Skip to main content
The OneLake destination writes data to Microsoft Fabric Lakehouses and Warehouses. It authenticates via a Service Principal, uploads Parquet files to OneLake storage, and commits Delta Lake transaction logs so data is immediately queryable through Fabric’s SQL analytics endpoint.

​
Architecture

​
Write path by item type

Fabric Lakehouses expose a read-only SQL analytics endpoint. DDL operations (CREATE TABLE, ALTER TABLE) are not supported via SQL. The OneLake destination automatically detects this and uses the Delta Lake write path instead.

​
Prerequisites

Before creating the connection, set up a Service Principal in Azure:
1

Register an application

In the Azure Portal, navigate to Microsoft Entra ID → App registrations → New registration. Name it (e.g., planasonix-onelake) and register.
2

Create a client secret

Under Certificates & secrets → New client secret, create a secret and copy the Value immediately — it is only shown once. Note the expiration date.
3

Copy identifiers

From the app registration Overview page, copy the Application (client) ID and Directory (tenant) ID.
4

Enable Fabric API access

In the Fabric Admin Portal → Tenant settings, enable “Service principals can use Fabric APIs” for your security group or the entire organization.
5

Grant workspace access

In your Fabric Workspace, click Manage access and add the service principal as a Contributor or Member. This grants write access to OneLake storage.
6

Find workspace and item IDs

Open your Lakehouse or Warehouse in the Fabric portal. The URL contains both IDs:https://app.fabric.microsoft.com/groups/{workspaceId}/lakehouses/{itemId}

​
Connection fields

Client secrets expire. Set a calendar reminder to rotate them before expiration. An expired secret silently breaks all pipelines using the connection.

​
Write modes

New Parquet files and Delta log versions are added without modifying existing data. Each pipeline run produces a new version number.Best for: incremental loads, event streams, CDC pipelines.

​
Schema management

  • Column types are inferred automatically from the first batch
  • Types are cached and reused for all subsequent batches in the same run
  • When auto schema migration is enabled, new columns in later batches trigger schema expansion
  • Supported types: string, long, double, boolean, timestamp, date

​
Performance considerations

The 1–3 minute SQL endpoint discovery delay is a Microsoft Fabric limitation. Data is physically present in OneLake immediately after the write — it just takes time for the SQL analytics endpoint to index the new Delta log version.

​
Troubleshooting

​
OneLake vs Fabric connector

The OneLake destination is functionally identical to the existing Fabric connector for write operations. It exists as a separate connection type to support independent configuration, future OneLake-specific optimizations, and clearer naming for OneLake-focused data engineering workflows.

Data warehouses

Overview of all warehouse and lakehouse connections.

Delta Lake destination

Cloud-agnostic Delta Lake writes to S3, GCS, or Azure Blob.

Credentials

Securely store and rotate service principal secrets.

Destination nodes

Write modes, pre-flight checks, and all destination node types.