Print, subscribe, snapshot, and monitor with the new Pivot table

The new pivot table is now getting support for download as image and PDF, storytelling, subscriptions, and monitoring. We know many have been waiting for this and we’re glad to announce general availability. The new tables are rolled out with new features incrementally and will soon be moved from the visualization bundle to the native chart section

Visualization bundle

AI processor for transformation flows

A new processor has been introduced in transformation flows to make it easy to leverage your Databricks platform’s AI capabilities. The AI processor includes seven state-of-the-art generative AI functions:

• sentiment analysis

• grammar issues fix

• translation

• text summary

• string similarity

• data masking

• classification.

AI processor

New Direct Access gateway version 1.6.8

Direct Access gateway 1.6.8 introduces new capabilities and resolves several issues. To benefit from the enhancements and resolved issues, customers are encouraged to upgrade at the earliest opportunity. For upgrade instructions and a list of resolved issues, see Upgrading the Direct Access gateway installation

• Data chunks recovery mechanism

  This version introduces support for sending missing or corrupted data chunks. To control the size in memory of the reloaded data chunks, a new property has been added to the configuration.properties file: RELOAD_CACHE_MEMORY_MB

  To learn how to set this property, see Caching data for reload in the event of a problem

• .NET 6.0.x no longer required

  From this version, running Direct Access gateway no longer requires .NET 6.0.x. Prior to installing Direct Access gateway, new customers should install .NET 4.8 (needed for the installation) as well as .NET 8.0 and ASP.NET Core Runtime (for working with Direct Access gateway) only.

  For more information about Direct Access gateway prerequisites, see System prerequisites

• Improved batch operations for dependent processes

  From this version, dependent processes will be checked in parallel instead of sequentially, thereby avoiding exceptions due to modification of the process list during runtime.

  For more information on process isolation, see Mitigating connector crashes during reload

• Improved troubleshooting

  More context (for example such as “process ID” and “commit size”) was added to the resource usage warnings in the logs.