# Data Quality Check

> A rule data must pass. The pipeline tests it each time.

**Field:** Data Engineering  
**Category:** Governance & Quality  
**Also known as:** Data Test, Data Quality Test, Data Quality Rule

## Think of it like

It is like the guard at the school gate. Each child must show a pass before going in.

## Definition

A data quality check is a rule that data must pass, run automatically in the pipeline. It might say a column has no blanks, or a key has no copies. It might say a value sits in a range, or a table is not empty. Checks usually run right after a load or a transform. A failed check can stop the next step, or just warn. Most data tools let you write checks in a few lines.

## Example

Jack adds a check that no price is below zero. A week later it stops a load with a bad currency fix, before any report sees it.

## Example code

```yaml
models:
  - name: orders
    columns:
      - name: order_id
        data_tests:
          - unique
          - not_null
```

Two dbt tests on one column. The build stops if any order ID is blank or appears twice.

## Why it matters

Checks catch bad data at the door, when it is cheap to fix. Without them, users find the problem in a meeting.

## Common mistake

Writing hundreds of checks that only warn. People learn to ignore the noise, and the one that matters is missed.

## Three principles

- **A clear rule.** Pass or fail, nothing in between.
- **Runs in the pipeline.** Right after data lands or changes.
- **Stop or warn.** Decide what a failure blocks.

## Related terms

- [Data Quality](https://learn.benibix.com/term/data/data-quality)
- [Great Expectations (GX)](https://learn.benibix.com/term/data/great-expectations)
- [dbt (Data Build Tool)](https://learn.benibix.com/term/data/dbt)
- [Data Contract](https://learn.benibix.com/term/data/data-contract)
- [dbt Test](https://learn.benibix.com/term/data/dbt-test)
- [Data Profiling](https://learn.benibix.com/term/data/data-profiling)
- [Data Reconciliation](https://learn.benibix.com/term/data/data-reconciliation)

## Told apart from

- [Data Quality Check vs Data Observability](https://learn.benibix.com/compare/data-quality-check-vs-data-observability): People think data observability is just more quality checks. It is not. A check tests a rule you wrote, like no blank IDs. It runs each time data passes. Observability learns what a normal day looks like. It flags a day that looks odd, even when no rule covers it. Checks catch what you saw coming, and it catches the rest.
- [Data Profiling vs Data Quality Check](https://learn.benibix.com/compare/data-profiling-vs-data-quality-check): People think profiling data is the same as testing it. A profile only counts what is there: blanks, values and ranges. It passes no judgment. A quality check tests a rule, like no blank emails, and fails when the rule breaks. You profile first to learn the data. Then you write checks for the problems you found.

![Top: a pipeline with a gate between load and transform, where checks for not null, unique and price above zero let good rows through and stop bad ones. Bottom: monitors watch a table over a week for row count, freshness and schema, and raise an alert when Thursday's row count drops.](https://learn.benibix.com/diagrams/quality-checks-vs-observability.svg)

---

Source: https://learn.benibix.com/term/data/data-quality-check
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-18
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
