Why Running tests with dbt test? - Purpose & Use Cases

The Big Idea

What if you could catch data mistakes before they ruin your reports, without endless manual checks?

The Scenario

Imagine you have a big spreadsheet with thousands of rows of sales data. You want to check if any rows have missing values or wrong totals. Doing this by scanning each row manually or writing separate scripts for each check can take hours and still miss errors.

The Problem

Manually checking data is slow and tiring. It's easy to overlook mistakes or forget to check some parts. Writing many custom scripts for each test is confusing and hard to maintain. When data changes, you must redo all checks again, wasting time and risking errors.

The Solution

Using dbt test lets you write simple, reusable tests that automatically check your data quality every time you run your project. It runs all tests consistently, finds errors quickly, and helps you fix problems before they cause bigger issues.

Before vs After

✗ Before

SELECT * FROM sales WHERE total IS NULL OR total < 0;

✓ After

dbt test --select model:sales

What It Enables

It enables you to trust your data by catching problems early and saving hours of manual checking.

Real Life Example

A data analyst uses dbt test to automatically verify that daily sales data has no missing values or duplicates before creating reports for the management team.

Key Takeaways

Manual data checks are slow and error-prone.

dbt test automates and standardizes data quality checks.

This saves time and improves trust in your data.

Practice

1. What is the main purpose of running dbt test in a dbt project?

easy

A. To deploy the dbt project to production

B. To build new data models from raw data

C. To check data quality and find errors in your data models

D. To generate documentation for your data models

Solution

1. Step 1: Understand the role of dbt test

   dbt test runs tests defined in your project to check data quality and catch errors.

2. Step 2: Differentiate from other dbt commands

   Building models is done with dbt run, deployment is outside dbt test, and documentation is generated with dbt docs.

3. Final Answer:

   To check data quality and find errors in your data models -> Option C

4. Quick Check:

   dbt test = data quality checks [OK]

Hint: Remember: test = check data quality, run = build models [OK]

Common Mistakes:

• Confusing dbt test with dbt run
• Thinking dbt test deploys models
• Assuming dbt test generates docs

2. Which of the following is the correct command to run tests only on a specific model named customers?

easy

A. dbt test --select customers

B. dbt run --models customers

C. dbt test --models customers

D. dbt test --only customers

Solution

1. Step 1: Identify the correct flag for running tests on specific models

   The flag --select is used with dbt test to specify which models to test.

2. Step 2: Check other options for correctness

   dbt run builds models, not tests. --models is not a valid flag for dbt test. --only is not a valid flag.

3. Final Answer:

   dbt test --select customers -> Option A

4. Quick Check:

   Use --select to target tests [OK]

Hint: Use --select flag to run tests on specific models [OK]

Common Mistakes:

• Using dbt run instead of dbt test
• Using invalid flags like --models or --only
• Confusing --select with other flags

3. Given this schema.yml test definition:

models:
  - name: orders
    tests:
      - unique:
          column_name: order_id
      - not_null:
          column_name: order_date

What will dbt test check for the orders model?

medium

A. It checks that order_id is unique and order_date has no null values

B. It checks that order_id has no null values and order_date is unique

C. It checks that both order_id and order_date are unique

D. It checks that both order_id and order_date have no null values

Solution

1. Step 1: Read the test types for each column

   The test unique applies to order_id, ensuring no duplicates. The test not_null applies to order_date, ensuring no missing values.

2. Step 2: Match tests to their meaning

   unique means no duplicates; not_null means no nulls. So the checks are: order_id is unique and order_date has no null values.

3. Final Answer:

   It checks that order_id is unique and order_date has no null values -> Option A

4. Quick Check:

   unique = no duplicates, not_null = no nulls [OK]

Hint: unique = no duplicates, not_null = no nulls [OK]

Common Mistakes:

• Mixing up unique and not_null tests
• Assuming both columns have the same test
• Ignoring the column_name key in test definitions

4. You run dbt test but get an error: Compilation Error: Could not find test 'uniquee'. What is the likely cause?

medium

A. The test passed with no errors

B. The model name is incorrect

C. The database connection is missing

D. A typo in the test name in schema.yml

Solution

1. Step 1: Analyze the error message

   The error says it cannot find test 'uniquee', which looks like a misspelled test name.

2. Step 2: Identify common causes of compilation errors

   Typos in test names in schema.yml cause dbt to fail to find the test. Model name or connection errors produce different messages.

3. Final Answer:

   A typo in the test name in schema.yml -> Option D

4. Quick Check:

   Compilation errors often mean typos [OK]

Hint: Check spelling of test names in schema.yml [OK]

Common Mistakes:

• Ignoring typo errors and rerunning blindly
• Assuming connection issues cause compilation errors
• Confusing model name errors with test name errors

5. You want to ensure that the email column in your users model is unique and not null. You also want to run tests only on this model. Which schema.yml snippet and command combination is correct?

hard

A. models:
- name: users
tests:
- unique
- not_nullCommand: dbt test --models users

B. models:
- name: users
columns:
- name: email
tests:
- unique
- not_nullCommand: dbt test --select users

C. models:
- name: users
columns:
- name: email
tests:
- unique
- not_nullCommand: dbt run --models users

D. models:
- name: users
tests:
- unique
- not_nullCommand: dbt test --select users

Solution

1. Step 1: Identify correct test syntax in schema.yml

   Tests on columns are defined under columns with name and tests list.

   models:
     - name: users
       columns:
         - name: email
           tests:
             - unique
             - not_null

   
   Command: dbt test --select users shows the correct format.

2. Step 2: Choose the correct command to run tests on the users model

   dbt test --select users runs tests only on the users model.

   models:
     - name: users
       columns:
         - name: email
           tests:
             - unique
             - not_null

   
   Command: dbt test --select users uses this command correctly.

3. Final Answer:

   Option B with column-level tests and dbt test --select users -> Option B

4. Quick Check:

   Column tests + --select flag = correct [OK]

Hint: Define tests under columns, run with --select flag [OK]

Common Mistakes:

• Defining tests directly under model without column_name (invalid syntax)
• Using dbt run instead of dbt test
• Using incorrect flags like --models