Prefect Docs

# Databricks Tasks

Verified by Prefect

These tasks have been tested and verified by Prefect.

This module contains a collection of tasks for interacting with Databricks resources.

# DatabricksSubmitRun

class

prefect.tasks.databricks.databricks_submitjob.DatabricksSubmitRun

(databricks_conn_secret=None, json=None, spark_jar_task=None, notebook_task=None, new_cluster=None, existing_cluster_id=None, libraries=None, run_name=None, timeout_seconds=None, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1, **kwargs)[source]

Submits a Spark job run to Databricks using the api/2.0/jobs/runs/submit API endpoint.

Currently the named parameters that DatabricksSubmitRun task supports are

  • spark_jar_task - notebook_task - new_cluster - existing_cluster_id - libraries - run_name - timeout_seconds

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • json (dict, optional): A JSON object containing API parameters which will be passed directly to the api/2.0/jobs/runs/submit endpoint. The other named parameters (i.e. spark_jar_task, notebook_task..) to this task will be merged with this json dictionary if they are provided. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys. (templated) For more information about templating see :ref:jinja-templating. https://docs.databricks.com/api/latest/jobs.html#runs-submit
  • spark_jar_task (dict, optional): The main class and parameters for the JAR task. Note that the actual JAR is specified in the libraries. EITHER spark_jar_task OR notebook_task should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobssparkjartask
  • notebook_task (dict, optional): The notebook path and parameters for the notebook task. EITHER spark_jar_task OR notebook_task should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobsnotebooktask
  • new_cluster (dict, optional): Specs for a new cluster on which this task will be run. EITHER new_cluster OR existing_cluster_id should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobsclusterspecnewcluster
  • existing_cluster_id (str, optional): ID for existing cluster on which to run this task. EITHER new_cluster OR existing_cluster_id should be specified. This field will be templated.
  • libraries (list of dicts, optional): Libraries which this run will use. This field will be templated. https://docs.databricks.com/api/latest/libraries.html#managedlibrarieslibrary
  • run_name (str, optional): The run name used for this task. By default this will be set to the Prefect task_id. This task_id is a required parameter of the superclass Task. This field will be templated.
  • timeout_seconds (int, optional): The timeout for this run. By default a value of 0 is used which means to have no timeout. This field will be templated.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
  • **kwargs (dict, optional): additional keyword arguments to pass to the Task constructor
Examples: There are two ways to instantiate this task.

In the first way, you can take the JSON payload that you typically use to call the api/2.0/jobs/runs/submit endpoint and pass it directly to our DatabricksSubmitRun task through the json parameter.

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksSubmitRun

    json = {
        'new_cluster': {
            'spark_version': '10.4.x-scala2.12',
            'num_workers': 2,
            'node_type_id': "m4.large",
            'aws_attributes': {
                'ebs_volume_type': "GENERAL_PURPOSE_SSD",
                'ebs_volume_count': 3,
                'ebs_volume_size': 100,
            }
        },
        'notebook_task': {
            'notebook_path': '/Users/andrew.h@prefect.io/records',
        },
    }

    with Flow("my flow") as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksSubmitRun(json=json)
        notebook_run(databricks_conn_secret=conn)

Another way to accomplish the same thing is to use the named parameters of the DatabricksSubmitRun directly. Note that there is exactly one named parameter for each top level parameter in the runs/submit endpoint. In this method, your code would look like this:

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksSubmitRun

    new_cluster = {
        'spark_version': '10.4.x-scala2.12',
        'num_workers': 2,
        'node_type_id': "m4.large",
        'aws_attributes': {
            'ebs_volume_type': "GENERAL_PURPOSE_SSD",
            'ebs_volume_count': 3,
            'ebs_volume_size': 100,
        }
    }
    notebook_task = {
        'notebook_path': '/Users/prefect@example.com/PrepareData',
    }

    with Flow("my flow") as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksSubmitRun(
            new_cluster=new_cluster,
            notebook_task=notebook_task)
        notebook_run(databricks_conn_secret=conn)

