Blog Post

Power BI Updates Blog
1 MIN READ

Amazon Redshift V2 Connector for Microsoft Data Factory

MichaelaIsaacs's avatar
MichaelaIsaacs
Icon for Microsoft Employee rankMicrosoft Employee
16 hours ago

Microsoft is updating our Amazon Redshift connector to use the new Amazon Redshift ODBC 2.x driver.

What’s new and why upgrade

Microsoft and Amazon Redshift have been collaborating to provide our shared customers with an industry-leading, improved security, connectivity experience. The new driver will ensure compatibility, stability, security, and support for your data connectivity. This improvement aligns with Microsoft’s continued commitment to enhancing and providing secure data connectivity solutions for customers, enabling the mitigation of data connectivity threats.

To support this initiative, existing connectors within Data Factory are transitioning to the Amazon Redshift ODBC 2.x driver for all Redshift connectors; the new connector is rolling out automatically across regions.

This improvement will benefit connectivity between Redshift and various Microsoft products, ranging from Fabric Copyjob, Fabric Copy pipeline, and Dataflows Gen2 to Semantic models, Power BI Desktop, Gateway, and more.

Key capabilities in this new connector include (and are not limited to):

  • Metadata visibility for external objects.
  • Support for Redshift Data Sharing reads.
  • Enhanced support for new features launched in Amazon Redshift.
  • Ability to audit connections between Microsoft products and Amazon Redshift.

Rollout details

The new connector is rolling out automatically across regions. To be best prepared, we recommend that customers who want to implement and test the connector sooner should enable the preview features.

To learn more, please reference the Microsoft Redshift documentation.

Updated 6 days ago
Version 1.0

1 Comment

  • rafal_c's avatar
    rafal_c
    Frequent Visitor

    Are there any plans to support redshift connections on gateways in copy data in pipelines? currently we need to use postgresql connector for that. And copy parallelism is not available in such scenario...