Forum Discussion

keataunooi's avatar
keataunooi
Regular Visitor
2 years ago
Solved

How to set default Lakehouse in the notebook programmatically?

Hi All,   Here are my workspace setup:   Workspace Fabric's Artifacts Git branch folder Fabric-Sandbox notebook: landingtoBronze lakehouse: Bronze feature/bronze123 Fabric-Dev ...
data engineering
one lake
  • Anonymous's avatar
    Anonymous
    2 years ago

    Hi keataunooi 
    Thanks for using Fabric Community.
    To programmatically attach a lakehouse for a session level, you can use this below code in your notebook:

    # Welcome to your new notebook
# Type here in the cell editor to add code!
%%configure
{
    "defaultLakehouse": {  # This overwrites the default lakehouse for current session
        "name": "your-lakehouse-name",
        #"id": "<(optional) lakehouse-id>",
        #"workspaceId": "<(optional) workspace-id-that-contains-the-lakehouse>" # Add workspace ID if it's from another workspace
    }
}


    Hope this helps. Please let me know if you have any further questions.

Anonymous's avatar
Anonymous
Not applicable
2 years ago

Hi keataunooi 
Thanks for using Fabric Community.
To programmatically attach a lakehouse for a session level, you can use this below code in your notebook:

# Welcome to your new notebook
# Type here in the cell editor to add code!
%%configure
{
    "defaultLakehouse": {  # This overwrites the default lakehouse for current session
        "name": "your-lakehouse-name",
        #"id": "<(optional) lakehouse-id>",
        #"workspaceId": "<(optional) workspace-id-that-contains-the-lakehouse>" # Add workspace ID if it's from another workspace
    }
}


Hope this helps. Please let me know if you have any further questions.

raviraoin's avatar
raviraoin
Regular Visitor
1 year ago

Thanks for sharing!

The approach works when you need to attach a single lakehouse—whether from the same workspace or an external one. However, we have a scenario that requires multiple lakehouses in one notebook:

  • Platform workspace holds our base shortcut tables.

  • Reporting workspace hosts notebooks that consume those shortcuts and persist transformed output.

  • In each notebook, we must attach:

    1. The Reporting workspace lakehouse (for writing results)

    2. One or more Platform workspace lakehouses (as input sources)

Because Fabric notebooks only support a single configuration cell, we currently rely on a post-deployment automation script to rebind the additional lakehouses for UAT/Prod. This extra step is manual and adds complexity.

Question: Is there a recommended workaround or best practice for attaching multiple lakehouses—across the same or different workspaces—directly within a Fabric notebook, without triggering a Spark session restart?

Any insights or suggestions would be greatly appreciated. Thank you!