Microsoft Fabric

AnalyticsCreator generates Microsoft Fabric warehouse structures, orchestration pipelines and Power BI semantic-model definitions from the metadata-driven project design. The generated implementation is deployed to and executed by Microsoft Fabric without requiring an AnalyticsCreator production runtime.

Overview

AnalyticsCreator supports Microsoft Fabric as a target environment for governed data warehouse development, pipeline orchestration and analytical model generation.

Source structures, transformations, historisation rules, dependencies and analytical definitions are maintained in the AnalyticsCreator project. AnalyticsCreator uses this metadata to generate the corresponding Microsoft Fabric implementation.

AnalyticsCreator provides the design and generation layer. Microsoft Fabric provides the warehouse, OneLake storage, pipeline execution and Power BI analytical environment used by the deployed solution.

Supported Fabric Components

AnalyticsCreator can participate in Microsoft Fabric architectures using:

Fabric Warehouse for generated SQL-based data warehouse structures and processing logic
OneLake as the storage foundation used by Fabric
Fabric Data Factory pipelines for orchestration and data movement
Power BI semantic models for analytical consumption
Direct Lake where the selected semantic-model architecture uses OneLake-backed Fabric data
SQL analytics endpoints as query and analytical access surfaces where applicable

The exact Fabric components used by a project depend on the selected architecture. AnalyticsCreator does not require every Fabric implementation to use every Fabric workload.

What AnalyticsCreator Generates

Depending on the project design and deployment configuration, AnalyticsCreator can generate:

warehouse tables and schemas
views and transformation structures
stored procedures for generated loading, historisation and persisting logic
DACPAC deployment output for supported SQL deployment scenarios
Fabric Data Factory pipeline definitions
dependency-aware orchestration
Power BI semantic-model definitions
measures, relationships and other supported analytical metadata
Power BI Project output with semantic-model definitions in native TMDL

The generated artifacts are derived from the metadata maintained in AnalyticsCreator and can be regenerated when AnalyticsCreator-owned definitions change.

Fabric Warehouse

Fabric Warehouse provides the primary SQL-based warehouse target for an AnalyticsCreator Fabric implementation.

AnalyticsCreator can generate the database structures and processing logic required by the warehouse design, including tables, views and stored procedures where supported by the target Fabric SQL environment.

The generated warehouse structure can represent staging, historisation, transformation, dimensional and analytical structures according to the project design.

These layers are determined by the AnalyticsCreator model. They should not be interpreted as a fixed Fabric architecture that every project must use.

DACPAC and Database Deployment

AnalyticsCreator can generate a DACPAC from the maintained data warehouse definition.

The deployment configuration controls the target database, generated objects and relevant deployment options. The generated database implementation can then be applied to the supported Fabric SQL target using the configured deployment process.

Database generation and deployment should remain separate from pipeline execution. Deploying the warehouse structure creates or updates the technical objects; it does not itself load the business data.

OneLake

Fabric Warehouse data is stored within the Microsoft Fabric and OneLake architecture. AnalyticsCreator generates the warehouse implementation, while Fabric manages the underlying storage and execution environment.

This means AnalyticsCreator does not need to introduce a proprietary production storage layer between the generated warehouse and Microsoft Fabric.

The generated implementation remains based on native Microsoft Fabric structures and can be consumed by other Fabric and Power BI capabilities according to the permissions and interfaces provided by Microsoft.

Lakehouse and SQL Analytics Endpoints

Microsoft Fabric Lakehouses expose their Delta tables through a SQL analytics endpoint for T-SQL querying and analytical access.

The Lakehouse SQL analytics endpoint is a read-only interface for the underlying table data. It should therefore not be treated as equivalent to Fabric Warehouse for SQL-based data warehouse loading and modification.

A Lakehouse SQL analytics endpoint can still participate in an AnalyticsCreator solution as a query or consumption surface where the architecture requires it.

Where an implementation requires AnalyticsCreator-generated SQL tables, data loading and database processing, Fabric Warehouse provides the appropriate SQL-based target.

Fabric Data Factory Pipelines

AnalyticsCreator can generate Fabric Data Factory pipelines from the processing definitions and dependencies maintained in the project.

Generated pipelines can coordinate data loading and the execution of the processing steps required by the warehouse design.

Depending on the project, this can include:

source ingestion
data movement
execution of generated database processing logic
historisation where configured
persisting where configured
dependency-aware execution of downstream processing

The exact pipeline structure is determined by the AnalyticsCreator project rather than by a fixed Fabric workflow.

Pipeline Connections and Runtime Configuration

Generated pipeline definitions and environment configuration are separate concerns.

The target Fabric environment must provide the connections, credentials, permissions and other environment-specific configuration required by the generated workflow.

AnalyticsCreator generates the orchestration structure from the project metadata. Microsoft Fabric executes the deployed pipeline and manages the runtime environment.

Power BI Semantic Models

AnalyticsCreator can generate Power BI semantic-model definitions from the analytical structures maintained in the project.

Depending on the configured model, generated semantic metadata can include:

facts and dimensions
relationships
measures
hierarchies and other supported analytical definitions

For Power BI Project workflows, AnalyticsCreator can generate the semantic model and measures as native TMDL.

Report pages and visuals remain part of Power BI development and are authored separately from the AnalyticsCreator-generated semantic model.

Direct Lake

Microsoft Fabric supports Direct Lake semantic models over data stored in OneLake.

Where the AnalyticsCreator-generated Fabric implementation provides the required OneLake-backed analytical tables, these structures can be used as the data foundation for a Direct Lake semantic model.

