Connect Python to SQL Server Azure and run basic queries

 Connect Python to SQL Server Azure and run basic queries

Prerequisites
pyodbc
ODBC Driver 18 for SQL Server

1.install pyodbc

from command prompt run the below command, pyodbc will add into .venv  
c:\project\database_agent\uv add pyodbc

validation
uv run python -c "import pyodbc; print(pyodbc.version)"

2. install ODBC Driver 18 for SQLServer

Download ODBC Driver for SQL Server - ODBC Driver for SQL Server | Microsoft Learn

python script to connect toSQL Azure and run basic queries

a). create python script sql_connect.py

import pyodbc

SERVER = "your-server-name.database.windows.net"
DATABASE = "your-database-name"
USERNAME = "your-username"
PASSWORD = "your-password"

conn_str = (
    "DRIVER={ODBC Driver 18 for SQL Server};"
    f"SERVER={SERVER};"
    f"DATABASE={DATABASE};"
    f"UID={USERNAME};"
    f"PWD={PASSWORD};"
    "Encrypt=yes;"
    "TrustServerCertificate=no;"
    "Connection Timeout=30;"
)

with pyodbc.connect(conn_str) as conn:
    cursor = conn.cursor()
    cursor.execute("SELECT @@VERSION")
    print(cursor.fetchone()[0])

b) Run the command from Visual Code terminal
(database-agent) C:\project\database_agent> python sql_connect.py

It shows the result as below

(database-agent) C:\project\database_agent> python sql_connect.py
Microsoft SQL Azure (RTM) - 12.0.2000.8 
        Oct  2 2025 00:38:42
        Copyright (C) 2025 Microsoft Corporation

Setup local dev environment 2 (UV, python virtual environment, python in Visual Code)

Setup local dev environment 2 (UV, python virtual environment, python in Visual Code)

UV

We use uv instead of pip for package management because uv is faster, safer, and manages dependencies per project automatically.

UV creates virtual environments, installs packages, locks dependencies, and runs Python — automatically and safely per project.

uv replaces pip + venv + pip-tools with a single, much faster tool.

Key idea:

You work in a project folder, and uv makes sure the right Python + the right packages + the right environment are always used — even without activating a virtual environment.

1. download and install UV for window 

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Reference Installation | uv

2. create folder prepare for the project

C:\project\database_agent

3) Inside c:\project\database_agent folder, run the below

UV init

it Creates a modern Python project folder structure as below inside c:\project\database_agent folder:

database_agent/
├── .venv/
├── pyproject.toml
├── README.md
└── .gitignore

4) Creates the project’s Python virtual environment (.venv) and prepares it to be used by uv.
uv init

database_agent/
├── .venv/
├── pyproject.toml
├── README.md
└── .gitignore
└── .venv

5) Open Visual code, file->open folder c:\project\database_agent

6) in Visual Code- >extention - > install python and pylance

7) Setup python interprate
   visual code -> view- > palatte -> python: select interpreter

8) select interpreter from environment create by UV just now;
   python 3.13.11 is the python version i create with UV 
   python 3.13.11 (database agent) .\venv\scripts\python.exe

9) Validate we use the right virtual environment
   Visual code-> Terminal -> New terminal
   If the following is shown, the environment is set up correctly and the virtual environment we created is being used. 

   (database-agent) C:\project\database_agent>
 

Deploy Azure SQL or SQL Server on Azure VM

 Deploy Azure SQL or SQL Server on Azure VM

from Azure database, select create Database - > SQL database (free offer)

if server not exists yet, create new server

back to resource group, that is one sql database service and window service create under the resource group

get SQL database connection 

download and install ssms

Install SQL Server Management Studio | Microsoft Learn

test connection connect to SQL Azure 

Create Azure account, resource groups, access mode

 

Create Azure account, resource groups, access mode

1.create subscription

2, Create Azure account

Create Your Azure Free Account Or Pay As You Go | Microsoft Azure

Try Azure for free to create account

3. Create resource group inside Azure portal

4. click on resource name

5 . click on access control (IAM) to explore access mode; 

Setup local dev environment (Python, IDE, Git)

 

Setup local dev environment (Python, IDE, Git)

1. download and install Visual Code

Download Visual Studio Code - Mac, Linux, Windows

2. download GIT

Git

from command prompt - Run the below

C:\>git --version

git version 2.51.1.windows.1

3. Install and install Python

Python Releases for Windows | Python.org

🗓 30-Day High-Level Plan Build an AI-enabled Cloud Database Agent in 30 days

📅 Week 1 – Cloud, Database & Automation Foundation

Day	Focus Area	High-Level Task	status
1	Environment	Setup local dev environment (Python, IDE, Git)	done 6/Jan
2	Azure Basics	Create Azure account, resource groups, access model	done 6/Jan
3	Database	Deploy Azure SQL or SQL Server on Azure VM	done 6/Jan
4	Connectivity	Connect Python to SQL Server and run basic queries	done 7/Jan
5	Automation	Automate core DB operations (create, query, maintain)
6	Azure Automation	Automate DB lifecycle using Azure CLI / PowerShell
7	Stabilization	Review scripts, logging, error handling

---

📅 Week 2 – GenAI + Prompt → SQL + RAG

Day	Focus Area	High-Level Task
8	LLM Setup	Enable Azure OpenAI and call LLM from Python
9	Prompt Design	Design prompts for natural language → SQL
10	SQL Copilot	Build basic SQL Copilot (NL → SQL → Execute)
11	Embeddings	Create embeddings for schema & runbooks
12	RAG	Implement retrieval-augmented SQL generation
13	Reasoning	Enable multi-step reasoning for complex queries
14	Optimization	Refine prompts, add execution logging

---

📅 Week 3 – AI Agent + Automation

Day	Focus Area	High-Level Task
15	Agent Core	Initialize AI Agent framework and control loop
16	Tooling	Register SQL, PowerShell, Azure CLI as Agent tools
17	Provisioning	Agent-driven database provisioning
18	DR / Backup	Agent-driven backup and restore operations
19	Incident Ops	Agent-assisted incident diagnosis and recommendations
20	End-to-End	Chain reasoning → action → feedback
21	Governance	Add guardrails, approval flow, safety controls

---

📅 Week 4 – UI, Demo & Presentation

Day	Focus Area	High-Level Task
22	UI	Build minimal UI for interacting with the Agent
23	Integration	Connect UI to Agent workflows
24	Observability	Add logging, transparency, execution feedback
25	Security	Apply role separation and safe-execution rules
26	Knowledge	Enhance Agent with runbooks and SOP knowledge
27	Scenarios	Prepare real-world demo scenarios
28	Demo	Record demo video and walkthrough
29	Documentation	Create architecture and usage documentation
30	Finalization	Polish solution and prepare interview narratives