In the case where both the json parameter AND the named parameters are provided, they will be merged together. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys. This task requires a Databricks connection to be specified as a Prefect secret and can be passed to the task like so:

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksSubmitRun

    with Flow("my flow") as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksSubmitRun(json=...)
        notebook_run(databricks_conn_secret=conn)
methods:                                                                                                                                                       

prefect.tasks.databricks.databricks_submitjob.DatabricksSubmitRun.run

(databricks_conn_secret=None, json=None, spark_jar_task=None, notebook_task=None, new_cluster=None, existing_cluster_id=None, libraries=None, run_name=None, timeout_seconds=None, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1)[source]

Task run method.

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based, authentication provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • json (dict, optional): A JSON object containing API parameters which will be passed directly to the api/2.0/jobs/runs/submit endpoint. The other named parameters (i.e. spark_jar_task, notebook_task..) to this task will be merged with this json dictionary if they are provided. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys. (templated) For more information about templating see :ref:jinja-templating. https://docs.databricks.com/api/latest/jobs.html#runs-submit
  • spark_jar_task (dict, optional): The main class and parameters for the JAR task. Note that the actual JAR is specified in the libraries. EITHER spark_jar_task OR notebook_task should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobssparkjartask
  • notebook_task (dict, optional): The notebook path and parameters for the notebook task. EITHER spark_jar_task OR notebook_task should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobsnotebooktask
  • new_cluster (dict, optional): Specs for a new cluster on which this task will be run. EITHER new_cluster OR existing_cluster_id should be specified. This field will be templated. https://docs.databricks.com/api/latest/jobs.html#jobsclusterspecnewcluster
  • existing_cluster_id (str, optional): ID for existing cluster on which to run this task. EITHER new_cluster OR existing_cluster_id should be specified. This field will be templated.
  • libraries (list of dicts, optional): Libraries which this run will use. This field will be templated. https://docs.databricks.com/api/latest/libraries.html#managedlibrarieslibrary
  • run_name (str, optional): The run name used for this task. By default this will be set to the Prefect task_id. This task_id is a required parameter of the superclass Task. This field will be templated.
  • timeout_seconds (int, optional): The timeout for this run. By default a value of 0 is used which means to have no timeout. This field will be templated.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
Returns:
  • run_id (str): Run id of the submitted run


# DatabricksRunNow

class

prefect.tasks.databricks.databricks_submitjob.DatabricksRunNow

(databricks_conn_secret=None, job_id=None, json=None, notebook_params=None, python_params=None, spark_submit_params=None, jar_params=None, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1, **kwargs)[source]

Runs an existing Spark job run to Databricks using the api/2.1/jobs/run-now API endpoint.

Currently the named parameters that DatabricksRunNow task supports are

  • job_id - json - notebook_params - python_params - spark_submit_params - jar_params

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • job_id (str, optional): The job_id of the existing Databricks job. https://docs.databricks.com/api/latest/jobs.html#run-now
  • json (dict, optional): A JSON object containing API parameters which will be passed directly to the api/2.0/jobs/run-now endpoint. The other named parameters (i.e. notebook_params, spark_submit_params..) to this operator will be merged with this json dictionary if they are provided. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys. (templated) https://docs.databricks.com/api/latest/jobs.html#run-now
  • notebook_params (dict, optional): A dict from keys to values for jobs with notebook task, e.g. "notebook_params": {"name": "john doe", "age": "35"}. The map is passed to the notebook and will be accessible through the dbutils.widgets.get function. See Widgets for more information. If not specified upon run-now, the triggered run will use the job’s base parameters. notebook_params cannot be specified in conjunction with jar_params. The json representation of this field (i.e. {"notebook_params":{"name":"john doe","age":"35"}}) cannot exceed 10,000 bytes. https://docs.databricks.com/user-guide/notebooks/widgets.html
  • python_params (list[str], optional): A list of parameters for jobs with python tasks, e.g. "python_params": ["john doe", "35"]. The parameters will be passed to python file as command line parameters. If specified upon run-now, it would overwrite the parameters specified in job setting. The json representation of this field (i.e. {"python_params":["john doe","35"]}) cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • spark_submit_params (list[str], optional): A list of parameters for jobs with spark submit task, e.g. "spark_submit_params": ["--class", "org.apache.spark.examples.SparkPi"]. The parameters will be passed to spark-submit script as command line parameters. If specified upon run-now, it would overwrite the parameters specified in job setting. The json representation of this field cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • jar_params (list[str], optional): A list of parameters for jobs with JAR tasks, e.g. "jar_params": ["john doe", "35"]. The parameters will be used to invoke the main function of the main class specified in the Spark JAR task. If not specified upon run-now, it will default to an empty list. jar_params cannot be specified in conjunction with notebook_params. The JSON representation of this field (i.e. {"jar_params":["john doe","35"]}) cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • timeout_seconds (int, optional): The timeout for this run. By default a value of 0 is used which means to have no timeout. This field will be templated.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
  • **kwargs (dict, optional): additional keyword arguments to pass to the Task constructor
Examples: There are two ways to instantiate this task.

