Skip to content
  • There are no suggestions because the search field is empty.

Fabric Lakehouse - Target environment

Applies to
Platform: Fabric Lakehouse

Prepare a Microsoft Fabric workspace for a Fabric Lakehouse project: a lakehouse with schemas, a service principal for the deployment and the source data. Afterwards you can deploy and load the generated artifacts. The setup applies to all Fabric Lakehouse generator configurations with the project property Target Platform set to Fabric, the default.

This article shows one possible setup. Microsoft Fabric offers many other configurations, which its documentation describes. For Target Platform Azure Synapse, see Azure Synapse - Target environment.

Prerequisites

  • Access to Microsoft Fabric with a capacity on which you can create a workspace and lakehouses.
  • A service principal, an application in Microsoft Entra ID, with a client secret.

Set up the environment

Create the workspace

  1. In Fabric, select Workspaces, then New workspace.
    Workspaces panel in Fabric with the New workspace button
  2. Enter a name, for example bgDocumentation, and select Apply.
    Create a workspace dialog with the name bgDocumentation
  3. Select Manage access, then Add people or groups. Search for the service principal, choose the role Member and select Add.
    Add people dialog with the role Member

The Member role is sufficient for deploy.ps1. To keep the notebooks in Git, connect the workspace in its workspace settings under Git integration, as the Microsoft Fabric documentation describes.

Create the lakehouse

  1. In the workspace, select New item, then Lakehouse.
    Lakehouse item type
  2. Enter a name, for example bg_lakehouse_doc, select Lakehouse schemas and create the lakehouse.

The deployment creates a schema for each layer and one for the load control, with the names of the placeholder values. deploy.ps1 needs at least one lakehouse in the workspace. Targets in a lakehouse with schemas need a default lakehouse with schemas, see Fabric Lakehouse - Deployment. If the default lakehouse that deploy.ps1 sets for the notebooks has no schemas, make a lakehouse with schemas their default lakehouse.

For targets in a lakehouse without schemas, leave schema_name empty for every layer and for the load control. Also clear Generate Create Schema If Not Exists Statements in Deployment Options.

Provide source data

A Fabric Lakehouse project gets its source data in one of three ways:

  • Delta or Parquet files in OneLake, discovered directly in the project, see Discover Parquet or Delta files on OneLake.
  • Files loaded by a Fabric Lakehouse Stage Files project, imported as a linked project.
  • Database tables loaded over JDBC by a Fabric Lakehouse Stage JDBC project, imported as a linked project.

Files

Upload the source files into the Files folder of a lakehouse. In the context menu of the folder, choose Upload, then Upload folder.

Context menu of the Files folder with Upload and Upload folder

The Stage loader reads the folder <file_path><source name>, where <source name> is the name of the source model object. Put each source in its own folder with that name, for example Files/Parquet/CreditCard/CreditCard.parquet for the source CreditCard:

Files folder Parquet with one folder per source, the folder CreditCard holding CreditCard.parquet

A Delta source is a folder with the Delta table, named the same way:

Folder of a Delta source with its Delta files

The path up to the source folders is the placeholder value file_path.

Database

A Stage JDBC project reads tables of a database such as Microsoft SQL Server or PostgreSQL. The notebooks must reach the database from Fabric. For an Azure SQL Database, select Allow Azure services and resources to access this server in the networking settings of the server:

Server networking setting Allow Azure services and resources to access this server, selected

Check the result

The workspace contains the lakehouse with schemas, and Manage access lists the service principal as Member:

Manage access with the service principal as Member

After the first deployment, the explorer of the lakehouse shows the schemas below Tables:

Lakehouse bg_lakehouse_doc with the schemas dbo, DataStore, DIM_Core, DV_RDV, LoadControl, StageFiles, StageJDBC and others below Tables

Upload artifacts in Microsoft Fabric

Uploading and running the generated notebooks with deploy.ps1 is described in Fabric Lakehouse - Deployment.

Failure symptoms and recovery

Failures of the deployment, such as a default lakehouse without schemas, are described in Fabric Lakehouse - Deployment. Load failures are described in Fabric Lakehouse - Load data - Native Load Control.

Related articles