dbt-expectations for data quality - Step-by-Step Execution

Concept Flow - dbt-expectations for data quality

Write dbt model SQL

↓

Add dbt-expectations tests

↓

Run dbt test command

↓

dbt-expectations runs tests

↓

Tests pass?

No→Report errors

Yes↓

Data quality confirmed

This flow shows how you add dbt-expectations tests to your dbt models, run tests, and check if data quality passes or errors are reported.

Execution Sample

dbt

select * from {{ ref('my_table') }}

-- Add test for no nulls in id
{{ expect_column_values_to_not_be_null('id') }}

-- Add test for unique ids
{{ expect_column_values_to_be_unique('id') }}

This code selects data from a table and adds two dbt-expectations tests: no nulls in 'id' column and unique values in 'id'.

Execution Table

Step	Action	Test Run	Result	Notes
1	Run dbt test	expect_column_values_to_not_be_null('id')	Pass	No null values found in 'id' column
2	Run dbt test	expect_column_values_to_be_unique('id')	Fail	Duplicate values found in 'id' column
3	Report errors	Summary	Fail	Data quality test failed due to duplicates
4	Fix data or model	N/A	N/A	User updates data or model to fix duplicates
5	Re-run dbt test	All tests	Pass	All data quality tests pass

💡 Tests stop when all pass or errors are reported for fixing

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 5
id column null check	Not run	Pass	Pass	Pass
id column uniqueness	Not run	Not run	Fail	Pass
data quality status	Unknown	Unknown	Fail	Pass

Key Moments - 2 Insights

Why does the uniqueness test fail even if the null check passes?

Because null values and duplicates are different issues. The null check (row 1) passes since no nulls exist, but duplicates cause the uniqueness test (row 2) to fail.

What happens after a test fails?

The failure is reported (row 3), and you need to fix the data or model before re-running tests (rows 4 and 5).

Visual Quiz - 3 Questions

Look at the execution table, what is the result of the null check test at step 1?

A Not run

B Fail

C Pass

D Error

💡 Hint

Check the 'Result' column in row 1 of the execution table.

At which step does the uniqueness test fail?

A Step 1

B Step 2

C Step 3

D Step 5

💡 Hint

Look at the 'Test Run' and 'Result' columns in the execution table rows.

If the duplicates were fixed before step 5, what would be the data quality status after step 5?

A Pass

B Fail

C Unknown

D Error

💡 Hint

Refer to the 'data quality status' variable in the variable tracker after step 5.

Concept Snapshot

dbt-expectations adds data quality tests to dbt models.
Write tests like expect_column_values_to_not_be_null or expect_column_values_to_be_unique.
Run tests with 'dbt test' command.
Tests pass or fail based on data.
Fix data if tests fail and re-run.
Ensures reliable, clean data in your warehouse.

Full Transcript

This visual execution shows how dbt-expectations helps check data quality in dbt projects. First, you write your dbt model SQL. Then, you add dbt-expectations tests to check things like no nulls or unique values in columns. When you run 'dbt test', these tests run on your data. If tests pass, data quality is confirmed. If tests fail, errors are reported so you can fix data or models. After fixing, you re-run tests until all pass. The execution table traces each test step, showing pass or fail results. The variable tracker shows how test results change over time. Key moments clarify why some tests fail while others pass and what to do next. The quiz checks your understanding of test results and flow. This process helps keep your data clean and trustworthy.