Pipeline testing
Check business correctness before the pipeline goes live.
SQACE verifies data pipelines before go-live against a business rulebook and delivers an audit trail for BCBS 239, MaRisk and NIS2.
30 minutes, on your own use case.
The problem
Business errors survive the standard tests.
ETL and ELT routes in Informatica, Databricks or stored procedures are usually safeguarded with synthetic test data and regression comparisons. Both are useful, yet both miss an entire class of errors.
Synthetic test data rarely covers rare combinations from the production population. Regression tests confirm that nothing has changed, even when the starting state was already wrong. Errors in valuation logic, mapping or complex aggregations only become visible after the release, often in reporting or in reconciliation with the business.
The consequence: downstream corrections, explanations to auditors and a loss of trust with the business.
The SQACE solution
Expected business results as a second, independent calculation.
SQACE produces independent expected results directly from the business rulebook. Business experts describe the logic in a rule-based language that stays readable without programming skills. SQACE computes the rules on the input data and compares the result with the output of the pipeline under test.
Deviations are classified automatically and traced down to the individual-record level. When a rulebook is adjusted, verification runs again, with no new set of test data needed.
Scenarios
Where teams use release verification.
Approve pipeline releases before go-live
Every new version is verified against the current rulebook before the release pull request is approved. Deviations appear with drill-down to the causing record.
Automate baseline comparison after code changes
After changes to transformations, SQACE runs across the unchanged and the changed paths. Hits in the difference point to unintended side effects.
Accelerate hotfix verification in emergencies
When corrections happen under time pressure, SQACE replaces a multi-day manual test with an automated run and a clear audit trail.
In concrete terms
- Up to 80% less effort in the business test layer, depending on scope and rulebook reuse.
- Higher test quality through a second, independent calculation.
- Strong acceptance from the business, because the rules are in business language.
At a glance
| Aspect | SQACE verification | Standard regression |
|---|---|---|
| Reference | Comparison with a business rulebook | Comparison with the previous run |
| Data basis | Representative production or full population | Synthetic test data |
| Finds | Logic errors in mapping and valuation | Behavioral drift |
| Audit fitness | Directly defensible | Limited |
Compliance
Compliance evidence as a by-product.
Every SQACE run produces an audit trail with the input data used, the rules applied, the expected result, the actual result and the classified deviation. This documentation is directly usable, and it is created in the same run rather than as a parallel compliance track.
How SQACE isolates data and stores evidence in an audit-proof way is described in Security & architecture.
FAQ
Frequently asked questions
Which pipeline technologies are supported?
SQACE works at the data level and is independent of the pipeline tool: it checks the result, not the technology behind it. That lets you safeguard routes built with Informatica, Databricks, Snowflake, stored procedures, dbt, Talend or Apache Spark, without SQACE requiring any particular tool.
How quickly is a first verification productive?
A first productive pipeline verification typically takes two to four weeks, depending on the scope of the rulebook and on data access.
Do business experts need to know SQL?
No. The rule language is readable for business experts. SQL knowledge helps but is not required.
How large can the data set be?
SQACE works on full populations of several hundred million records; the platform scales horizontally.
Can SQACE be embedded in CI/CD pipelines?
Yes. Via the REST API, SQACE integrates with Jenkins, GitLab CI, GitHub Actions and Azure DevOps. Releases are only approved once verification is green.
What sets SQACE apart from tools like Great Expectations or Soda?
Those tools check technical data quality such as value ranges and null rates. SQACE checks business correctness against an independently formulated rulebook. The two approaches complement each other.