Skip to content
DataSquares
Start free

DataSquares Data Quality

Check your data before you publish

Monitor data quality with scheduled checks and clear issue reports. Critical failures can hold publication while readers continue using the last published version.

Start freeBook a demo
Read the Data Quality documentation
  • 14 rule types, from not-null and uniqueness to custom SQL
  • Failing critical checks hold the publication of dashboards, reports and models
  • Overrides need a reason, and they are recorded in the audit log

Quality checks, scores and issues are on every plan. Publication gates, which hold publishing while a critical check fails, are available on the Business plan. See plans

A failing check holds the release, not the readers

When a critical check on the underlying model or source is failing, publication is held and viewers keep the last good version. An override needs a reason, and the audit log keeps it with the failing tests.

What you can do with Data Quality

  • Rules for columns and tables

    Not null, unique, duplicates, foreign keys, value and date ranges, no future dates, row counts, email and phone formats, allowed values, regular expressions and custom SQL.

  • Run when it matters

    On demand, on a cron schedule, or automatically whenever a source syncs. Each result reads pass, warn, fail or error.

  • Severity that means something

    Critical checks gate publication. High, medium, low and info surface warnings without blocking anything.

  • Column profiles

    Distributions, null ratios and distinct counts for the columns you scan, so you can see what good data looks like before you write a rule.

  • Freshness

    Freshness monitoring flags tables that are stale or late against the schedule you expect.

  • AI scan

    An AI scan looks across your models for drift, outliers and missing relationships, and suggests fixes you apply or dismiss with one click.
14
rule types
Gated
Critical failures hold publication
Audited
Every override records who, why and which tests

Inside Data Quality

Gates that fail safe

Only critical checks gate publication. A check that errors counts as a failure, so unverified data can’t slip through, and a new check that hasn’t run yet doesn’t freeze anything retroactively.

  • The previously published version keeps serving viewers
  • Overrides are audited with the author, reason and failing tests
  • Deleting a rule closes its open issues with a reason

A score you can explain

The overview rolls completeness, validity, uniqueness and freshness into a quality score. A table nobody has scanned reads Not scanned, never a perfect score.

  • Issues gathers failures, anomalies and schema discrepancies in one list
  • Scan history shows runs, durations and score trends
  • Relationship health checks integrity across model dependencies

How it works

  1. Step 1: Profile a table

    See distributions, nulls and distinct counts

  2. Step 2: Create rules

    Pick a rule type, a column and a severity

  3. Step 3: Run them

    On demand, on a schedule or on every sync

  4. Step 4: Gate publication

    Critical failures hold the publish

Questions, answered

Something else? Ask our team or search the documentation.

What happens when a critical check fails?
Publishing a dashboard, report or model built on that data is held, and the previously published version keeps serving viewers. An administrator can override with a mandatory reason, which is recorded in the audit log with the failing tests.
Which sources can be checked?
Column profiling and source tests run on MySQL, MariaDB, PostgreSQL, Redshift, CockroachDB, SQL Server and Azure SQL, and against staged data for uploaded and pulled sources. Other dialects return a clear support error before any query runs.
Can pipelines check data before it lands?
Yes. In Pipelines, an assert step holds named checks. Failing rows either fail the step before anything is published, or go to a quarantine port for inspection.
Who can create and run rules?
Creating and running rules needs the data quality permission, which company admins have and custom roles can grant.

Works well with Data Quality

All products
GovernDataSquares CatalogA shared catalog for trusted dataIntegrateDataSquares PipelinesMove and shape data between your systemsAnalyzeDataSquares BIInteractive dashboards. Shared definitions.

See Data Quality on your own data.

Start a free trial in minutes, or book a walkthrough with our team.

Start freeBook a demo