In the first way, you can take the JSON payload that you typically use to call the api/2.1/jobs/run-now endpoint and pass it directly to our DatabricksRunNow task through the json parameter.

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksRunNow

    json = {
        "job_id": 42,
        "notebook_params": {
            "dry-run": "true",
            "oldest-time-to-consider": "1457570074236"
        }
    }

    with Flow("my flow") as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksRunNow(json=json)
        notebook_run(databricks_conn_secret=conn)

Another way to accomplish the same thing is to use the named parameters of the DatabricksRunNow task directly. Note that there is exactly one named parameter for each top level parameter in the run-now endpoint. In this method, your code would look like this:

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksRunNow

    job_id = 42

    notebook_params = {
        "dry-run": "true",
        "oldest-time-to-consider": "1457570074236"
    }

    python_params = ["douglas adams", "42"]
    spark_submit_params = ["--class", "org.apache.spark.examples.SparkPi"]
    jar_params = ["john doe","35"]

    with Flow("my flow') as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksRunNow(
            notebook_params=notebook_params,
            python_params=python_params,
            spark_submit_params=spark_submit_params,
            jar_params=jar_params
        )
        notebook_run(databricks_conn_secret=conn)

In the case where both the json parameter AND the named parameters are provided, they will be merged together. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys.

This task requires a Databricks connection to be specified as a Prefect secret and can be passed to the task like so:

    from prefect import Flow
    from prefect.tasks.secrets import PrefectSecret
    from prefect.tasks.databricks import DatabricksRunNow

    with Flow("my flow") as flow:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        notebook_run = DatabricksRunNow(json=...)
        notebook_run(databricks_conn_secret=conn)
methods:                                                                                                                                                       

prefect.tasks.databricks.databricks_submitjob.DatabricksRunNow.run

(databricks_conn_secret=None, job_id=None, json=None, notebook_params=None, python_params=None, spark_submit_params=None, jar_params=None, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1)[source]

Task run method.

