#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
"""This module contains Google BigQuery sensors."""
from __future__ import annotations
import warnings
from collections.abc import Sequence
from datetime import timedelta
from typing import TYPE_CHECKING, Any
from google.api_core.exceptions import NotFound
from google.cloud.bigquery import DatasetReference, TableReference
from airflow.exceptions import AirflowProviderDeprecationWarning
from airflow.providers.common.compat.sdk import AirflowException, BaseSensorOperator, conf, timezone
from airflow.providers.google.cloud.hooks.bigquery import BigQueryHook
from airflow.providers.google.cloud.triggers.bigquery import (
BigQueryStreamingBufferEmptyTrigger,
BigQueryTableExistenceTrigger,
BigQueryTablePartitionExistenceTrigger,
)
from airflow.providers.google.version_compat import AIRFLOW_V_3_0_PLUS
if TYPE_CHECKING:
from datetime import datetime
from airflow.providers.common.compat.sdk import Context
[docs]
class BigQueryTableExistenceSensor(BaseSensorOperator):
"""
Checks for the existence of a table in Google Bigquery.
:param project_id: The Google cloud project in which to look for the table.
The connection supplied to the hook must provide
access to the specified project.
:param dataset_id: The name of the dataset in which to look for the table.
storage bucket.
:param table_id: The name of the table to check the existence of.
:param gcp_conn_id: (Optional) The connection ID used to connect to Google Cloud.
:param impersonation_chain: Optional service account to impersonate using short-term
credentials, or chained list of accounts required to get the access_token
of the last account in the list, which will be impersonated in the request.
If set as a string, the account must grant the originating account
the Service Account Token Creator IAM role.
If set as a sequence, the identities from the list must grant
Service Account Token Creator IAM role to the directly preceding identity, with first
account from the list granting this role to the originating account (templated).
"""
[docs]
template_fields: Sequence[str] = (
"project_id",
"dataset_id",
"table_id",
"impersonation_chain",
"gcp_conn_id",
)
def __init__(
self,
*,
project_id: str,
dataset_id: str,
table_id: str,
gcp_conn_id: str = "google_cloud_default",
impersonation_chain: str | Sequence[str] | None = None,
deferrable: bool = conf.getboolean("operators", "default_deferrable", fallback=False),
**kwargs,
) -> None:
if deferrable and "poke_interval" not in kwargs:
# TODO: Remove once deprecated
if "polling_interval" in kwargs:
kwargs["poke_interval"] = kwargs["polling_interval"]
warnings.warn(
"Argument `poll_interval` is deprecated and will be removed "
"in a future release. Please use `poke_interval` instead. Planned removal date: October 5, 2026.",
AirflowProviderDeprecationWarning,
stacklevel=2,
)
else:
kwargs["poke_interval"] = 5
super().__init__(**kwargs)
[docs]
self.project_id = project_id
[docs]
self.dataset_id = dataset_id
[docs]
self.table_id = table_id
[docs]
self.gcp_conn_id = gcp_conn_id
[docs]
self.impersonation_chain = impersonation_chain
[docs]
self.deferrable = deferrable
[docs]
def poke(self, context: Context) -> bool:
table_uri = f"{self.project_id}:{self.dataset_id}.{self.table_id}"
self.log.info("Sensor checks existence of table: %s", table_uri)
hook = BigQueryHook(
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
)
return hook.table_exists(
project_id=self.project_id, dataset_id=self.dataset_id, table_id=self.table_id
)
[docs]
def execute(self, context: Context) -> None:
"""Airflow runs this method on the worker and defers using the trigger."""
if not self.deferrable:
super().execute(context)
else:
if not self.poke(context=context):
self.defer(
timeout=timedelta(seconds=self.timeout),
trigger=BigQueryTableExistenceTrigger(
dataset_id=self.dataset_id,
table_id=self.table_id,
project_id=self.project_id,
poll_interval=self.poke_interval,
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
hook_params={
"impersonation_chain": self.impersonation_chain,
},
),
method_name="execute_complete",
)
[docs]
def execute_complete(self, context: dict[str, Any], event: dict[str, str] | None = None) -> str:
"""
Act as a callback for when the trigger fires - returns immediately.
