micrate

Micrate is a database migration tool written in Crystal.

It is inspired by goose. Some code was ported from there too, so check it out.

Micrate currently supports migrations for Postgres, Mysql and SQLite3, but it should be easy to add support for any other database engine with an existing crystal-db API driver.

Command line

To install the standalone binary tool check out the releases page, or use homebrew:

$ brew tap amberframework/micrate
$ brew install micrate

Execute micrate help for usage instructions. Micrate will connect to the database specified by the DATABASE_URL environment variable.

To create a new migration use the scaffold subcommand. For example, micrate scaffold add_users_table will create a new SQL migration file with a name such as db/migrations/20160524162446_add_users_table.sql that looks like this:

-- +micrate Up
-- SQL in section 'Up' is executed when this migration is applied


-- +micrate Down
-- SQL section 'Down' is executed when this migration is rolled back

Comments that start with +micrate are interpreted by micrate when running your migrations. In this case, the Up and Down directives are used to indicate which SQL statements must be run when applying or reverting a migration. You can now go along and write your migration like this:

-- +micrate Up
CREATE TABLE users(id INT PRIMARY KEY, email VARCHAR NOT NULL);

-- +micrate Down
DROP TABLE users;

Now run it using micrate up. This command will execute all pending migrations:

$ micrate up
Migrating db, current version: 0, target: 20160524162947
OK   20160524162446_add_users_table.sql

$ micrate dbversion # at any time you can find out the current version of the database
20160524162446

If you ever need to roll back the last migration, you can do so by executing micrate down. There's also micrate redo which rolls back the last migration and applies it again. Last but not least: use micrate status to find out the state of each migration:

$ micrate status
Applied At                  Migration
=======================================
2016-05-24 16:31:07 UTC  -- 20160524162446_add_users_table.sql
Pending                  -- 20160524163425_add_address_to_users.sql

If using complex statements that might contain semicolons, you must give micrate a hint on how to split the script into separate statements. You can do this with StatementBegin and StatementEnd directives: (thanks goose for this!)

-- +micrate Up
-- +micrate StatementBegin
CREATE OR REPLACE FUNCTION histories_partition_creation( DATE, DATE )
returns void AS $$
DECLARE
  create_query text;
BEGIN
  FOR create_query IN SELECT
      'CREATE TABLE IF NOT EXISTS histories_'
      || TO_CHAR( d, 'YYYY_MM' )
      || ' ( CHECK( created_at >= timestamp '''
      || TO_CHAR( d, 'YYYY-MM-DD 00:00:00' )
      || ''' AND created_at < timestamp '''
      || TO_CHAR( d + INTERVAL '1 month', 'YYYY-MM-DD 00:00:00' )
      || ''' ) ) inherits ( histories );'
    FROM generate_series( $1, $2, '1 month' ) AS d
  LOOP
    EXECUTE create_query;
  END LOOP;  -- LOOP END
END;         -- FUNCTION END
$$
language plpgsql;
-- +micrate StatementEnd

API

To use the Crystal API, add this to your application's shard.yml:

dependencies:
  micrate:
    github: amberframework/micrate

This allows you to programatically use micrate's features. You'll see the Micrate module has an equivalent for every CLI command. If you need to use micrate's CLI without installing the tool (which could be convenient in a CI environment), you can write a runner script as follows:

#! /usr/bin/env crystal
#
# To build a standalone command line client, require the
# driver you wish to use and use `Micrate::Cli`.
#

require "micrate"
require "pg"

Micrate::DB.connection_url = "postgresql://..."
Micrate::Cli.run