Why testing ensures data quality in dbt - Visual Breakdown

Concept Flow - Why testing ensures data quality

Define data tests

↓

Run tests on data

↓

Check test results

↓

Data is

↓

trusted

↓

Improve data quality

↓

Re-run tests

This flow shows how defining and running tests on data helps catch errors early, leading to trusted data or fixing issues to improve quality.

Execution Sample

dbt

select * from users where email is null;
-- test to find missing emails

select count(*) from orders where order_date > current_date;
-- test to find future order dates

These tests check for missing emails and future order dates, which indicate data quality problems.

Execution Table

Step	Test Query	Test Purpose	Result	Action
1	select * from users where email is null;	Check for missing emails	2 rows found	Fail - investigate missing emails
2	select count(*) from orders where order_date > current_date;	Check for future order dates	0 rows found	Pass - no future orders
3	Re-run tests after fix	Verify fixes	0 rows found	Pass - data quality improved

💡 Tests stop when all checks pass, ensuring data quality is verified.

Variable Tracker

Variable	Start	After Test 1	After Fix	Final
missing_emails_count	unknown	2	0	0
future_orders_count	unknown	0	0	0

Key Moments - 2 Insights

Why do we run tests before trusting data?

Because tests reveal problems like missing or incorrect data early, as shown in execution_table step 1 where missing emails were found.

What happens if a test fails?

We investigate and fix the data issues, then re-run tests to confirm the fix, as shown in execution_table steps 1 and 3.

Visual Quiz - 3 Questions

Look at the execution table, what was the result of the test checking for future order dates?

A 2 rows found

B Test not run

C 0 rows found

D Error in query

💡 Hint

Check execution_table row 2 under the Result column.

At which step did the missing emails count become zero?

A After Test 1

B After Fix

C Start

D Final

💡 Hint

Look at variable_tracker row for missing_emails_count.

If the test for missing emails still found 1 row after fix, what would happen next?

A Investigate and fix again

B Stop testing

C Data is trusted

D Ignore the test

💡 Hint

Refer to concept_flow where failing tests lead to investigation and fixing.

Concept Snapshot

Testing in dbt means writing queries that check data for errors.
Run tests regularly to catch problems early.
If tests fail, fix data and re-run tests.
Passing tests mean data is trusted and high quality.
Testing ensures reliable data for decisions.

Full Transcript

Testing ensures data quality by running queries that check for data problems like missing or incorrect values. When tests find issues, we fix them and run tests again to confirm the fix. This cycle helps keep data trustworthy and reliable for analysis and decisions.