Databricks SDKs added Delta Sharing replication configuration while Procore moved to hourly data sharing

The one product change specific to Delta Sharing in October 2026 came in the Databricks SDKs. Releases v0.167.0 and v0.153.0, both published on October 15, added support for replication configuration for Delta Sharing [5][6]. Both releases also added a realtime warehouse type and issue tracking methods for ML experiments. Both removed the effective serverless compute ID field from job run tasks, which was a breaking change for code that read it [5][6]. Separately, Delta Lake 4.4.1 on October 5 added safeguards for catalog-managed tables and fixed snapshot construction errors caused by unratified staged commits [10].

Most of the month's announcements were about how organizations used sharing in production. Procore moved its construction analytics product from an API-driven, once-a-day refresh to OpenSharing on Databricks, so customers could pull data into downstream systems hourly [1]. CARIAD published a customer story on unifying global automotive data with OpenSharing and Unity Catalog [7]. A video on the Databricks and SAP partnership described sharing SAP data products into Databricks using Delta Sharing, with a supply chain example of agents automating predictive maintenance and procurement [9]. A Genie tutorial also covered receiving shared data locally, using a dataset obtained from the Databricks Marketplace [2].

The community threads were mostly about access design. One asked how to handle permissions and access resources that a Databricks Asset Bundles-based CI/CD pipeline cannot declare [3]. Another asked about a cross-subscription and cross-workspace access pattern for PII data in Azure Databricks [4]. A third post shared notes on whether Delta Sharing replaces external ETL, comparing open sharing with Databricks-to-Databricks sharing [8].