Skip to main content
Version: 0.14.13

How to create a new Checkpoint

This guide will help you create a new CheckpointThe primary means for validating data in a production deployment of Great Expectations., which allows you to couple an Expectation SuiteA collection of verifiable assertions about data. with a data set to ValidateThe act of applying an Expectation Suite to a Batch..

Note: As of Great Expectations version 0.13.7, we have updated and improved the Checkpoints feature. You can continue to use your existing legacy Checkpoint workflows if you’re working with concepts from the Batch Kwargs (v2) API. If you’re using concepts from the BatchRequest (v3) API, please refer to the new Checkpoints guides.

Steps for Checkpoints (>=0.13.12)​

Prerequisites:

This how-to guide assumes you have already:

1. Use the CLI to open a Jupyter Notebook for creating a new Checkpoint​

To assist you with creating Checkpoints, our CLICommand Line Interface has a convenience method that will open a Jupyter Notebook with all the scaffolding you need to easily configure and save your Checkpoint. Simply run the following CLI command from your Data ContextThe primary entry point for a Great Expectations deployment, with configurations and methods for all supporting components.: 

great_expectations checkpoint new my_checkpoint

2. Configure and save your Checkpoint in the provided Jupyter Notebook​

The Jupyter Notebook which was opened in the previous step will guide you through the steps of creating a Checkpoint. It will also include a default configuration that you can edit to suite your use case. The following sections of this document walk you through an example of how to do this.

3. Configuring a SimpleCheckpoint (Example)​

3a. Edit the configuration​

For this example, we’ll demonstrate using a basic Checkpoint configuration with the SimpleCheckpoint class, which takes care of some defaults. Replace all names such as my_datasource with the respective DatasourceProvides a standard API for accessing and interacting with data from a wide variety of source systems., Data ConnectorProvides the configuration details based on the source data system which are needed by a Datasource to define Data Assets., Data AssetA collection of records within a Datasource which is usually named based on the underlying data system and sliced to correspond to a desired specification., and Expectation Suite names you have configured in your great_expectations.yml. 

config = """
name: my_checkpoint
config_version: 1
class_name: SimpleCheckpoint
validations:
  - batch_request:
      datasource_name: my_datasource
      data_connector_name: my_data_connector
      data_asset_name: MyDataAsset
      data_connector_query:
        index: -1
    expectation_suite_name: my_suite
"""

This is the minimum required to configure a Checkpoint that will run the Expectation Suite my_suite against the Data Asset MyDataAsset. See How to configure a new Checkpoint using test_yaml_config for advanced configuration options.

3b. Test your config using context.test_yaml_config​ 

context.test_yaml_config(yaml_config=config)

When executed, test_yaml_config will instantiate the component and run through a self_check procedure to verify that the component works as expected.

In the case of a Checkpoint, this means

  1. validating the yaml configuration,
  2. verifying that the Checkpoint class with the given configuration, if valid, can be instantiated, and
  3. printing warnings in case certain parts of the configuration, while valid, may be incomplete and need to be better specified for a successful Checkpoint operation.

The output will look something like this:

Attempting to instantiate class from config...
Instantiating as a SimpleCheckpoint, since class_name is SimpleCheckpoint
Successfully instantiated SimpleCheckpoint


Checkpoint class name: SimpleCheckpoint

If something about your configuration wasn’t set up correctly, test_yaml_config will raise an error.

2c. Store your Checkpoint config​

After you are satisfied with your configuration, save it by running the appropriate cells in the Jupyter Notebook.

2d. (Optional) Check your stored Checkpoint config​

If the StoreA connector to store and retrieve information about metadata in Great Expectations. backend of your Checkpoint StoreA connector to store and retrieve information about means for validating data in a production deployment of Great Expectations. is on the local filesystem, you can navigate to the checkpoints store directory that is configured in great_expectations.yml and find the configuration files corresponding to the Checkpoints you created.

2e. (Optional) Test run the new Checkpoint and open Data Docs​

Now that you have stored your Checkpoint configuration to the Store backend configured for the Checkpoint Configuration store of your Data Context, you can also test context.run_checkpoint, right within your Jupyter Notebook by running the appropriate cells.

Before running a Checkpoint, make sure that all classes and Expectation Suites referred to in the configuration exist.

When run_checkpoint returns, the checkpoint_run_result can then be checked for the value of the success field (all validations passed) and other information associated with running the specified ActionsA Python class with a run method that takes a Validation Result and does something with it.

For more advanced configurations of Checkpoints, please see How to configure a new Checkpoint using test_yaml_config.

Additional Resources​