Skip to main content
With the workspace deployed, connect your data so you can query it. This has two parts: granting the engine access to your data lake, and registering a catalog in the Compute Plane.

​
Grant data-lake access

The In-VPC engine reads your data through IAM. To grant access to data lakes in your account (or in another account), create the cross-account IAM role, S3 read policy, and - if you use Glue - the Glue catalog access policy. These are the same policies used for a Serverless deployment; see Configure AWS access for the full role, trust policy, and S3/Glue policy definitions. The workspace engine role already includes sts:AssumeRole and sts:TagSession, so it can assume a cross-account role you create for data access.

​
Register the catalog

  1. Open the workspace’s Compute Plane and go to the Catalogs section.
  2. Select Create Catalog and choose your Catalog Type - AWS Glue, Hive Metastore, Databricks Unity, Apache Polaris, Iceberg REST, or Amazon S3 Tables.
  3. Provide the connection details and, where your data is in another account, the Target AWS Account ID and IAM Role ARN.
  4. Select Test Connection, choose the schemas to include, and select Create Catalog.
For the full set of catalog types and their connection settings, see Catalogs.

​
Next

Continue to Create a cluster.
Last modified on September 29, 2026