Use mssql-python with Flask

Flask is a lightweight Python web framework that gives you full control over application structure. Combined with mssql-python, you can build web applications and REST APIs backed by Microsoft SQL and Azure SQL Database with minimal overhead.

Prerequisites

• Python 3.10 or later.
• The mssql-python and flask packages. Install both with pip install flask mssql-python.
• Install one-time operating system specific prerequisites. Windows users can skip this step. For full platform details, see Install mssql-python.
  • Alpine
  • Debian/Ubuntu
  • RHEL
  • SUSE
  • openSUSE
  • macOS

  apk add libtool krb5-libs krb5-dev
  

Create a SQL database

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

The examples in this article use the AdventureWorksLT sample database, specifically the SalesLT.Product table. If you don't have AdventureWorksLT installed, see AdventureWorks sample databases.

Project setup

Install dependencies

Install the required packages with pip:

pip install flask mssql-python

Project structure

Organize your project with separate modules for configuration, connection management, routes, and tests:

my_app/
├── app.py            # Flask app and routes
├── config.py         # database settings
├── database.py       # connection lifecycle
├── test_app.py       # pytest tests
└── blueprints/       # optional: routes grouped into modules
    ├── __init__.py
    └── products.py

Database connection management

Flask doesn't include a built-in database layer, so you manage connections directly. The pattern in this section stores one connection per request on Flask's g object and closes it automatically when the request ends.

Create config.py

Centralize database settings in a configuration class. Environment variables let you override defaults without changing code.

# config.py
import os

class Config:
    """Application configuration."""
    DATABASE_SERVER = os.getenv("DB_SERVER", "<server>.database.windows.net")
    DATABASE_NAME = os.getenv("DB_NAME", "<database>")
    POOL_SIZE = int(os.getenv("DB_POOL_SIZE", "10"))

Create database.py

The database.py module manages the connection lifecycle. Flask's g object is a per-request namespace, so storing the connection there ensures each request gets its own connection that's cleaned up when the request finishes.

The get_connection_string() function builds the connection string from the app config. The get_db() function creates a connection on first call and reuses it for the rest of the request. The close_db() function runs automatically at the end of each request, rolling back the transaction if an exception occurred and committing otherwise. The init_app() function registers this teardown behavior with the Flask app.

# database.py
import mssql_python
from flask import g, current_app

def get_connection_string() -> str:
    """Build connection string from Flask app config."""
    cfg = current_app.config
    return (
        f"Server={cfg['DATABASE_SERVER']};"
        f"Database={cfg['DATABASE_NAME']};"
        "Authentication=ActiveDirectoryDefault;"
        "Encrypt=yes"
    )

def get_db():
    """Get a database cursor for the current request.

    The connection is stored on Flask's g object so it persists
    for the duration of the request and is reused across calls.
    """
    if "db_conn" not in g:
        g.db_conn = mssql_python.connect(get_connection_string())
        g.db_cursor = g.db_conn.cursor()
    return g.db_cursor

def close_db(exception=None):
    """Close the database connection at the end of the request."""
    cursor = g.pop("db_cursor", None)
    conn = g.pop("db_conn", None)

    if cursor is not None:
        cursor.close()
    if conn is not None:
        if exception:
            conn.rollback()
        else:
            conn.commit()
        conn.close()

def init_app(app):
    """Register database teardown with the Flask app."""
    app.teardown_appcontext(close_db)

Flask application

The following example shows a complete Flask application with routes for listing, retrieving, creating, updating, and deleting products.

Create app.py

The application module creates the Flask app, loads configuration, and registers the database teardown. Each route function calls get_db() to get a cursor, executes queries with parameterized SQL (using %(name)s placeholders and a dictionary of values), and returns JSON responses.

