Configure a dbt job in Microsoft Fabric

When you create or edit a dbt job, select the dbt configurations button to open the profile setup page. Here, you define how your dbt job connects to your data warehouse. You can also change the adapter if needed and configure advanced settings to fine-tune execution behavior.

Project structure and job settings

Each dbt job in Fabric includes key tabs to help manage your project:

Explorer: View and organize files such as models, seeds, and YAML configs.
Settings: Adjust adapter configurations like schema, connection, and concurrency.
Output Panel: View run logs, job output, and error messages in real time.

The project follows a standard dbt layout:

my-dbt_project/ 
├── dbt_project.yml     # Project configuration 
├── models/             # SQL models for transformations 
│   ├── staging/ 
│   ├── marts/ 
│   └── analytics/ 
├── schema.yml          # Model tests, descriptions, and metadata 
└── seeds/              # Optional CSV data sources

dbt_project.yml defines project-level settings like model paths and configurations.
models/ contains your SQL files, each representing a model built on top of source data.
schema.yml stores tests, documentation, and relationships.
seeds/ lets you upload CSVs to use as static reference data.

Configure profile settings

Use dbt configurations to set or review your dbt profile:

Adapter: DataWarehouse (default in Fabric)
Connection name: For example, dbtsampledemowarehouse
Schema (required): For example, jaffle_shop_demo
Seed data: Optionally enable loading CSVs from /seeds as managed tables

Open your dbt job and select dbt configurations.
Confirm the Adapter (default is DataWarehouse).
Verify Connection name.
Enter Schema (for example, jaffle_shop_demo).
(Optional) Check Seed data if you want to load CSVs on dbt seed or dbt build.
Select Apply.

Change adapter

The change adapter control at the top-left of the dbt configurations page lets you change the dbt adapter used by the job's profile.

When to use it

Your workspace connection changes (for example, moving to a different Fabric Data Warehouse).
You're creating demos that contrast adapters (for example, a future PostgreSQL path), or you cloned a job and need to point it to a new target.
You're standardizing schemas across environments (dev → test → prod) and need a different connection behind the scenes.

What changes when you switch

The adapter and connection that back the profile.
Dependent fields (for example, Schema) might need revalidation.
Runtime behavior must align with the adapter’s SQL dialect and capabilities.

Advanced settings

After you configure your dbt job's profile, select Advanced Settings to fine-tune execution and run behavior. The Advanced Settings panel is split into two tabs:

General settings
Run settings

General settings

Here you can adjust project-level execution options:

Threads: Set the number of parallel threads for dbt execution (for example, 4 for medium workloads).
Fail fast: If enabled, dbt stops immediately if any resource fails to build.
Full refresh: Forces dbt to rebuild all models from scratch, ignoring incremental logic.

Select Advanced Settings > General.
Set the desired number of threads.
(Optional) Enable Fail fast or Full refresh as needed.
Select Apply to save.

Run settings

The run settings tab lets you control which models run and how to select them. There are two settings:

Run mode
Advanced selector configuration

Run mode

Run only selected models: Choose specific models to include in the run (for example, orders, stg_customers, and so on).
Run with advanced selectors: Use dbt selectors for granular control (unions, intersections, exclusions).

Advanced selector configuration

Selectors let you target parts of your pipeline for faster iteration during development or testing.

Select Advanced Settings > Run settings.
Choose your run mode:
- For simple runs, select models from the tree.
- For advanced runs, configure selectors for targeted execution.
Select Apply to save.

You can selectively run or exclude specific models by using selectors.

For example:

dbt run --select my_model 
dbt build --select staging.* 
dbt build --exclude deprecated_models

Build custom selectors by specifying:

Selector: Name your selector.
Select: Specify resources (models, tags, packages).
Exclude: List resources to skip.

Feedback

Was this page helpful?

Last updated on 2025-12-11