migrate4j

Run transactional Neo4j data migrations written with Typescript.

Note: this is (obviously) under active development. Use at your own risk.

neo4j-migrations/202008122203_add-stuff-to-things.ts

import { Migration } from 'migrate4j';

export default class AddStuffToThings extends Migration {
  static description = "adds stuff to things"

  up(session) {
    session.run(`
      MATCH (n:Things)
      SET n.stuff = 'hi'
    `)
  }

  down(session) {
    session.run(`
      MATCH(n:Things)
      REMOVE n.stuff
    `)
  }
}

Migrations are run in named order. All migrations in the migrations folder will be run once.

To create a new migration

$ migrate4j create name-of-the-migration

this will create a file matching the pattern current-date-timestamp_name-of-the-migration.ts for example: 202008122203_add-stuff-to-things.ts

What happens when a migration file with a date in the past is added?

For example, you have migrations 0.ts, 2.ts, 3.ts and migrate. Then you add 1.ts and 4.ts and migrate once more.

The migrations that have yet to run will be run in order of their filenames.

This results in migrations running in the following order:

0.ts -> 2.ts -> 3.ts -> 1.ts -> 4.ts

If all of the migrations were reverted and re-ran, they would run in this order:

0.ts -> 1.ts -> 2.ts -> 3.ts -> 4.ts

Keep this in mind when order of migrations is important!

Migrate4j creates a graph representing the migration chain, so you can easily tell the order in which migrations have run.

(:Migration)->[:MIGRATED_TO]->(:Migration)