# app.py
from flask import Flask, jsonify, request, abort
from config import Config
from database import init_app, get_db

app = Flask(__name__)
app.config.from_object(Config)
init_app(app)

@app.route("/")
def index():
    return jsonify({"message": "Product API", "docs": "/products"})

@app.route("/products")
def list_products():
    """List products with pagination."""
    page = request.args.get("page", 1, type=int)
    page_size = request.args.get("page_size", 10, type=int)
    skip = (page - 1) * page_size

    cursor = get_db()

    cursor.execute("SELECT COUNT(*) FROM SalesLT.Product")
    total = cursor.fetchval()

    cursor.execute("""
        SELECT ProductID, Name, ProductNumber, ListPrice, Color, ProductCategoryID
        FROM SalesLT.Product
        ORDER BY ProductID
        OFFSET %(skip)s ROWS
        FETCH NEXT %(limit)s ROWS ONLY
    """, {"skip": skip, "limit": page_size})

    items = [{
        "id": row.ProductID,
        "name": row.Name,
        "product_number": row.ProductNumber,
        "price": float(row.ListPrice),
        "color": row.Color,
        "category_id": row.ProductCategoryID
    } for row in cursor.fetchall()]

    return jsonify({
        "items": items,
        "total": total,
        "page": page,
        "page_size": page_size,
        "pages": (total + page_size - 1) // page_size
    })

@app.route("/products/<int:product_id>")
def get_product(product_id):
    """Get a single product by ID."""
    cursor = get_db()
    cursor.execute("""
        SELECT ProductID, Name, ProductNumber, ListPrice, Color, ProductCategoryID
        FROM SalesLT.Product
        WHERE ProductID = %(id)s
    """, {"id": product_id})

    row = cursor.fetchone()
    if not row:
        abort(404)

    return jsonify({
        "id": row.ProductID,
        "name": row.Name,
        "product_number": row.ProductNumber,
        "price": float(row.ListPrice),
        "color": row.Color,
        "category_id": row.ProductCategoryID
    })

@app.route("/products", methods=["POST"])
def create_product():
    """Create a new product."""
    data = request.get_json()
    if not data:
        abort(400)

    cursor = get_db()

    # OUTPUT INSERTED returns the new row's columns in the same statement,
    # so you don't need a separate SELECT to get the generated ID and defaults.
    # ProductNumber is required and unique. StandardCost and SellStartDate are
    # also NOT NULL in SalesLT.Product, so supply values for them.
    cursor.execute("""
        INSERT INTO SalesLT.Product
            (Name, ProductNumber, ListPrice, Color, Size, ProductCategoryID, StandardCost, SellStartDate)
        OUTPUT INSERTED.ProductID, INSERTED.Name, INSERTED.ProductNumber, INSERTED.ListPrice,
               INSERTED.Color, INSERTED.ProductCategoryID
        VALUES (%(name)s, %(product_number)s, %(price)s, %(color)s, %(size)s, %(category_id)s, 0, GETDATE())
    """, {
        "name": data["name"],
        "product_number": data["product_number"],
        "price": data["price"],
        "color": data.get("color"),
        "size": data.get("size"),
        "category_id": data["category_id"]
    })

    row = cursor.fetchone()
    return jsonify({
        "id": row.ProductID,
        "name": row.Name,
        "product_number": row.ProductNumber,
        "price": float(row.ListPrice),
        "color": row.Color,
        "category_id": row.ProductCategoryID
    }), 201

@app.route("/products/<int:product_id>", methods=["PUT"])
def update_product(product_id):
    """Update an existing product."""
    data = request.get_json()
    if not data:
        abort(400)

    cursor = get_db()

    updates = []
    params = {"id": product_id}

    for field in ("name", "product_number", "price", "color", "category_id"):
        if field in data:
            col = {"name": "Name", "product_number": "ProductNumber",
                   "price": "ListPrice", "color": "Color",
                   "category_id": "ProductCategoryID"}[field]
            updates.append(f"{col} = %({field})s")
            params[field] = data[field]

    if not updates:
        abort(400)

