Aug 20, 2026: Snowflake Data Clean Rooms updates¶
Clean Rooms API Version: 17.6¶
The following updates are now available in Snowflake Data Clean Rooms:
- External and Apache Iceberg™ tables in cross-region collaborations: Collaborators can now share data offerings that reference external or Iceberg tables with parties in the same region as the collaboration owner, even when the collaboration itself is leveraging Cross-Cloud Auto-Fulfillment. These offerings can’t be shared with collaborators in other clouds or regions, because replication isn’t supported for these object types in Data Clean Rooms.
- Preset tables in clean room templates (public preview): A template author can now preset a specific provider dataset in a template with a
preset_tablesblock. The template logic can use this dataset and the analysis runner doesn’t need to supply it at run time. Learn more. - Legacy identity connectors removed from the web application: As part of the ongoing deprecation of the legacy Data Clean Rooms web application, the LiveRamp, Acxiom, Merkury, and PAIR third-party identity connectors have been removed. TransUnion remains available. Available identity connectors are listed in the connectors documentation.
- Legacy activation connectors removed from the web application: As part of the same deprecation, the following third-party activation connectors have been removed from the provider and consumer web interface: Google DV360 (PAIR and Customer Match), LiveRamp (SFTP and Share), The Trade Desk (UID 2.0 and CRM), and Yahoo DSP. Google Ads and Meta Ads Manager remain available and continue to be supported through the Collaboration API. Available activation connectors are listed in the connectors documentation.
- Consecutive update requests no longer block each other: Consecutive Add Template, Remove Template, Link Data Offering, and Unlink Data Offering requests can now be approved in a row in any order. Previously, approving one of them would cause other pending requests to fail a version check. This doesn’t apply to Edit Collaboration requests, which are still subject to the version check described in the next update.
- Conflicting Edit Collaboration requests fail immediately: An Edit Collaboration request is built against a specific version of the collaboration spec and can’t be applied after another change updates that version. Previously the request stayed pending until someone approved it and only then failed. It now moves to the
FAILEDstatus as soon as the other change is applied, and thedetailscolumn ofVIEW_UPDATE_REQUESTSreports the reason. Resubmit the edit request against the updated collaboration if you face this error. - Automatic version rollover when approaching the patch limit: When a collaboration approaches the per-version patch limit, Snowflake automatically increments the version and resets the patch counter. This removes the effective limit on the number of patches for a given collaboration.
- General performance improvements and bug fixes.
- Updates to private preview features.