Relies on trigger to throw an exception, otherwise it assumes execution was successful.
"""
table_uri = f"{self.project_id}:{self.dataset_id}.{self.table_id}"
self.log.info("Sensor checks existence of table: %s", table_uri)
if event:
if event["status"] == "success":
return event["message"]
raise AirflowException(event["message"])
message = "No event received in trigger callback"
raise AirflowException(message)
[docs]
class BigQueryRoutineExistenceSensor(BaseSensorOperator):
"""
Checks for the existence of a routine (UDF, procedure, or TVF) in a BigQuery dataset.
:param project_id: The Google Cloud project that owns the dataset.
:param dataset_id: The dataset that owns the routine.
:param routine_id: The identifier of the routine to check.
:param gcp_conn_id: (Optional) The connection ID used to connect to Google Cloud.
:param impersonation_chain: Optional service account to impersonate using short-term
credentials, or chained list of accounts required to get the access_token
of the last account in the list, which will be impersonated in the request.
If set as a string, the account must grant the originating account
the Service Account Token Creator IAM role.
If set as a sequence, the identities from the list must grant
Service Account Token Creator IAM role to the directly preceding identity, with first
account from the list granting this role to the originating account (templated).
"""
[docs]
template_fields: Sequence[str] = (
"project_id",
"dataset_id",
"routine_id",
"impersonation_chain",
"gcp_conn_id",
)
def __init__(
self,
*,
project_id: str,
dataset_id: str,
routine_id: str,
gcp_conn_id: str = "google_cloud_default",
impersonation_chain: str | Sequence[str] | None = None,
**kwargs,
) -> None:
super().__init__(**kwargs)
[docs]
self.project_id = project_id
[docs]
self.dataset_id = dataset_id
[docs]
self.routine_id = routine_id
[docs]
self.gcp_conn_id = gcp_conn_id
[docs]
self.impersonation_chain = impersonation_chain
[docs]
def poke(self, context: Context) -> bool:
routine_uri = f"{self.project_id}.{self.dataset_id}.{self.routine_id}"
self.log.info("Sensor checks existence of routine: %s", routine_uri)
hook = BigQueryHook(
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
)
try:
hook.get_routine(
project_id=self.project_id,
dataset_id=self.dataset_id,
routine_id=self.routine_id,
)
except NotFound:
return False
return True
[docs]
class BigQueryTablePartitionExistenceSensor(BaseSensorOperator):
"""
Checks for the existence of a partition within a table in Google Bigquery.
:param project_id: The Google cloud project in which to look for the table.
The connection supplied to the hook must provide
access to the specified project.
:param dataset_id: The name of the dataset in which to look for the table.
storage bucket.
:param table_id: The name of the table to check the existence of.
:param partition_id: The name of the partition to check the existence of.
:param gcp_conn_id: (Optional) The connection ID used to connect to Google Cloud.
:param impersonation_chain: Optional service account to impersonate using short-term
credentials, or chained list of accounts required to get the access_token
of the last account in the list, which will be impersonated in the request.
If set as a string, the account must grant the originating account
the Service Account Token Creator IAM role.
If set as a sequence, the identities from the list must grant
Service Account Token Creator IAM role to the directly preceding identity, with first
account from the list granting this role to the originating account (templated).
