24. How to Load SQL Query Results into a pandas DataFrame

24.2 Connection Basics: Read-Only Mindset

By Ravindra Bagale · Last updated: 23 October 2026

Why. One wrong UPDATE or DELETE from a notebook can hurt a live Blinkit-style ops table. Learners practise read first. Treat production like a museum: look, do not rearrange the exhibits.

What.

• A connection is a live pipe from Python to the database.
• Prefer a read-only DB user when your admin can create one.
• Keep host, user, and password out of Git — use environment variables or a local untracked file.
• For learning on a laptop, SQLite (one .db file) is safest: no server, easy reset.

How.

Worked example A — SQLite classroom file (no password).

import sqlite3
from pathlib import Path

db_path = Path("data/blinkit_lab.db")
db_path.parent.mkdir(parents=True, exist_ok=True)
con = sqlite3.connect(db_path)
print("connected", db_path.resolve())
con.close()

Worked example B — Secrets via environment (shape only).

import os

# Set once in the terminal (example — use YOUR practice values):
#   export DS_DB_URL="sqlite:///data/blinkit_lab.db"
url = os.environ.get("DS_DB_URL", "sqlite:///data/blinkit_lab.db")
print("using", url.split("@")[-1])  # avoid printing passwords

Worked example C — Who gets which access (fictional ops).