@farjs/better-sqlite3-migrate

Automates Sqlite DB schema versioning.

Uses wrapper around better-sqlite3 and bun:sqlite to allow cross- runtime/engine usage.

How to use it

1. Add SQL Migrations Scripts

Add your SQL migrations scripts under dedicated project sub-folder, for ex. ./migrations.

See example SQL migrations scripts:

• V001__initial_db_structure.sql
• V002__rename_db_field.sql

2. Generate bundle.json migrations file

To generate singe migrations bundle file you can use sql-bundle script and run it as part of build process.

For example, in your package.json:

1"scripts": {
2  "sql-bundle": "npx sql-bundle ./migrations"
3}

Then run it as follows:

1npm run sql-bundle

bundle.json file will be (re)generated in the same ./migrations folder.

3. Exclude bundle.json file from git

Since it can be easily generated from the input SQL migrations files there is no need to commit this file.

You can exclude it by adding the following line to your .gitignore file:

1migrations/bundle.json

4. Load and run migrations bundle

The final setup is to actually load and apply migrations on your Sqlite DB during app startup:

1import Database from "@farjs/better-sqlite3-wrapper";
2import { readBundle, runBundle } from "@farjs/better-sqlite3-migrate";
3
4// connect to your DB
5const db = new Database(":memory:");
6
7const bundleUrl = new URL("./migrations/bundle.json", import.meta.url);
8
9// load migrations bundle
10const bundle = await readBundle(bundleUrl);
11
12// apply it
13await runBundle(db, bundle);

This will create (or update existing) schema_versions DB table with the all SQL migrations that were applied.

So, next time it will only apply new SQL migrations, that were not applied before (not exist in schema_versions table yet).