"""
[docs]
template_fields: Sequence[str] = (
"project_id",
"dataset_id",
"table_id",
"partition_id",
"impersonation_chain",
"gcp_conn_id",
)
def __init__(
self,
*,
project_id: str,
dataset_id: str,
table_id: str,
partition_id: str,
gcp_conn_id: str = "google_cloud_default",
impersonation_chain: str | Sequence[str] | None = None,
deferrable: bool = conf.getboolean("operators", "default_deferrable", fallback=False),
**kwargs,
) -> None:
if deferrable and "poke_interval" not in kwargs:
kwargs["poke_interval"] = 5
super().__init__(**kwargs)
[docs]
self.project_id = project_id
[docs]
self.dataset_id = dataset_id
[docs]
self.table_id = table_id
[docs]
self.partition_id = partition_id
[docs]
self.gcp_conn_id = gcp_conn_id
[docs]
self.impersonation_chain = impersonation_chain
[docs]
self.deferrable = deferrable
[docs]
def poke(self, context: Context) -> bool:
table_uri = f"{self.project_id}:{self.dataset_id}.{self.table_id}"
self.log.info('Sensor checks existence of partition: "%s" in table: %s', self.partition_id, table_uri)
hook = BigQueryHook(
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
)
return hook.table_partition_exists(
project_id=self.project_id,
dataset_id=self.dataset_id,
table_id=self.table_id,
partition_id=self.partition_id,
)
[docs]
def execute(self, context: Context) -> None:
"""Airflow runs this method on the worker and defers using the triggers if deferrable is True."""
if not self.deferrable:
super().execute(context)
else:
if not self.poke(context=context):
self.defer(
timeout=timedelta(seconds=self.timeout),
trigger=BigQueryTablePartitionExistenceTrigger(
dataset_id=self.dataset_id,
table_id=self.table_id,
project_id=self.project_id,
partition_id=self.partition_id,
poll_interval=self.poke_interval,
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
hook_params={
"impersonation_chain": self.impersonation_chain,
},
),
method_name="execute_complete",
)
[docs]
def execute_complete(self, context: dict[str, Any], event: dict[str, str] | None = None) -> str:
"""
Act as a callback for when the trigger fires - returns immediately.
Relies on trigger to throw an exception, otherwise it assumes execution was successful.
"""
table_uri = f"{self.project_id}:{self.dataset_id}.{self.table_id}"
self.log.info('Sensor checks existence of partition: "%s" in table: %s', self.partition_id, table_uri)
if event:
if event["status"] == "success":
return event["message"]
raise AirflowException(event["message"])
message = "No event received in trigger callback"
raise AirflowException(message)
[docs]
class BigQueryStreamingBufferEmptySensor(BaseSensorOperator):
"""
Wait for the streaming buffer of a BigQuery table to be empty.
BigQuery DML statements (UPDATE, DELETE, MERGE) cannot run against rows that
are still in the streaming buffer; the buffer is flushed within ~90 minutes.
Use this sensor between a streaming insert and a DML step to avoid
``UPDATE/MERGE/DELETE statement over table ... would affect rows in the
streaming buffer`` errors.
.. warning::
BigQuery's ``table.streaming_buffer`` metadata is eventually consistent.
The sensor mitigates this by requiring multiple consecutive empty
observations before reporting the buffer empty (configurable via
``empty_confirmations``), but this may delay success by up to one
``poke_interval`` compared to earlier releases. ``mode="reschedule"`` with
``empty_confirmations > 1`` needs a running triggerer once an empty reading is seen;
``empty_confirmations=1`` keeps the worker-only behaviour.
:param project_id: Google Cloud project containing the table.
:param dataset_id: Dataset of the table to monitor.
:param table_id: Table to monitor.
:param gcp_conn_id: Airflow connection ID for GCP.
:param impersonation_chain: Optional service account to impersonate, or a
chained list of accounts. See the Google provider docs for details.
:param empty_confirmations: Number of consecutive empty readings required
before considering the streaming buffer empty. Must be at least 1.
Defaults to 2.
:param deferrable: Run in deferrable mode using
:class:`BigQueryStreamingBufferEmptyTrigger`.
