Changed behavior

The July 2026 release of Data Governance and Catalog includes the following changed behavior.

Prior to the July 2026 release, unstructured files such as PDF, Text, DOC, and DOCX files were catalogued as structured file assets.

If you have existing PDF, Text, DOC, or DOCX files in your catalog and your organization administrator re-runs the catalog source job after upgrading to the July 2026 release, the files get catalogued as unstructured files with specific asset types. For more information, see Unstructured data discovery and classification.

As a result, you might see duplicate assets in the catalog. Previously discovered unstructured files are deprecated and their lifecycle status is marked as Obsolete. The new assets don't retain the existing enrichments from the deprecated assets. You can add enrichments to the new assets. To remove duplicate assets from the catalog, your organization administrators can purge deprecated assets using the 'Purge Obsolete Objects option on the Explore page for the catalog source in Metadata Command Center.

You can't create business assets in the Obsolete lifecycle through the bulk import process.

Previously, you could create obsolete assets through bulk import if you selected the Skip Errors or Warnings option.

To cancel a job that another user started, or to download its export file, the organization administrator must assign you the Manage Access Control feature privilege in Administrator. You can cancel a job that you started and download its export file without additional privileges.

Previously, you didn't need additional privileges to cancel a job that another user started or to download its export file.

For more information about managing jobs, see Jobs.

The system jobs that propagate updated data quality scores to data elements run once every 12 hours. When a rule occurrence is created from a scorecard and you manually associate a data element with the rule occurrence, the aggregated data quality score on the data element is updated within 12 hours, at the next scheduled run.

Previously, the system jobs ran automatically after the profiling job completed.