SQLite

Notes on SQLite3

Author

Published

July 9, 2024

Disclaimer: if you’re reading this, I’m not a database expert. These are kinda notes to myself (and potentially others who stumble upon them), and they’re intended to organize my thoughts. Make of them what you will

What is SQLite?

SQLite is basically a small, self-contained database that can be embedded in other applications. It’s useful for applications or devices that don’t need or want to connect to an external database like Postgres or some cloud offering.

SQLite is nice for lots of reasons – it’s easy to install (see below), it’s typically just something you link into your application via a library, it’s minimalistic but also feature-rich (e.g. it has CTEs, full-text search, JSON compatibility). It reminds me of DuckDB is a lot of ways, but DuckDB is more intended for analytic, column-oriented workflows whereas SQLite is more intended for row-oriented operations.

That said, one of SQLite’s strengths (that it’s bundled with the server) is also one of its weaknesses. That is, if the application server explodes, so does the database. Pre-iCloud phones are a useful analogy. Before there were cloud backups for your phone, if you fell into a pool, all of your data on the phone was probably lost. So I guess don’t throw your SQLite-powered application server into a pool…

Installing SQLite

The easiest way to install sqlite – and the way you’ll do it on the vast vast majority of systems – is via apt:

#sudo apt update && sudo apt upgrade
sudo apt install sqlite

Then verify the installation via:

sqlite3 --version

Creating a Database

There are a few ways to create a database. The most straightforward is as follows:

sqlite3 demo.db

This will open demo.db if it exists or create it if it doesn’t.

But you can also open the sqlite application from the shell, then create a database from there:

sqlite3
.open demo.db

*Note that just running the sqlite3 command without any arguments will connect to an in-memory database. So we’re first opening an in-memory database, then saying “actually I want to connect to demo.db instead”

Another way is to use a language with an SQLite3 library to do that, e.g. Python:

import sqlite3

con = sqlite3.connect('demo.db')

Create and Manage Tables

You create tables just like you would with any other SQL database (e.g. Postgres). That is, it uses the same SQL statements. So to create a table, we could run

sqlite3

then run the following code in the terminal:

CREATE TABLE users (
    id INTEGER PRIMARY KEY AUTOINCREMENT
    username TEXT UNIQUE,
    email TEXT UNIQUE
);

A better approach, though, is to save the above to a .sql file, then run this file from the shell, so you don’t have to type a long SQL statement into the terminal. If that statement is saved in my_file.sql, then we could run:

sqlite3 demo.db < path/to/my_file.sql

Check Existing Tables

To see what tables exist in an sqlite database, run:

sqlite3 demo.db
.tables

Check the Schema of a Table

To see the schema of a table:

sqlite3 demo.db
.schema tablename