How to Create a Lakehouse in Microsoft Fabric: Step by Step

João Barros 04 de July de 2026 3 min read

A Lakehouse combines the flexibility of a data lake with the organisation of a data warehouse, and in Microsoft Fabric it is the ideal starting point for storing and analysing data. Creating a Lakehouse in Microsoft Fabric, loading a CSV file into a table and querying it with SQL are tasks you can complete in a few minutes, through the interface and without writing any infrastructure code. This guide walks through each step, simply, for anyone just getting started.

Prerequisites

An account with access to Microsoft Fabric and an active capacity (the free trial works).
A workspace where you have permission to create items.
A simple CSV file on your computer, with a header on the first line (for example, sales.csv).

Step 1: Create the Lakehouse

In your workspace, open the Data Engineering experience and click + New item. Choose Lakehouse, give it a clear name such as sales_lakehouse and confirm. Within a few seconds, Fabric creates three things at once: the Lakehouse itself, a SQL analytics endpoint (read-only) and a default semantic model for Power BI.

How to Create a Lakehouse in Microsoft Fabric: Step by Step

Every piece of data you place here is stored in OneLake, Fabric's single storage layer, in an open format. When the Lakehouse opens, the Explorer on the left shows two sections: Tables (Delta tables, ready for analysis) and Files (raw files, such as your CSV).

Step 2: Upload the CSV file

Hover over the Files section, click the three dots (...) and choose Upload and then Upload files. Select your sales.csv and confirm. The file now appears inside Files. Note that, at this point, it is just text stored in OneLake: it is not yet a table you can query with SQL.

Step 3: Load the CSV into a Delta table

On the sales.csv file, click the three dots and choose Load to Tables and then New table. Confirm the table name (for example, sales) and click load. Fabric reads the CSV, converts it to Delta format and places it in the Tables section. This is the most important step in the tutorial: only Delta tables become available to query in the SQL analytics endpoint.

Tip: make sure the CSV has a header row. Fabric uses it to name the table columns automatically.

Step 4: Query the data with SQL

In the top-right corner, switch from Lakehouse mode to SQL analytics endpoint. Open a New SQL query and run a simple query to see the first rows:

SELECT TOP (10) *
FROM sales;

Next, try an aggregation to draw a quick conclusion from the data, for example the number of sales per city:

SELECT city, COUNT(*) AS total_sales
FROM sales
GROUP BY city
ORDER BY total_sales DESC;

Adjust the column names to match your file. Remember that this endpoint is read-only: it is for querying data and creating views, not for inserting or deleting rows.

Check the result

You know it worked if: the sales table appears in the Tables section with a table icon; the first query returns 10 rows with your CSV columns; and the aggregation shows counts per city. If the table does not appear, you most likely left the file only in Files — repeat Step 3 using Load to Tables.

Conclusion

With four steps you created a Lakehouse, turned a CSV into a Delta table and queried it with SQL — the foundation of almost any data project in Fabric. The natural next step is to build a report in Power BI from the Lakehouse's semantic model, or to automate the load with a data pipeline. Which of your real data files will you load first?