OneLake Shortcuts Explained

OneLake shortcuts in Microsoft Fabric create virtual references to data stored in external locations — Azure Data Lake Storage, Amazon S3, Google Cloud Storage, or other Fabric workspaces — making that data appear as native Lakehouse tables or files without copying a single byte, which eliminates the cost, latency, and pipeline complexity of traditional data movement approaches. If your organization has data scattered across multiple cloud storage accounts and you need unified analytics without building ETL pipelines, shortcuts are the answer.

In my 25+ years implementing enterprise data platforms, I have watched organizations spend millions building and maintaining data movement pipelines whose sole purpose is copying data from one storage system to another so that an analytics tool can read it. OneLake shortcuts eliminate this entire class of infrastructure for read-only analytics use cases. Our Microsoft Fabric consulting team has used shortcuts to connect existing multi-million-dollar investments in ADLS Gen2 and S3 storage to Fabric without moving a single byte, saving clients 40-60% of their ETL pipeline maintenance costs.

How Shortcuts Work

Architecture

A shortcut is a metadata pointer stored in OneLake that redirects read requests to the actual data location. When a Spark notebook, SQL query, or Direct Lake Power BI model references a shortcut table, the Fabric compute engine reads data directly from the source storage through configured authentication credentials. No intermediate copy is created — reads go straight to the source.

Key Characteristics

Tables vs Files Shortcuts

Shortcuts can be created in two locations within a Lakehouse:

Tables section: The shortcut target must contain Delta-formatted data. The data appears as a queryable table with schema, column types, and support for SQL analytics endpoint queries, Spark DataFrame operations, and Direct Lake Power BI models. This is the most common shortcut type for analytics workloads.

Files section: The shortcut target can contain any file format — Parquet, CSV, JSON, images, documents, or any other binary format. Files appear in the Lakehouse file browser but are not automatically queryable as tables. You must read them explicitly in Spark notebooks or register them as external tables using CREATE TABLE with a LOCATION clause.

Supported Data Sources

Azure Data Lake Storage Gen2 (ADLS)

The most common shortcut source for organizations with existing Azure data lakes:

• Authentication options: Account key, SAS token, or service principal (OAuth). Service principal is recommended for production — it supports credential rotation and RBAC without sharing storage keys
• Path configuration: Specify the storage account URL, container name, and folder path. You can shortcut to a specific subfolder within a container
• DFS endpoint: Use the DFS endpoint format (https://account.dfs.core.windows.net) rather than the blob endpoint for optimal performance

Common pattern: Organizations with years of data in ADLS Gen2 create shortcuts to their existing Delta Lake tables, making them instantly queryable in Fabric without any data migration. I have implemented this pattern for a healthcare client with 8 TB of patient analytics data in ADLS — within 2 hours, the data was accessible through Fabric SQL analytics endpoints and Power BI reports.

Amazon S3

For multi-cloud organizations with data in AWS:

• Authentication: IAM access key and secret key, or cross-account IAM role (recommended for production)
• Region considerations: S3 data in us-east-1 accessed by a Fabric capacity in East US will have minimal latency. Cross-region or cross-continent shortcuts will have noticeable latency for interactive queries
• Format requirement: S3 data must be in Delta format for Tables shortcuts or any format for Files shortcuts
• Egress costs: AWS charges for data egress from S3 to Azure. For large-scale, frequent access, consider whether the egress cost justifies not copying the data

Google Cloud Storage

• Similar pattern to S3 with HMAC keys for authentication
• Same egress cost considerations as S3
• Less commonly used than ADLS and S3 shortcuts in our enterprise engagements

OneLake to OneLake (Cross-Workspace)

The zero-cost shortcut option for data sharing within your Fabric tenant:

• No egress charges: OneLake-to-OneLake shortcuts have no data movement or egress costs
• Read-write support: Unlike external shortcuts, OneLake shortcuts support both read and write operations
• Use case: Share a curated Gold-layer table from the data engineering workspace with the analytics workspace without duplicating data. The analytics team's Power BI reports and SQL queries access the data through the shortcut, and the data engineering team maintains the single source