    cursor.execute(f"""
        UPDATE SalesLT.Product SET {', '.join(updates)}
        OUTPUT INSERTED.ProductID, INSERTED.Name, INSERTED.ProductNumber, INSERTED.ListPrice,
               INSERTED.Color, INSERTED.ProductCategoryID
        WHERE ProductID = %(id)s
    """, params)

    row = cursor.fetchone()
    if not row:
        abort(404)

    return jsonify({
        "id": row.ProductID,
        "name": row.Name,
        "product_number": row.ProductNumber,
        "price": float(row.ListPrice),
        "color": row.Color,
        "category_id": row.ProductCategoryID
    })

@app.route("/products/<int:product_id>", methods=["DELETE"])
def delete_product(product_id):
    """Delete a product."""
    cursor = get_db()
    cursor.execute("DELETE FROM SalesLT.Product WHERE ProductID = %(id)s", {"id": product_id})
    if cursor.rowcount == 0:
        abort(404)
    return "", 204

@app.route("/health")
def health_check():
    """Check database connectivity."""
    try:
        cursor = get_db()
        cursor.execute("SELECT 1")
        return jsonify({"status": "healthy", "database": "connected"})
    except Exception as e:
        return jsonify({"status": "unhealthy", "error": str(e)}), 503

Run the application

Start the development server:

flask --app app run --debug --port 5000

The server listens on http://localhost:5000. Open a second terminal and call the endpoints by using curl to confirm the app is talking to your database:

# Check database connectivity
curl http://localhost:5000/health

# List the first page of products
curl "http://localhost:5000/products?page_size=5"

# Get a single product by ID
curl http://localhost:5000/products/680

Each endpoint returns JSON. You can also open http://localhost:5000/products in a browser to view the paginated list.

Connection pooling

Without connection pooling, each request opens and closes a TCP connection to Microsoft SQL, which adds latency. Connection pooling keeps a set of idle connections ready for reuse. To enable connection pooling call mssql_python.pooling() once at module level. With pooling enabled, conn.close() in the close_db teardown returns the connection to the pool instead of closing it.

Enable connection pooling

Enable pooling by calling mssql_python.pooling() at module level before any connections are opened:

# database.py with connection pooling
import mssql_python
from flask import g, current_app

# Configure pool at module level
mssql_python.pooling(max_size=20, idle_timeout=300)

def get_db():
    """Get a database cursor with connection pooling."""
    if "db_conn" not in g:
        g.db_conn = mssql_python.connect(get_connection_string())
        g.db_cursor = g.db_conn.cursor()
    return g.db_cursor

Error handling

Flask lets you register handlers for specific exception types. Catching mssql_python.DatabaseError and mssql_python.IntegrityError lets you return structured JSON error responses instead of default HTML error pages.

Register error handlers

Add these handlers to the existing app.py, after the app = Flask(__name__) line. Because the handlers reference the app object, they must come after the app is created. app.py needs import mssql_python at the top. The handlers return structured JSON responses instead of default HTML error pages:

# app.py
import mssql_python

@app.errorhandler(mssql_python.DatabaseError)
def handle_database_error(error):
    """Handle database errors."""
    return jsonify({"error": "Database error occurred"}), 500

@app.errorhandler(mssql_python.IntegrityError)
def handle_integrity_error(error):
    """Handle integrity constraint violations."""
    error_msg = str(error)
    if "UNIQUE" in error_msg:
        return jsonify({"error": "Resource already exists"}), 409
    if "FOREIGN KEY" in error_msg:
        return jsonify({"error": "Referenced resource not found"}), 400
    return jsonify({"error": "Data integrity error"}), 400

@app.errorhandler(404)
def not_found(error):
    return jsonify({"error": "Resource not found"}), 404

@app.errorhandler(400)
def bad_request(error):
    return jsonify({"error": "Bad request"}), 400

Blueprints

As your application grows, putting all routes in a single file becomes hard to maintain. Flask Blueprints let you group related routes into separate modules that are registered with the app.