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • job_id (str, optional): The job_id of the existing Databricks job. https://docs.databricks.com/api/latest/jobs.html#run-now
  • json (dict, optional): A JSON object containing API parameters which will be passed directly to the api/2.0/jobs/run-now endpoint. The other named parameters (i.e. notebook_params, spark_submit_params..) to this operator will be merged with this json dictionary if they are provided. If there are conflicts during the merge, the named parameters will take precedence and override the top level json keys. (templated) https://docs.databricks.com/api/latest/jobs.html#run-now
  • notebook_params (dict, optional): A dict from keys to values for jobs with notebook task, e.g. "notebook_params": {"name": "john doe", "age": "35"}. The map is passed to the notebook and will be accessible through the dbutils.widgets.get function. See Widgets for more information. If not specified upon run-now, the triggered run will use the job’s base parameters. notebook_params cannot be specified in conjunction with jar_params. The json representation of this field (i.e. {"notebook_params":{"name":"john doe","age":"35"}}) cannot exceed 10,000 bytes. https://docs.databricks.com/user-guide/notebooks/widgets.html
  • python_params (list[str], optional): A list of parameters for jobs with python tasks, e.g. "python_params": ["john doe", "35"]. The parameters will be passed to python file as command line parameters. If specified upon run-now, it would overwrite the parameters specified in job setting. The json representation of this field (i.e. {"python_params":["john doe","35"]}) cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • spark_submit_params (list[str], optional): A list of parameters for jobs with spark submit task, e.g. "spark_submit_params": ["--class", "org.apache.spark.examples.SparkPi"]. The parameters will be passed to spark-submit script as command line parameters. If specified upon run-now, it would overwrite the parameters specified in job setting. The json representation of this field cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • jar_params (list[str], optional): A list of parameters for jobs with JAR tasks, e.g. "jar_params": ["john doe", "35"]. The parameters will be used to invoke the main function of the main class specified in the Spark JAR task. If not specified upon run-now, it will default to an empty list. jar_params cannot be specified in conjunction with notebook_params. The JSON representation of this field (i.e. {"jar_params":["john doe","35"]}) cannot exceed 10,000 bytes. https://docs.databricks.com/api/latest/jobs.html#run-now
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
Returns:
  • run_id (str): Run id of the submitted run


# DatabricksSubmitMultitaskRun

class

prefect.tasks.databricks.databricks_submitjob.DatabricksSubmitMultitaskRun

(databricks_conn_secret=None, tasks=None, run_name=None, timeout_seconds=None, idempotency_token=None, access_control_list=None, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1, git_source=None, **kwargs)[source]

Creates and triggers a one-time run via the Databricks submit run API endpoint. Supports the execution of multiple Databricks tasks within the Databricks job run. Note: Databricks tasks are distinct from Prefect tasks. All tasks configured will run as a single Prefect task.

For more information about the arguments of this task, refer to the Databricks submit run API documentation

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}'
  • tasks (List[JobTaskSettings]): A list containing the Databricks task configuration. Should contain configuration for at least one task.
  • timeout_seconds (int, optional): An optional timeout applied to each run of this job. The default behavior is to have no timeout.
  • run_name (str, optional): An optional name for the run. The default value is "Job run created by Prefect flow run {flow_run_name}".
  • idempotency_token (str, optional): An optional token that can be used to guarantee the idempotency of job run requests. Defaults to the Python object id.
  • access_control_list (List[AccessControlRequest]): List of permissions to set on the job.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
  • git_source (GitSource): A git source for the source code of the jobs (see https://databricks.com/blog/2022/06/21/build-reliable-production-data-and-ml-pipelines-with-git-support-for-databricks-workflows.html)
  • **kwargs (dict, optional): Additional keyword arguments to pass to the Task constructor
Examples: Trigger an ad-hoc multitask run

    from prefect import Flow
    from prefect.tasks.databricks import DatabricksSubmitMultitaskRun
    from prefect.tasks.databricks.models import (
        AccessControlRequestForUser,
        AutoScale,
        AwsAttributes,
        AwsAvailability,
        CanManage,
        JobTaskSettings,
        Library,
        NewCluster,
        NotebookTask,
        SparkJarTask,
        TaskDependency,
    )

