snix/playbooks/sqlite3.md

SQLite3

Creating a reference for SQLite that I can access when I'm offline (e.g. traveling in an airplane).

Benefits

I enjoy using SQLite because it's lightweight and simple. Instead of networking microservices, I can oftentimes just create a simple db.sqlite3 file and get significant mileage without much tooling overhead.

Limitations

SQLite has some limitations; here are some of the limitations that I have encountered.

• SQLite disables support for FOREIGN KEY by default. Enable it with:

sqlite> PRAGMA foreign_keys = ON;

• SQLite has no BOOLEAN type; it uses 0 and 1 instead.

sqlite> SELECT TRUE;
TRUE
----------
1
sqlite> SELECT FALSE;
FALSE
----------
0

• SQLite has no DATETIME type; it uses TEXT instead.

sqlite> SELECT datetime('now');
datetime('now')
-------------------
2020-07-26 09:52:32

Reference

The following should serve as a useful reference for working with SQLite.

Schema

CREATE TABLE IF NOT EXISTS Movies (
  title TEXT NOT NULL,
  year INTEGER,
  PRIMARY KEY (title)
);

ALTER TABLE Movies ADD COLUMN rating DEFAULT 0.0;

DROP TABLE Movies;

Queries

The following queries should come in handy as a reference:

sqlite> -- I'm using an intentionally incorrect date here for the subsequent UPDATE.
sqlite> INSERT INTO Movies (title, year) VALUES ('Toy Story 3', 2100);
sqlite> SELECT * FROM Movies WHERE year IS NULL;
sqlite> UPDATE Movies SET year = 2010 WHERE title = 'Toy Story 3';
sqlite> -- % is like .* in a regex
sqlite> DELETE FROM Movies WHERE title LIKE 'Toy Story%';

Command Line

• Create a ~/.sqliterc file with the following contents:

.mode column
.headers on

• To start an interactive session:

$ sqlite3 db.sqlite3

• To create a SQLite database from a .sql file:

$ sqlite3 db.sqlite3 <db.sql

• To reload changes to a .sql file while in an interactive session:

sqlite> .read db.sql

Miscellaneous

• For a web-browser-based SQLite viewer, run the following:

$ sqlite_web db.sqlite3

• To import a CSV:

sqlite> .mode csv <table-name>
sqlite> .import path/to/file.csv <table-name>