Manage Data
Data Quality
Data Quality lets you attach a set of rules — an "expectation suite" — to any dataset, so bad data gets caught instead of silently accepted. Rules run whenever you ask, and automatically after every refresh, with a clear pass/fail breakdown and an email alert if something's wrong.
Quick Start
- 1
Create a suite
Open Data Quality, click "New suite", name it, and pick the dataset it should check.
- 2
Add expectations
Add rules like "this column can't be empty", "values must be one of these", or "this column must be unique". Mark each as an Error or a Warning.
- 3
Turn on auto-validate
Leave "Auto-validate after every refresh" checked so the suite reruns automatically whenever the dataset updates — no manual step needed.
- 4
Run it
Click Validate to check the data right now. You'll see how many rules passed, failed, or triggered a warning.
- 5
Drill into failures
Open a failed check to see exactly which rows broke the rule.
- 6
Get notified automatically
If a check fails, the dataset owner gets an email — no one has to go looking for problems.
- A dataset needs to already exist in the Catalog before you can attach a quality suite to it.
- Auto-validation also fires after a fresh import lands in a dataset, not just after a transform.
- Deleting a suite doesn't touch the underlying data — only the rules are removed.