    submit_multitask_run = DatabricksSubmitMultitaskRun(
        tasks=[
            JobTaskSettings(
                task_key="Sessionize",
                description="Extracts session data from events",
                existing_cluster_id="0923-164208-meows279",
                spark_jar_task=SparkJarTask(
                    main_class_name="com.databricks.Sessionize",
                    parameters=["--data", "dbfs:/path/to/data.json"],
                ),
                libraries=[Library(jar="dbfs:/mnt/databricks/Sessionize.jar")],
                timeout_seconds=86400,
            ),
            JobTaskSettings(
                task_key="Orders_Ingest",
                description="Ingests order data",
                existing_cluster_id="0923-164208-meows279",
                spark_jar_task=SparkJarTask(
                    main_class_name="com.databricks.OrdersIngest",
                    parameters=["--data", "dbfs:/path/to/order-data.json"],
                ),
                libraries=[Library(jar="dbfs:/mnt/databricks/OrderIngest.jar")],
                timeout_seconds=86400,
            ),
            JobTaskSettings(
                task_key="Match",
                description="Matches orders with user sessions",
                depends_on=[
                    TaskDependency(task_key="Orders_Ingest"),
                    TaskDependency(task_key="Sessionize"),
                ],
                new_cluster=NewCluster(
                    spark_version="10.4.x-scala2.12",
                    node_type_id="m4.large",
                    spark_conf={"spark.speculation": True},
                    aws_attributes=AwsAttributes(
                        availability=AwsAvailability.SPOT,
                        zone_id="us-west-2a",
                        ebs_volume_type="GENERAL_PURPOSE_SSD",
                        ebs_volume_count=3,
                        ebs_volume_size=100,
                    ),
                    autoscale=AutoScale(min_workers=1, max_workers=2),
                ),
                notebook_task=NotebookTask(
                    notebook_path="/Users/user.name@databricks.com/Match",
                    base_parameters={"name": "John Doe", "age": "35"},
                ),
                timeout_seconds=86400,
            ),
        ],
        run_name="A multitask job run",
        timeout_seconds=86400,
        access_control_list=[
            AccessControlRequestForUser(
                user_name="jsmith@example.com", permission_level=CanManage.CAN_MANAGE
            )
        ],
    )


    with Flow("my flow") as f:
        conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
        submit_multitask_run(databricks_conn_secret=conn)
methods:                                                                                                                                                       

prefect.tasks.databricks.databricks_submitjob.DatabricksSubmitMultitaskRun.convert_dict_to_kwargs

(input)[source]

Method to convert a dict that matches the structure of the Databricks API call into the required object types for the task input

Args:

  • input (Dict): A dictionary representing the input to the task
Returns:
  • A dictionary with values that match the input types of the class


Example: Use a JSON-like dict as input


from prefect import Flow
from prefect.tasks.databricks import DatabricksSubmitMultitaskRun

submit_multitask_run = DatabricksSubmitMultitaskRun()

databricks_kwargs = DatabricksSubmitMultitaskRun.convert_dict_to_kwargs({
    "tasks": [
        {
            "task_key": "Sessionize",
            "description": "Extracts session data from events",
            "depends_on": [],
            "existing_cluster_id": "0923-164208-meows279",
            "spark_jar_task": {
                "main_class_name": "com.databricks.Sessionize",
                "parameters": ["--data", "dbfs:/path/to/data.json"],
            },
            "libraries": [{"jar": "dbfs:/mnt/databricks/Sessionize.jar"}],
            "timeout_seconds": 86400,
        },
        {
            "task_key": "Orders_Ingest",
            "description": "Ingests order data",
            "depends_on": [],
            "existing_cluster_id": "0923-164208-meows279",
            "spark_jar_task": {
                "main_class_name": "com.databricks.OrdersIngest",
                "parameters": ["--data", "dbfs:/path/to/order-data.json"],
            },
            "libraries": [{"jar": "dbfs:/mnt/databricks/OrderIngest.jar"}],
            "timeout_seconds": 86400,
        },
        {
            "task_key": "Match",
            "description": "Matches orders with user sessions",
            "depends_on": [
                {"task_key": "Orders_Ingest"},
                {"task_key": "Sessionize"},
            ],
            "new_cluster": {
                "spark_version": "7.3.x-scala2.12",
                "node_type_id": "i3.xlarge",
                "spark_conf": {"spark.speculation": True},
                "aws_attributes": {
                    "availability": "SPOT",
                    "zone_id": "us-west-2a",
                },
                "autoscale": {"min_workers": 2, "max_workers": 16},
            },
            "notebook_task": {
                "notebook_path": "/Users/user.name@databricks.com/Match",
                "base_parameters": {"name": "John Doe", "age": "35"},
            },
            "timeout_seconds": 86400,
        },
    ],
    "run_name": "A multitask job run",
    "timeout_seconds": 86400,
    "access_control_list": [
        {
            "user_name": "jsmith@example.com",
            "permission_level": "CAN_MANAGE",
        }
    ],
})

with Flow("my flow") as f:
    conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
    submit_multitask_run(**databricks_kwargs, databricks_conn_secret=conn)

prefect.tasks.databricks.databricks_submitjob.DatabricksSubmitMultitaskRun.run

(databricks_conn_secret=None, tasks=None, run_name=None, timeout_seconds=None, idempotency_token=None, access_control_list=None, polling_period_seconds=None, databricks_retry_limit=None, databricks_retry_delay=None, git_source=None)[source]

Task run method. Any values passed here will overwrite the values used when initializing the task.

