Run Your First Test in 10 Minutes

Prev Next

To run your first test, open a table in the Catalog, click Add Test on its Tests tab, pick a quick test, fill in its values and run it. Once your data source has quick tests, that takes about 10 minutes and you never leave the Catalog.

When you finish, you'll have a row count test on one table, saved in a project, with a passing result you can open.

Before you start

  • A data source with its schema metadata ingested. The Catalog only lists tables Validatar has ingested. If you don't have a data source yet, start with Create a Data Source.
  • Permission to create tests. You need Create and Edit Tests on the data source and Create Tests in a project that uses it.
  • Quick tests for that data source. Read the next section first. It decides whether this page works for you today.

Where quick tests come from

A quick test is a template test that someone has offered to the Catalog: a Metadata Driven template test that uses Select Objects and has Enable Adding from Catalog checked, in a project that includes your data source. When you pick one in the Test Wizard, Validatar adds your table to that template and creates a test for it.

So the list you see depends on what your team has set up. Validatar ships one ready-made set, and it is for Snowflake.

Your data source How to get quick tests
Snowflake Create the built-in Validatar Data Quality Monitoring quick start project for the data source, once. It adds 97 quick tests (24 for tables, 73 for columns) covering row counts, freshness, nulls, uniqueness, value formats and schema changes, plus four daily jobs to schedule them in. Steps below.
Any other platform Someone who builds tests turns on Enable Adding from Catalog on a template test. See Add Quick Tests from the Catalog. Until then the wizard says "No available templates found". To test this data source today, follow Check That Row Counts Match Between Source and Target instead.

To create the Snowflake quick start project (you need the Security Admin role):

  1. Go to Settings > Data Sources and open your Snowflake data source.
  2. In the left panel, click Projects, then click New Quick Start Project.
  3. Click the Validatar Data Quality Monitoring tile.
  4. Keep the suggested Project Name (your data source's name followed by "- Validatar DQM") or type your own.
  5. Check Copy Current Data Source Permissions to New Project? to copy the data source's permissions to the new project.
  6. Click Create Project.

The example

This page adds a test to SALES_ORDERS in the Enterprise Data Warehouse, a Snowflake warehouse. SALES_ORDERS is the warehouse copy of the Sales.Orders table in the Wholesale Distribution System, loaded into the LOAD_SQLSERVER_WORLDWIDEIMPORTERS schema.

The quick test is table row count - in range. A load that half-fails leaves the orders table with far fewer rows than usual, and a load that runs twice leaves it with far more. A floor and a ceiling on the row count catch both before anyone builds a report on the table.

Step 1: Open the table

  1. Click Catalog in the top bar.
  2. Open Enterprise Data Warehouse, go to its Tables tab and open SALES_ORDERS. You can also type the table name in the Search box at the top of the page.
  3. On the Table Info tab, open the Table Statistics panel and note Number of Rows. You'll base your range on it in step 3. The number comes from the table's latest profile, so if it says N/A, use a count you already know.

Step 2: Pick a quick test

  1. Open the Tests tab and click Add Test. The Test Wizard opens and asks "What can we check for you?"
  2. Type row count in the search box at the top left. The Quality Dimensions and Projects filters below it narrow the list further.
  3. Click table row count - in range. The card opens to show its description, severity level and project. Your test will be created in that project.
  4. Click Next.

Step 3: Set the range

The wizard's Configure step ("Let's get the details:") lists the quick test's Test Variables.

  1. In min_row_count, enter the fewest rows the table should ever hold. In max_row_count, enter the most. The example uses 400000 and 500000.
  2. If the step also shows Table Metadata fields, leave them as they are.
  3. Click Next. The wizard says "Congratulations, the following 1 test will be created:" and lists the test's name.
  4. Click Create Test.

Validatar adds SALES_ORDERS to the quick test's template and creates one test for it, named table row count - between 400000 and 500000 - SAMPLE_DW.LOAD_SQLSERVER_WORLDWIDEIMPORTERS.SALES_ORDERS. The test is a child test of that template, so a later fix to the template reaches it too. The wizard moves on to its Run step and says "All set, 1 test created!"

Step 4: Run it and read the result

  1. In the test's row, click Run in the Preview column. After a few seconds the column shows Passed.
  2. Click Passed. The Result tab shows the Test Value, which is the table's row count, and the Control Range you entered. Close the result.
  3. Click Run All To Repository. The preview run in step 1 isn't saved; this run saves the result to the test's history.

On the demo the test passes: SALES_ORDERS holds 431,507 orders, inside the range.

Step 5 (optional): Schedule it

To run the test every day, click Schedule Test on the wizard's Run step before you close it. The Schedule step lists jobs that already have a schedule. The Validatar Data Quality Monitoring project comes with four daily jobs, from 1. Midnight - 12 AM CDT to 4. Afternoon - 5 pm CDT. Pick one and click Add To Schedule. To build your own job and schedule, see Schedule a Job.

Check that it worked

Close the wizard. The Tests tab on SALES_ORDERS now lists your test. The test also lives in the quick test's project: open Projects, then the project (in the example, Enterprise Data Warehouse - Validatar DQM). The Validatar Data Quality Monitoring quick tests put each test in a folder path built from the server, database, schema and table names.

Do it with Validatar AI

Do it with Validatar AI (Validatar Cloud, beta). On the table's Tests tab, click Recommend Table Tests: the Validatar AI panel opens with a request for test recommendations already filled in, and you send it. Then ask: "Add the table row count - in range quick test to SALES_ORDERS with a minimum of 400000 and a maximum of 500000." Review what it creates before you run it. You can undo the change from the chat history.

The AI works from the same quick tests as the Test Wizard, so it can't add one your data source doesn't have.

Troubleshooting

Symptom Cause Fix
The wizard says "No available templates found" No quick test at this level is set up for this data source in a project where you have Create Tests On Snowflake, create the quick start project above. Elsewhere, ask a test author to turn on Enable Adding from Catalog on a template test, or to give you Create Tests in the project that has one
The quick test you expect isn't listed It's a column-level quick test, or a filter is on Open a column (from the table's Columns tab) and use its Tests tab, or click Clear filters in the wizard
A card says "already exists - select to edit" This table already has that quick test Selecting it updates the existing test with the values you enter instead of adding a second one
No New Quick Start Project button, or it's disabled You don't have the Security Admin role Ask a Security Admin to create the project
No Validatar Data Quality Monitoring tile The data source doesn't use the Snowflake data source template That project is offered for Snowflake data sources only
The test failed The row count is outside your range Open the result. If the range was wrong, run the wizard again on the same quick test and change the values. If the count is wrong, you've found a problem in the data
The test shows an error instead of a result The query couldn't run, often a permission on the table See Why Is My Test Failing or Erroring?

Next steps

Engineers who work in an AI coding tool can add tests through the Validatar MCP server or the Validatar Agent Kits.