Quickstart: Connect with the mssql-python driver for Python

In this quickstart, you connect a Python script to a database that you created and loaded with sample data. You use the mssql-python driver for Python to connect to your database and perform basic operations, like reading and writing data.

The mssql-python driver doesn't require any external dependencies on Windows machines. The driver installs everything that it needs with a single pip install, allowing you to use the latest version of the driver for new scripts without breaking other scripts that you don't have time to upgrade and test.

Use the local SQL authentication example in this article only for local development against a SQL Server instance that you control. For Azure SQL Database, SQL database in Fabric, shared development environments, CI, and production deployments, start with Microsoft Entra authentication or another passwordless flow.

mssql-python documentation | mssql-python source code | Package (PyPI) | Visual Studio Code

Prerequisites

• Python 3.10 or later

• If you don't already have Python, install the Python runtime and pip package manager from python.org.

• Don't want to use your own environment? Follow Container and local development to create a reproducible devcontainer or GitHub Codespaces environment.

• (Optional) Azure Command-Line Interface (CLI)

Create or connect to a database on SQL Server, Azure SQL Database, or SQL database in Fabric. Use the following steps to set up a database with the AdventureWorks2025 sample schema, and keep the connection string for later.

Create a SQL database

Create or connect to a SQL database on one of the following platforms:

Setup

Follow these steps to configure your development environment to develop an application using the mssql-python Python driver.

Install the mssql-python package

Get the mssql-python package from PyPI.

1. Open a command prompt in an empty directory.

2. Install the mssql-python package.

   • Windows
   • Alpine
   • Debian/Ubuntu
   • RHEL
   • SUSE
   • openSUSE
   • macOS

   pip install mssql-python
   

Install the python-dotenv package

Get the python-dotenv package from PyPI.

1. In the same directory, install the python-dotenv package.

   pip install python-dotenv
   

Check installed packages

You can use the PyPI command-line tool to verify that your intended packages are installed.

1. Check the list of installed packages with pip list.

   pip list
   

Run the code

• Create a new file
• Execute a query
• Insert a row as a transaction

Create a new file

1. Create a new file named app.py.

2. Add a module docstring.

   """
   Connects to a SQL database using mssql-python
   """
   

3. Import packages, including mssql-python.

   from os import getenv
   from dotenv import load_dotenv
   from mssql_python import connect
   

4. Use the mssql-python.connect function to connect to a SQL database.

   load_dotenv()
   conn = connect(getenv("SQL_CONNECTION_STRING"))
   

5. In the current directory, create a new file named .env.

6. Within the .env file, add an entry for your connection string named SQL_CONNECTION_STRING. Use one of the following examples and replace the placeholders with your actual values.

   For Azure SQL Database or SQL database in Fabric, start with Microsoft Entra authentication:

   SQL_CONNECTION_STRING="Server=<server_name>;Database=<database_name>;Encrypt=yes;TrustServerCertificate=no;Authentication=ActiveDirectoryInteractive"
   

   For local SQL Server during development, start with SQL authentication:

   SQL_CONNECTION_STRING="Server=localhost,1433;Database=<database_name>;UID=<username>;PWD=<password>;Encrypt=yes;TrustServerCertificate=yes"
   

   Caution

   Treat .env as a local development convenience, not a deployment mechanism. Never commit it, never reuse this SQL authentication sample in shared or production environments, and keep certificate validation enabled outside local development.

   Use Connection strings to adapt the sample for named instances, containers, or advanced settings. If you're connecting to Azure SQL Database or SQL database in Fabric, use Microsoft Entra authentication for passwordless and interactive sign-in options. For broader secret and certificate guidance, see Security best practices.

Execute a query

Use a SQL query string to execute a query and parse the results.

1. Create a variable for the SQL query string.

   SQL_QUERY = """
   SELECT
   TOP 5 c.CustomerID,
   c.CompanyName,
   COUNT(soh.SalesOrderID) AS OrderCount
   FROM
   SalesLT.Customer AS c
   LEFT OUTER JOIN SalesLT.SalesOrderHeader AS soh ON c.CustomerID = soh.CustomerID
   GROUP BY
   c.CustomerID,
   c.CompanyName
   ORDER BY
   OrderCount DESC;
   """
   

2. Use cursor.execute to retrieve a result set from a query against the database.

   cursor = conn.cursor()
   cursor.execute(SQL_QUERY)
   

   Note

   This function essentially accepts any query and returns a result set. To iterate over the result set, use cursor.fetchone().

3. Use cursor.fetchall with a for loop to get all the records from the database. Then, print the records.

   records = cursor.fetchall()
   for r in records:
     print(f"{r.CustomerID}\t{r.OrderCount}\t{r.CompanyName}")
   

4. Save the app.py file.

   Tip

   On macOS, both ActiveDirectoryInteractive and ActiveDirectoryDefault work for Microsoft Entra authentication. ActiveDirectoryInteractive prompts you to sign in every time you run the script. To avoid repeated sign-in prompts, sign in once through the Azure CLI by running az login, then use ActiveDirectoryDefault, which reuses the cached credential.

5. Open a terminal and test the application.

   python app.py
   

   Here's the expected output.

   29485   1       Professional Sales and Service
   29531   1       Remarkable Bike Store
   29546   1       Bulk Discount Store
   29568   1       Coalition Bike Company
   29584   1       Futuristic Bikes
   

Insert a row as a transaction

Execute an INSERT statement safely and pass parameters. Passing parameters as values protects your application from SQL injection attacks.

1. Add an import for randrange from the random library to the top of app.py.

   from random import randrange
   

2. At the end of app.py add code to generate a random product number.

   productNumber = randrange(1000)
   

   Tip

   Generating a random product number here ensures that you can run this sample multiple times.

3. Create a SQL statement string.

   SQL_STATEMENT = """
   INSERT SalesLT.Product (
   Name,
   ProductNumber,
   StandardCost,
   ListPrice,
   SellStartDate
   ) OUTPUT INSERTED.ProductID
   VALUES (%(name)s, %(product_number)s, %(standard_cost)s, %(list_price)s, CURRENT_TIMESTAMP)
   """
   

4. Execute the statement using cursor.execute.

   cursor.execute(
      SQL_STATEMENT,
      {
         'name': f'Example Product {productNumber}',
         'product_number': f'EXAMPLE-{productNumber}',
         'standard_cost': 100,
         'list_price': 200
      }
   )
   

5. Fetch the single result using cursor.fetchone, print the result's unique identifier, and then commit the operation as a transaction using connection.commit.

   result = cursor.fetchone()
   print(f"Inserted Product ID : {result.ProductID}")
   conn.commit()
   

   Tip

   Optionally, you can use connection.rollback to roll back the transaction.

6. Close the cursor and connection using cursor.close and connection.close.

   cursor.close()
   conn.close()
   

7. Save the app.py file and test the application again.

   python app.py
   

   Here's the expected output.

   Inserted Product ID : 1001
   

Next steps

Use these articles to keep building:

• Connection strings to adapt the sample for local SQL Server, Azure SQL, containers, and named instances.
• Connection management to use context managers, pooling, and connection settings.
• Troubleshooting to diagnose authentication, certificate, and connectivity problems.