the search to XComs that were returned by other tasks (as opposed to those that were pushed manually). To remove this filter, pass key=None (or any desired value).

If a single task_id string is provided, the result is the value of the most recent matching XCom from that task_id. If multiple task_ids are provided, a tuple of matching values is returned. None is returned whenever no matches are found.

  • context (Any) – Execution Context Dictionary

  • key (str) – A key for the XCom. If provided, only XComs with matching keys will be returned. The default key is ‘return_value’, also available as a constant XCOM_RETURN_KEY. This key is automatically given to XComs returned by tasks (as opposed to being pushed manually). To remove the filter, pass key=None.

  • task_ids (str | list[str] | None) – Only XComs from tasks with matching ids will be pulled. Can pass None to remove the filter.

  • dag_id (str | None) – If provided, only pulls XComs from this DAG. If None (default), the DAG of the calling task is used.

  • include_prior_dates (bool | None) – If False, only XComs from the current execution_date are returned. If True, XComs from previous dates are returned as well.

classmethod get_serialized_fields()[source]

Stringified DAGs and operators contain exactly these fields.


Serialize; required by DAGNode.

defer(*, trigger, method_name, kwargs=None, timeout=None)[source]

Mark this Operator “deferred”, suspending its execution until the provided trigger fires an event.

This is achieved by raising a special exception (TaskDeferred) which is caught in the main _execute_task wrapper.

resume_execution(next_method, next_kwargs, context)[source]

This method is called when a deferred task is resumed.


Given a number of tasks, builds a dependency chain.

This function accepts values of BaseOperator (aka tasks), EdgeModifiers (aka Labels), XComArg, TaskGroups, or lists containing any mix of these types (or a mix in the same list). If you want to chain between two lists you must ensure they have the same length.

Using classic operators/sensors:

chain(t1, [t2, t3], [t4, t5], t6)

is equivalent to:

  / -> t2 -> t4 \
t1               -> t6
  \ -> t3 -> t5 /

Using task-decorated functions aka XComArgs:

chain(x1(), [x2(), x3()], [x4(), x5()], x6())

is equivalent to:

  / -> x2 -> x4 \
x1               -> x6
  \ -> x3 -> x5 /
x1 = x1()
x2 = x2()
x3 = x3()
x4 = x4()
x5 = x5()
x6 = x6()

Using TaskGroups:

chain(t1, task_group1, task_group2, t2)


It is also possible to mix between classic operator/sensor, EdgeModifiers, XComArg, and TaskGroups:

chain(t1, [Label("branch one"), Label("branch two")], [x1(), x2()], task_group1, x3())

is equivalent to:

  / "branch one" -> x1 \
t1                      -> task_group1 -> x3
  \ "branch two" -> x2 /
x1 = x1()
x2 = x2()
x3 = x3()
label1 = Label("branch one")
label2 = Label("branch two")

# or

x1 = x1()
x2 = x2()
x3 = x3()
t1.set_downstream(x1, edge_modifier=Label("branch one"))
t1.set_downstream(x2, edge_modifier=Label("branch two"))

tasks (airflow.models.taskmixin.DependencyMixin | Sequence[airflow.models.taskmixin.DependencyMixin]) – Individual and/or list of tasks, EdgeModifiers, XComArgs, or TaskGroups to set dependencies

airflow.models.baseoperator.cross_downstream(from_tasks, to_tasks)[source]

Set downstream dependencies for all tasks in from_tasks to all tasks in to_tasks.

Using classic operators/sensors:

cross_downstream(from_tasks=[t1, t2, t3], to_tasks=[t4, t5, t6])

is equivalent to:

t1 ---> t4
   \ /
t2 -X -> t5
   / \
t3 ---> t6

Using task-decorated functions aka XComArgs:

cross_downstream(from_tasks=[x1(), x2(), x3()], to_tasks=[x4(), x5(), x6()])

is equivalent to:

x1 ---> x4
   \ /
x2 -X -> x5
   / \
x3 ---> x6
x1 = x1()
x2 = x2()
x3 = x3()
x4 = x4()
x5 = x5()
x6 = x6()

It is also possible to mix between classic operator/sensor and XComArg tasks:

cross_downstream(from_tasks=[t1, x2(), t3], to_tasks=[x1(), t2, x3()])

is equivalent to:

t1 ---> x1
   \ /
x2 -X -> t2
   / \
t3 ---> x3
x1 = x1()
x2 = x2()
x3 = x3()
  • from_tasks (Sequence[airflow.models.taskmixin.DependencyMixin]) – List of tasks or XComArgs to start from.

  • to_tasks (airflow.models.taskmixin.DependencyMixin | Sequence[airflow.models.taskmixin.DependencyMixin]) – List of tasks or XComArgs to set as downstream dependencies.


Simplify task dependency definition.

E.g.: suppose you want precedence like so:

    ╭─op2─╮ ╭─op4─╮
op1─┤     ├─├─op5─┤─op7
    ╰-op3─╯ ╰-op6─╯

Then you can accomplish like so:

    [op2, op3],
    [op4, op5, op6],

elements (airflow.models.taskmixin.DependencyMixin | Sequence[airflow.models.taskmixin.DependencyMixin]) – a list of operators / lists of operators


Abstract base class that defines how we get an operator link.

abstract property name: str[source]

Name of the link. This will be the button name on the task UI.

operators: ClassVar[BaseOperatorClassList] = [][source]

This property will be used by Airflow Plugins to find the Operators to which you want to assign this Operator Link


List of Operator classes used by task for which you want to create extra link

Link to external system.

Note: The old signature of this function was (self, operator, dttm: datetime). That is still supported at runtime but is deprecated.


link to external system

Return type


Was this entry helpful?