Skip to content

27 August 2026 Release Notes

Enterprise Version 5.92.1

Feature highlights

Microsoft OneLake

DataOps TestGen Enterprise is now compatible with Microsoft OneLake (Microsoft Fabric)! Set up TestGen so you can better understand the health of your lakehouse data with intelligent profiling and automatically generated tests. TestGen connects through the Fabric SQL analytics endpoint and authenticates with an Entra ID Service Principal or a Managed Identity.

For connection setup details, see Database Access Requirements.

Commercial Pharma test types

TestGen Enterprise now includes 12 Validity test types for the identifiers that commercial pharmaceutical data is joined and reported on — NPI, DEA, NDC, ICD diagnosis and procedure codes, HCPCS J-code, CMS place of service, and NCT. They are grouped in a Commercial Pharma generation set, so they are only generated for the test suites you select it on.

For details, see Commercial Pharma Tests.

Upload your own test types

A test type tailored to your data — a proprietary identifier format, an industry code set, a domain-specific statistical check — can now reach a TestGen Enterprise instance without waiting for a release. Contact your DataKitchen representative to have a test type file authored for your use case; a system administrator then uploads it from System Settings. An uploaded test type behaves like any other: it appears in the test type list, can be added to a test suite, and auto-generates when the file includes selection criteria and a generation set.

For details, see Upload Test Types.

Tune monitors and manage access from AI assistants

The Model Context Protocol (MCP) server gains full lifecycle control over monitors and per-monitor threshold tuning.

Assistants authenticated as system administrators can also manage project membership, listing who has access to a project and setting their roles.

For the full tool reference, see Supported Tools.

REST API enhancements

  • A new endpoint returns the per-column profiles for a profiling run, ordered worst-first by hygiene severity.
  • New endpoints return a profiling run's hygiene issues, filterable by issue type, likelihood, and disposition, and the columns flagged as potential PII, filterable by risk and disposition.
  • New endpoints return the monitors on a table group with their anomaly counts, and the time series behind a single monitor.
  • A new endpoint returns a project's quality score rollup, filterable by catalog attribute, with an optional breakdown.

For the full API reference, see API.

Deprecations

Command-line interface

The testgen command-line interface is deprecated. Its commands for running and reading data quality work are no longer the recommended way to drive TestGen.

  • Use the API to automate TestGen from CI/CD pipelines, orchestration tools, and your own scripts.
  • Use the MCP Server for interactive and AI-assisted work.

Unlike the CLI, both authenticate as a specific user and are scoped to that user's permissions.

Other fixes and enhancements

  • Azure SQL Database and Azure Synapse Analytics connections can now authenticate with an Entra ID Service Principal.
  • The Add Test and Edit Test dialogs now show which fields are required and keep Save disabled until they are filled in, and place each match field beside the field it refers to.
  • A test suite can now draw its tests from more than one generation set at once, and TestGen remembers the selection, so regenerating later draws from the same sets unless you change them.
  • We fixed hygiene issue detection stopping partway through a run when a profiled value contained a line break or a pipe delimiter.
  • We fixed profiling failures due to text values containing a date-like fragment and pattern collection on sampled tables.
  • We fixed SQL Server connections being left open on the server holding an uncommitted transaction.
  • Empty tables are no longer reported as inaccessible when you verify table access.
  • We fixed clearing Reconfigure Freshness monitors after profiling having no effect.
  • We fixed UI display errors when a test run contained a Value Match All result.
  • We applied general fixes to reduce errors, increase security, and improve our user experience.