Improvado
Esc

Start typing to search.

Improvado
Sign In

Microsoft Fabric Data source

Updated on Jul 27, 2026

‍Note: This article covers the Microsoft Fabric data source setup process. If you instead want to load data into Microsoft Fabric, see the Microsoft Fabric destination guide.

Improvado connects to your Microsoft Fabric SQL analytics endpoint (Warehouse or Lakehouse) and extracts data from your tables. Authentication uses a Microsoft Entra service principal — there is no SQL username/password.

How to connect Microsoft Fabric integration

On the Data sources page, click the Microsoft Fabric tile. You can find it using the search field or open the SQL category.

Create a Microsoft Entra service principal that Improvado will use to read your data:

  1. In the Azure portal, go to Applications > App registrations > New registration.
  2. ~Register a new application (this creates a Service Principal). Leave the Redirect URI empty.
  3. ~On the application's details page, copy the Client ID and Tenant ID.
  4. ~Under Certificates & secrets, generate a new client secret and copy its value.
  5. In your Fabric workspace, open Manage access, click Add people or groups, add the service principal, and give it at least the Viewer role (Contributor also works).
  6. ~Ask your Fabric tenant administrator to enable Service principals can use Fabric APIs in the admin portal (tenant settings) for this principal (or its security group). Without it, the connection will be rejected even with the correct credentials.

Note

Ask your Microsoft Fabric administrator if you don't have permission to register applications or grant workspace access.

On the Microsoft Fabric authorization page, fill in the following fields and click Authorize:

  1. ~SQL Endpoint — the SQL connection string of your Fabric item (Fabric > your Warehouse/Lakehouse > Settings > SQL connection string), for example <id>.datawarehouse.fabric.microsoft.com.
    1. Warehouse / Lakehouse name — the item name (Initial Catalog).
    2. Schema — defaults to dbo.
  2. ~Client ID, Client Secret, and Tenant ID from Step 2.

Make sure the service principal has the following permissions:

Access level Scope
Read access INFORMATION_SCHEMA and its views (e.g. TABLES, COLUMNS)
The Warehouse/Lakehouse specified in the form and the tables inside it

How to extract data from Microsoft Fabric

Step 1. Select table

  1. Select a table from the schema you specified during authorization.
  2. Click Continue to proceed to the next step.

Note

Prefer to write your own query — for example to join, filter, or aggregate before extraction? Use the Custom SQL Query option instead of picking a single table. See Custom query for SQL Flat Data.

Step 2. Match

  1. Enter your Data Table name. Remember, you can't change this later.
  2. Configure your Extraction scheduling — how often your data is extracted to Improvado storage.
  3. Select the fields you need to import. By default, all fields are selected.
  4. Check the field type. Each field can be assigned a String, Number, or Date type.
  5. ~The date field is a unique date key and should be singular; choose from a range of date formats for it. When a date field is set, extraction is filtered by the selected date range.
  6. ~If you have a field with numeric values you don't want to include in calculations, change its type from Number to Text.
  7. The Skip rule lets you exclude empty rows or rows with specific values in a selected field.
  8. Choose the Write strategy used when uploading data into Improvado Storage. Learn more.
  9. Click the Import N fields button to continue.

Note

Learn more about working with dates in Flat Data here.

Step 3. Preview

  1. Check the structure of the data that will be extracted.
  2. Once you've reviewed the fields, click Continue to confirm the settings and start data extraction.

Connection Details page

You can find your Microsoft Fabric connection on the Connections page. The Overview tab shows the connection status, and the Settings tab lets you change the extraction scheduling and table meta for your data extraction.

Was this article helpful?