Direct Lake is a Power BI and Microsoft Fabric execution mode. AnalyticsCreator defines and generates the analytical structures and semantic-model metadata; Fabric and Power BI provide the Direct Lake runtime.

The selected semantic-model storage mode, Fabric configuration and target model compatibility should be validated for the specific implementation.

Warehouse and Semantic Model Relationship

AnalyticsCreator maintains the relationship between the governed warehouse design and the analytical model.

Facts, dimensions, relationships, measures and other AnalyticsCreator-owned definitions can therefore be generated from the same maintained project context rather than being reconstructed independently in the reporting layer.

When the underlying design changes, the corresponding warehouse and semantic-model definitions can be regenerated and reviewed together.

Regeneration does not imply bidirectional synchronisation of every change made directly in Fabric or Power BI back into AnalyticsCreator.

Generation, Deployment and Execution

AnalyticsCreator separates design and generation from Fabric deployment and execution.

Design. Maintain source metadata, warehouse structures, transformations, historisation rules, dependencies and analytical definitions in AnalyticsCreator.
Generate. Generate the configured database, pipeline and semantic-model artifacts from the approved design.
Deploy. Apply the generated warehouse structures and orchestration artifacts to the appropriate Microsoft Fabric environment.
Configure. Apply or validate Fabric-specific connections, credentials, permissions and environment settings.
Execute. Run the deployed Fabric pipelines and warehouse processing.
Consume. Use the populated warehouse and semantic model from Power BI or other supported analytical consumers.

Generation and deployment do not themselves mean that business data has been processed. Data movement and transformation occur when the deployed Fabric workflows are executed.

Typical Fabric Workflow

A typical AnalyticsCreator Fabric implementation can follow this sequence:

import or maintain source metadata in AnalyticsCreator
define and refine the warehouse model and business logic
generate the required Fabric Warehouse structures
generate the Fabric Data Factory orchestration
deploy and configure the target Fabric environment
execute the generated pipeline workflow
validate the populated warehouse structures
generate or update the Power BI semantic model
validate the model and downstream reports

This is an example delivery flow. The exact implementation depends on the source systems, warehouse design and Fabric components selected for the project.

Changes and Regeneration

When AnalyticsCreator-owned definitions change, the corresponding warehouse, orchestration and analytical artifacts can be regenerated from the updated metadata.

This can include changes to:

tables and columns
transformations
historisation rules
persisting logic
pipeline dependencies
facts and dimensions
relationships
measures

Generated changes should be reviewed before deployment, particularly where manually maintained Fabric or Power BI objects depend on the generated structures.

Git, CI/CD and Fabric Lifecycle Management

AnalyticsCreator project definitions and generated deployment artifacts can be incorporated into controlled Git and CI/CD processes.

Teams can use this approach to:

version the AnalyticsCreator project definition
review design changes before generation
review generated database and analytical changes
use branches and pull requests for controlled development
promote approved artifacts between environments
retain an auditable history of implementation changes

Microsoft Fabric also provides its own Git integration and deployment-pipeline capabilities for supported Fabric items. These Fabric lifecycle features are separate from AnalyticsCreator generation and should be configured according to the target workspace and deployment architecture.

Source control, generation, Fabric deployment and workload execution remain separate activities.

Connectors and Source Systems

Source systems are defined through the appropriate AnalyticsCreator connectors. The generated Fabric implementation then uses the configured orchestration and target environment to load and process the required data.

Refer to the AnalyticsCreator connector documentation for supported source types and their individual configuration requirements.

Prerequisites, Limitations and Notes

A Microsoft Fabric workspace and the required Fabric capacity, permissions and target items must be available.
The selected Fabric Warehouse features must support the generated SQL objects and data types required by the project.
The Lakehouse SQL analytics endpoint is read-only for underlying table data and should not be treated as a general warehouse write target.
Pipeline connections, authentication and permissions must be configured for the target Fabric environment.
Power BI semantic-model behaviour should be validated against the selected Fabric and Power BI model configuration.
Direct Lake compatibility should be validated for the target semantic model and OneLake structures.
Generated changes should be reviewed before promotion between environments.
Manual changes to generated definitions require an agreed ownership model to avoid conflicts during regeneration.
Performance depends on the warehouse design, data volume, processing strategy, semantic-model configuration and available Fabric capacity.

Design Considerations

Use Fabric Warehouse where the implementation requires AnalyticsCreator-generated SQL tables and database processing.
Use Lakehouse SQL analytics endpoints as query and analytical surfaces rather than assuming write access to Lakehouse tables.
Keep AnalyticsCreator-owned warehouse and semantic-model definitions aligned through the maintained project design.
Select the appropriate Power BI storage mode, including Direct Lake where it fits the Fabric architecture.
Separate environment-specific Fabric configuration from the metadata-driven AnalyticsCreator design.
Review generated SQL against current Fabric Warehouse capabilities when changing the target environment or AnalyticsCreator version.

Example Use Cases

Generate a dimensional data warehouse for Fabric Warehouse.
Generate Fabric Data Factory pipelines for warehouse loading and orchestration.
Generate historised warehouse structures and their processing logic.
Generate Power BI semantic models from governed warehouse metadata.
Use OneLake-backed Fabric data with Power BI Direct Lake.
Regenerate warehouse, pipeline and semantic-model definitions when the governed design changes.
Include generated Fabric artifacts in controlled Git and CI/CD workflows.

Key Takeaway

AnalyticsCreator provides the governed design and generation layer for Microsoft Fabric data warehouse development. It generates native Fabric Warehouse structures, orchestration and Power BI semantic-model definitions from maintained metadata, while Microsoft Fabric provides OneLake storage, pipeline execution, warehouse processing and analytical runtime services.