What it is
A Shortcut is a logical pointer that makes data in one location appear as if it lives inside a Lakehouse. The data isn't copied; queries read directly from the source location, but with OneLake's metadata, security, and discoverability layered on top.
Supported targets
- OneLake — point at another workspace's Lakehouse or Warehouse table
- Azure Data Lake Storage Gen2 — your existing ADLS containers
- Amazon S3 — including S3-compatible (R2, MinIO) endpoints
- Google Cloud Storage
- Dataverse — Power Platform data
- External Iceberg / Delta tables
Use patterns
- Cross-domain reuse. Domain A's Gold tables become visible (read-only) in Domain B's Lakehouse without duplicating data.
- Cross-cloud federation. An S3 bucket of public datasets sits next to your private Lakehouse tables, queryable in the same query.
- Vendor data on-ramp. A vendor drops files in their own ADLS; you shortcut them into your Lakehouse without an ingestion pipeline.
- Power Platform integration. Shortcut Dataverse tables into Fabric for analytics without copying CRM data.
Best practices
- Use Shortcuts before you build a pipeline. Pipelines move data; Shortcuts let it stay where it is.
- Document the source. Shortcut data looks local but isn't. The runbook should make that explicit.
- Watch egress on cross-cloud Shortcuts. Querying terabytes of S3 data from Fabric incurs egress cost — caching helps.
- Use Shortcuts to enforce data product boundaries. Each domain publishes Gold via Shortcut; consumers don't get write access.
Common pitfalls
!
Replacing every pipeline with a Shortcut. Shortcuts work when the source is already in a Fabric-readable format (Delta, Parquet, Iceberg). Raw JSON in S3 still needs transformation.