Litequery is a minimalist library for interacting with SQLite in Python. It lets you define your queries once and call them as methods. No ORM bloat, just raw SQL power, with the flexibility to operate in both asynchronous and synchronous modes.
- Simplicity: Define SQL queries in
.sql
files. No complex ORM layers. - Async first: Built for modern async Python, but also supports synchronous operations for traditional use cases.
- Flexible: Supports different SQL operations seamlessly.
pip install litequery
Create a queries.sql
file. Name your queries using comments and write them in
pure SQL.
-- name: get_all_users
SELECT * FROM users;
-- name: get_user_by_id^
SELECT * FROM users WHERE id = :id;
-- name: get_last_user_id$
SELECT MAX(id) FROM users;
-- name: insert_user<!
INSERT INTO users (name, email) VALUES (:name, :email);
-- name: delete_all_users!
DELETE FROM users;
Define your database and queries, and then call them as methods. Choose async or sync setup based on your needs. It's as straightforward as it sounds.
import litequery
import asyncio
async def main():
lq = litequery.setup("database.db", "queries.sql", use_async=True)
await lq.connect()
user_id = await lq.insert_user(name="Alice", email="[email protected]")
print(user_id)
users = await lq.get_all_users()
print(users)
user = await lq.get_user_by_id(id=user_id)
print(user)
rows_count = await lq.delete_all_users()
await lq.disconnect()
asyncio.run(main())
Litequery also supports transactions in both async and sync contexts, allowing you to execute multiple queries atomicaly.
import litequery
import asyncio
async def main():
lq = litequery.setup("database.db", "queries.sql")
await lq.connect()
try:
async with lq.transaction():
await lq.insert_user(name="Charlie", email="[email protected]")
raise Exception("Force rollback")
await lq.insert_user(name="Eve", email="[email protected]")
except Exception:
print("Transaction failed")
users = await lq.get_all_users()
print(users)
await lq.disconnect()
asyncio.run(main())
Litequery is all about simplicity and efficiency. Why wrestle with bloated ORMs when you can have raw SQL power? If you think there's a better way or have suggestions, let's hear them. Happy querying!