"""
[docs]
template_fields: Sequence[str] = (
"project_id",
"dataset_id",
"table_id",
"impersonation_chain",
"gcp_conn_id",
)
def __init__(
self,
*,
project_id: str,
dataset_id: str,
table_id: str,
gcp_conn_id: str = "google_cloud_default",
impersonation_chain: str | Sequence[str] | None = None,
empty_confirmations: int = 2,
deferrable: bool = conf.getboolean("operators", "default_deferrable", fallback=False),
**kwargs,
) -> None:
if deferrable and "poke_interval" not in kwargs:
kwargs["poke_interval"] = 30
super().__init__(**kwargs)
if empty_confirmations < 1:
raise ValueError("empty_confirmations must be at least 1")
[docs]
self.project_id = project_id
[docs]
self.dataset_id = dataset_id
[docs]
self.table_id = table_id
[docs]
self.gcp_conn_id = gcp_conn_id
[docs]
self.impersonation_chain = impersonation_chain
[docs]
self.empty_confirmations = empty_confirmations
[docs]
self.deferrable = deferrable
self._consecutive_empty = 0
[docs]
def execute(self, context: Context) -> None:
if not self.deferrable:
super().execute(context)
return
if self.poke(context=context):
return
self._defer_to_trigger(timeout=self.timeout)
def _defer_to_trigger(self, timeout: float) -> None:
self.defer(
timeout=timedelta(seconds=timeout),
trigger=BigQueryStreamingBufferEmptyTrigger(
project_id=self.project_id,
dataset_id=self.dataset_id,
table_id=self.table_id,
poll_interval=self.poke_interval,
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
empty_confirmations=self.empty_confirmations,
),
method_name="execute_complete",
)
[docs]
def execute_complete(self, context: dict[str, Any], event: dict[str, str] | None = None) -> str:
if event is None:
raise ValueError("No event received in trigger callback")
if event["status"] == "success":
return event["message"]
raise RuntimeError(event["message"])
[docs]
def poke(self, context: Context) -> bool:
table_uri = f"{self.project_id}:{self.dataset_id}.{self.table_id}"
self.log.info("Checking streaming buffer state for table: %s", table_uri)
# ``Client.get_table`` only accepts a ``TableReference`` or a standard-SQL
# ``project.dataset.table`` string -- not the legacy ``project:dataset.table``
# form -- so build an explicit reference here.
table_ref = TableReference(
dataset_ref=DatasetReference(self.project_id, self.dataset_id),
table_id=self.table_id,
)
hook = BigQueryHook(gcp_conn_id=self.gcp_conn_id, impersonation_chain=self.impersonation_chain)
try:
table = hook.get_client(project_id=self.project_id).get_table(table_ref)
except NotFound as err:
raise ValueError(f"Table {table_uri} not found") from err
if table.streaming_buffer is not None:
self._consecutive_empty = 0
return False
self._consecutive_empty += 1
self.log.info(
"Streaming buffer reported empty (%s/%s confirmations) for table: %s",
self._consecutive_empty,
self.empty_confirmations,
table_uri,
)
if self._consecutive_empty >= self.empty_confirmations:
return True
if self.reschedule and not self.deferrable:
# A reschedule rebuilds the operator and Airflow 3 clears XComs on each run start, so the
# confirmation count cannot survive it; the trigger repeats the full confirmation count instead.
first_reschedule_date = self._get_first_reschedule_date(context)
elapsed = (
(timezone.utcnow() - first_reschedule_date).total_seconds() if first_reschedule_date else 0
)
if self.timeout - elapsed > 0:
self._defer_to_trigger(timeout=self.timeout - elapsed)
return False
def _get_first_reschedule_date(self, context: Context) -> datetime | None:
if AIRFLOW_V_3_0_PLUS:
return context["ti"].get_first_reschedule_date(context)
# Same lookup as the Airflow 2.11 BaseSensorOperator.execute; `_orig_start_date` only exists in Airflow 2.
from airflow.sensors.base import _orig_start_date # type: ignore[attr-defined]
ti = context["ti"]
return _orig_start_date(
dag_id=ti.dag_id,
task_id=ti.task_id,
run_id=ti.run_id,
map_index=ti.map_index,
try_number=(ti.max_tries or 0) - (self.retries or 0) + 1,
)