Organize routes with blueprints

Create a blueprint module for product routes that imports get_db and defines endpoints under a shared URL prefix:

# blueprints/products.py
from flask import Blueprint, jsonify, request, abort
from database import get_db

products_bp = Blueprint("products", __name__, url_prefix="/api/products")

@products_bp.route("/")
def list_products():
    """List all products."""
    cursor = get_db()
    cursor.execute("""
        SELECT ProductID, Name, ListPrice, Color, ProductCategoryID
        FROM SalesLT.Product ORDER BY ProductID
    """)
    return jsonify([{
        "id": row.ProductID,
        "name": row.Name,
        "price": float(row.ListPrice),
        "color": row.Color,
        "category_id": row.ProductCategoryID
    } for row in cursor.fetchall()])

@products_bp.route("/<int:product_id>")
def get_product(product_id):
    """Get a product by ID."""
    cursor = get_db()
    cursor.execute(
        "SELECT ProductID, Name, ListPrice, Color FROM SalesLT.Product WHERE ProductID = %(id)s",
        {"id": product_id}
    )
    row = cursor.fetchone()
    if not row:
        abort(404)
    return jsonify({"id": row.ProductID, "name": row.Name, "price": float(row.ListPrice), "color": row.Color})

Register the blueprint

Save the blueprint as blueprints/products.py, and add an empty blueprints/__init__.py file so Python treats the folder as a package. Then, in app.py, import the blueprint with your other imports and register it after the app = Flask(__name__) line:

# app.py
from blueprints.products import products_bp

app.register_blueprint(products_bp)

Because the blueprint sets url_prefix="/api/products", its routes are served under that prefix. For example, the list route is available at http://localhost:5000/api/products/, separate from the /products routes defined directly in app.py.

Testing

Flask provides a test client that sends requests to your application without starting a real HTTP server. Use pytest fixtures to create the client and reuse it across tests.

Test setup with pytest

Create a pytest fixture that provides a test client and write tests to verify route behavior:

# test_app.py
import uuid

import pytest
from app import app

@pytest.fixture
def client():
    app.config["TESTING"] = True
    with app.test_client() as client:
        yield client

def test_health_check(client):
    response = client.get("/health")
    assert response.status_code == 200
    data = response.get_json()
    assert data["status"] == "healthy"

def test_list_products(client):
    response = client.get("/products")
    assert response.status_code == 200
    data = response.get_json()
    assert "items" in data
    assert "total" in data

def test_create_product(client):
    suffix = uuid.uuid4().hex[:8]
    name = f"Test Product {suffix}"
    response = client.post("/products", json={
        "name": name,
        "product_number": f"TEST-{suffix}",
        "price": 19.99,
        "category_id": 18
    })
    assert response.status_code == 201
    data = response.get_json()
    assert data["name"] == name

def test_get_product_not_found(client):
    response = client.get("/products/99999")
    assert response.status_code == 404

These tests run against your live database rather than mocks, so test_create_product inserts a real row into SalesLT.Product. In AdventureWorksLT, both Name and ProductNumber have unique constraints, so the test generates a unique value for each on every run. If you hardcode those values instead, the test fails with a conflict on the second run unless you delete the row first.

Run the tests

Save the tests as test_app.py in your project folder. With your virtual environment activated, install pytest and run it from that folder. Installing and running pytest inside the same virtual environment as flask and mssql-python ensures the tests import the packages your app uses. pytest automatically discovers test_app.py and reports the results:

pip install pytest
pytest

pytest discovers test_app.py automatically and reports the results:

==================== test session starts ====================
collected 4 items

test_app.py ....                                       [100%]

===================== 4 passed in 3.21s =====================

Related content

• Connection management
• Connection pooling
• Error handling
• FastAPI integration
• Microsoft Entra authentication