Data Quality

Learn how you can use OpenMetadata to define Data Quality tests and measure your data reliability.

You must have a running deployment of OpenMetadata to use this guide. OpenMetadata includes the following services:

  • OpenMetadata server supporting the metadata APIs and user interface
  • Elasticsearch for metadata search and discovery
  • MySQL as the backing store for all metadata
  • Airflow for metadata ingestion workflows

To deploy OpenMetadata checkout the deployment guide

Please use the following command to check the version of Python you have.

OpenMetadata is where all users share and collaborate around data. It is where you make your assets discoverable; with data quality you make these assets trustable.

This section will show you how to configure and run Data Quality pipelines with the OpenMetadata built-in tests.

Test Suites are containers allowing you to group related Test Cases together. Once configured, a Test Suite can easily be deployed to execute all the Test Cases it contains.

Test Definitions are generic tests definition elements specific to a test such as:

  • test name
  • column name
  • data type

Test Cases specify a Test Definition. It will define what condition a test must meet to be successful (e.g. max=n, etc.). One Test Definition can be linked to multiple Test Cases.

Note: you will need to make sure you have the right permission in OpenMetadata to create a test.

From your table service click on the profiler tab. From there you will be able to create table tests by clicking on the purple background Add Test top button or column tests by clicking on the white background Add Test button.

Write your first test

Write your first test

On the next page you will be able to either select an existing Test Suite or Create a new one. If you select an existing one your Test Case will automatically be added to the Test Suite

Create test suite

Create test suite

On the next page, you will create a Test Case. You will need to select a Test Definition from the drop down menu and specify the parameters of your Test Case.

Note: Test Case name needs to be unique across the whole platform. A warning message will show if your Test Case name is not unique.

Create test case

Create test case

If you have created a new test suite you will see a purple background Add Ingestion button after clicking submit. This will allow you to schedule the execution of your Test Suite. If you have selected an existing Test Suite you are all set.

After clicking Add Ingestion you will be able to select an execution schedule for your Test Suite (note that you can edit this later). Once you have selected the desired scheduling time, click submit and you are all set.

Create ingestion workflow

Create ingestion workflow

When creating a JSON config for a test workflow the source configuration is very simple.

The only section you need to modify here is the serviceName key. Note that this name needs to be unique across OM platform Test Suite name.

Once you have defined your source configuration you'll need to define te processor configuration.

The processor type should be set to "orm-test-runner". For accepted test definition names and parameter value names refer to the tests page.

sink and workflowConfig will have the same settings than the ingestion and profiler workflow.

To run the tests from the CLI execute the following command

As with the Ingestion or Profiler workflow, you can as well execute a Test Suite directly from Python. We are going to use Airflow as an example, but any orchestrator would achieve the same goal.

Let's prepare the DAG as usual, but importing a different Workflow class:

Note how we are using the TestSuiteWorkflow class to load and execute the tests based on the YAML configurations specified above.

From the home page click on the Test Suite menu in the left pannel.

Test suite home page

Test suite home page

This will bring you to the Test Suite page where you can select a specific Test Suite.

Test suite landing page

Test suite landing page

From there you can select a Test Suite and visualize the results associated with this specific Test Suite.

Test suite results page

Test suite results page

Navigate to your table and click on the profiler tab. From there you'll be able to see test results at the table or column level.

In the top pannel, click on the white background Data Quality button. This will bring you to a summary of all your quality tests at the table level

Test suite results table

Test suite results table

On the profiler page, click on a specific column name. This will bring you to a new page where you can click the white background Quality Test button to see all the tests results related to your column.

Test suite results table

Test suite results table

While OpenMetadata provides out of the box tests, you may want to write your test results from your own custom quality test suite. This is very easy to do using the API.

First, you'll need to create a Test Definition for your test. You can use the following endpoint /api/v1/testDefinition using a POST protocol to create your Test Definition. You will need to pass the following data in the body your request at minimum.

Here is a complete CURL request

Make sure to keep the UUID from the response as you will need it to create the Test Case.

You'll also need to create a Test Suite for your Test Case -- note that you can also use an existing one if you want to. You can use the following endpoint /api/v1/testSuite using a POST protocol to create your Test Definition. You will need to pass the following data in the body your request at minimum.

Here is a complete CURL request

Make sure to keep the UUID from the response as you will need it to create the Test Case.

Once you have your Test Definition created you can create a Test Case -- which is a specification of your Test Definition. You can use the following endpoint /api/v1/testCase using a POST protocol to create your Test Case. You will need to pass the following data in the body your request at minimum.

Important: for entityLink make sure to include the starting and ending <>

Here is a complete CURL request

Make sure to keep the UUID from the response as you will need it to create the Test Case.

Once you have your Test Case created you can write your results to it. You can use the following endpoint /api/v1/testCase/{test FQN}/testCaseResult using a PUT protocol to add Test Case Results. You will need to pass the following data in the body your request at minimum.

Here is a complete CURL request

You will now be able to see your test in the Test Suite or the table entity.