Enterprise Architecture Patterns

Data Mesh with Shortcuts

Shortcuts enable a data mesh architecture where domain teams own their data products but share them across the organization:

• Domain workspaces: Each domain (Sales, Finance, Operations, HR) owns a Fabric workspace with their curated data products (Delta tables in their Lakehouse)
• Consumer workspaces: Analytics teams create shortcuts to the domain data products they need, composing cross-domain datasets without data duplication
• Governance: Each domain team controls access to their workspace. Shortcut creation requires the consumer to have appropriate permissions on the source workspace

Hybrid Cloud Analytics

For organizations operating across Azure and AWS:

• Data engineering runs in Databricks on AWS, producing Delta tables in S3
• Analytics and BI runs in Microsoft Fabric on Azure
• Shortcuts connect S3 Delta tables to Fabric Lakehouses
• Power BI reports query the shortcut tables through Direct Lake mode or SQL analytics endpoints
• No ETL pipeline between clouds — shortcuts provide the bridge

Migration Acceleration

Shortcuts serve as a migration bridge when moving from ADLS-based architectures to Fabric:

1. Phase 1: Create shortcuts from Fabric Lakehouses to existing ADLS Delta tables. Reports immediately work against Fabric
2. Phase 2: Migrate ETL pipelines from Azure Data Factory/Synapse to Fabric notebooks and pipelines
3. Phase 3: Write output directly to Fabric Lakehouses instead of ADLS. Remove shortcuts as data is produced natively in Fabric
4. Phase 4: Decommission ADLS storage accounts

This phased approach eliminates the big-bang migration risk and allows the organization to validate Fabric workloads incrementally.

Performance Optimization

Minimizing Latency

• Co-locate storage and compute: Create shortcuts to storage in the same Azure region as your Fabric capacity
• Use Delta format: Delta tables with statistics enable predicate pushdown, reducing the amount of data read from the source
• Partition alignment: If the source data is partitioned by date, Fabric leverages partition pruning to read only relevant partitions
• Cache warming: For frequently accessed shortcuts, initial queries populate the OneLake cache, improving subsequent query performance

Monitoring

Monitor shortcut performance through Fabric capacity metrics:

• Track read latency per shortcut to identify slow external sources
• Monitor data egress volume to manage cloud provider costs
• Alert on authentication failures that indicate expired credentials

Security and Governance

Credential Management

• Store shortcut credentials in a managed credential store (not in plain text)
• Rotate access keys and SAS tokens on a schedule (90-day rotation recommended)
• Use service principals with short-lived tokens where possible
• Audit credential access through Azure AD sign-in logs

Access Control

Shortcut access is governed by Fabric workspace permissions:

• Users with Viewer role in the workspace can read shortcut data through reports
• Users with Contributor or higher role can create new shortcuts
• Source-side permissions are enforced by the shortcut credentials (the shortcut authenticates to the source, and source-side RBAC applies)

For a complete security framework, see our Fabric security guide and Power BI governance framework.

Ready to eliminate data silos with OneLake shortcuts? Contact our Fabric team for architecture planning and implementation.

OneLake Shortcut Performance Considerations

Shortcuts are not free — understanding their performance characteristics prevents surprises:

• Cross-region shortcuts add 50-200ms latency per query depending on distance. Keep shortcuts within the same Azure region when possible.
• External shortcuts (S3, GCS) incur egress costs from the source cloud provider. A single Power BI report querying 10GB from S3 via shortcut costs approximately $0.90 per full scan.
• Shortcut refresh: Shortcuts to ADLS Gen2 and S3 reflect source changes within minutes. Shortcuts to other Fabric lakehouses reflect changes in near-real-time.
• Security inheritance: Shortcuts inherit the security context of the requesting workspace, not the source. Configure access carefully to prevent unintended data exposure.

For help designing your OneLake shortcut architecture, contact our team.