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}'
  • tasks (List[JobTaskSettings]):" A list containing the Databricks task configuration. Should contain configuration for at least one task.
  • timeout_seconds (int, optional): An optional timeout applied to each run of this job. The default behavior is to have no timeout.
  • run_name (str, optional): An optional name for the run. The default value is "Job run created by Prefect flow run {flow_run_name}".
  • idempotency_token (str, optional): An optional token that can be used to guarantee the idempotency of job run requests. Defaults to the Python object id.
  • access_control_list (List[AccessControlRequest]): List of permissions to set on the job.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
  • git_source (GitSource): A git source for the source code of the jobs (see https://databricks.com/blog/2022/06/21/build-reliable-production-data-and-ml-pipelines-with-git-support-for-databricks-workflows.html)
Returns:
  • run_id (str): Run id of the submitted run


# DatabricksGetJobID

class

prefect.tasks.databricks.databricks_get_job_id.DatabricksGetJobID

(databricks_conn_secret=None, search_limit=25, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1, **kwargs)[source]

Finds a job_id corresponding to a job name on Databricks using the api/2.1/jobs/list API endpoint.

Args:

  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • search_limit (int, optional): Controls the number of jobs to return per API call, This value must be greater than 0 and less than or equal to 25.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
  • **kwargs (dict, optional): Additional keyword arguments to pass to the Task constructor.
Returns:
  • job_id (int): Job id of the job name.
Examples: You can use the task to feed in the job_id for DatabricksRunNow.

    conn = PrefectSecret('DATABRICKS_CONNECTION_STRING')
    get_job_id = DatabricksGetJobID(databricks_conn_secret=conn)
    dbx_job_id = get_job_id(job_name="dbx")

    notebook_run = DatabricksRunNow(
        job_id=dbx_job_id,
        notebook_params=notebook_params,
        python_params=python_params,
        spark_submit_params=spark_submit_params,
        jar_params=jar_params
    )
    notebook_run(databricks_conn_secret=conn)
methods:                                                                                                                                                       

prefect.tasks.databricks.databricks_get_job_id.DatabricksGetJobID.run

(job_name, databricks_conn_secret=None, search_limit=25, polling_period_seconds=30, databricks_retry_limit=3, databricks_retry_delay=1)[source]

Task run method.

Args:

  • job_name (str): The job_name of an existing Databricks job.
  • databricks_conn_secret (dict, optional): Dictionary representation of the Databricks Connection String. Structure must be a string of valid JSON. To use token based authentication, provide the key token in the string for the connection and create the key host. PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "login": "ghijklmn", "password": "opqrst"}' OR PREFECT__CONTEXT__SECRETS__DATABRICKS_CONNECTION_STRING= '{"host": "abcdef.xyz", "token": "ghijklmn"}' See documentation of the DatabricksSubmitRun Task to see how to pass in the connection string using PrefectSecret.
  • search_limit (int, optional): Controls the number of jobs to return per API call, This value must be greater than 0 and less or equal to 25. this run. By default the task will poll every 30 seconds.
  • polling_period_seconds (int, optional): Controls the rate which we poll for the result of this run. By default the task will poll every 30 seconds.
  • databricks_retry_limit (int, optional): Amount of times retry if the Databricks backend is unreachable. Its value must be greater than or equal to 1.
  • databricks_retry_delay (float, optional): Number of seconds to wait between retries (it might be a floating point number).
Returns:
  • job_id (int): Job id of the job name.


This documentation was auto-generated from commit bd9182e
on July 31, 2024 at 18:02 UTC

← DBT Tasks →