Test Results Import

How data-quality test results land in ODD Platform — push-client integrations with Great Expectations and dbt, statistical profiles via odd-collector-profiler, and the custom-framework ingestion endpo

Monitor test suite results in ODD Platform without thinking about masking or removing sensitive data — your datasets do not migrate to your ODD Platform installation, the platform gathers test results only.

The platform ingests test results from four canonical paths today, plus a custom-framework escape hatch.

Great Expectations

Push-client integration via odd-great-expectations. Add the action to your Great Expectations checkpoint configuration; on each validation run the action emits the validation results into ODD's Ingestion API. Test outcomes surface on the dataset's detail page Test reports tab and aggregate into the Quality Dashboard.

dbt tests

Push-client integration via odd-dbt. Run dbt test (or your normal dbt build); the adapter emits the test outcomes into ODD's Ingestion API. The same surface conventions as Great Expectations — per-dataset Test reports tab + dashboard aggregation.

Statistical profiles via odd-collector-profiler

Pull-strategy collector that uses Capital One's DataProfiler under the hood. See odd-collector-profiler for the adapter's per-source coverage. Profile results contribute to the platform's quality signal alongside framework-driven test outcomes — the dashboard renders them as part of the same anomaly-class breakdown.

Custom frameworks

For frameworks not covered by an existing adapter, push results directly through the POST /ingestion/entities/datasets/stats endpoint of the ODD Specification. This is the same Ingestion API contract that every other producer uses; you author a small client that maps your framework's outcomes onto ODD's DataEntityList schema with DataQualityTest and DataQualityTestRun entity types. See the specification's Ingestion API documentation for the wire contract.

Where to next

• Quality Dashboard — where ingested test results aggregate into the cross-catalog quality view.

• Dataset Quality Statuses (SLA) — operator-set Minor / Major / Critical severities on test results, plus the BI-report SLA endpoint.

• Visibility for Data Quality Engineer — the end-to-end DQ-engineer workflow.

• Alerting — where DQ-test-failed alerts surface and how they route to operators.

• odd-great-expectations, odd-dbt, odd-collector-profiler — the per-integration setup pages.