migsql is a simple, lightweight up/down sql migration manager.
- Store migrations with unix time stamps, to ensure they're applied sequentially, even if multiple people are working on the same project.
- Enable both up and down migrations.
- Store the "state" of the database (ie its current migration) in a migration table on the server itself.
- Support multiple database confgurations per project.
- Be a globally accessible binary executable, installed via gem for ease of use.
- Supports both linux and windows
NOTE: migsql is still under development, in need of a nice 'Green Refactor'.
Still to be done:
- Configuration file to override defaults (such as ./db location)
- Apply migrations in a transactional manner, rolling back all migrations part of that batch on failure
- Some sad-path tests + fixes
Linux: You'll need the freedts package + development libraries, so (depending on your distro):
- yum install freedts
- yum install freedts-devel
I have also pulled out the tiny_tds requirement (because of Cygwin).
As a result - if you plan to apply the migrations from your machine, you'll need to do
gem install tiny_tds
gem install migsql
From there, create an initial config with:
This will create a ./db/config.yml - you need to edit this with your database parameters
To create a migration do:
migsql create-migration <friendly name>
You'll then get up/down scripts for this migration created. Simply stick your SQL in them.
To execute a migration to the most recent one available:
Or to migrate to a specific version (up, or down):
migsql migrate to <friendlyname>
If you want to forcefully apply a particular up/down script (for testing purposes):
migsql apply <friendlyname>
If your config.yml contains multiple databases, you will need to specify which db you're targeting, like so:
migsql create-migration <friendly name> <dbname> migsql migrate <dbname> migsql migrate <dbname> to <friendlyname> migsql apply <friendlyname> to <dbname>
This project has been developed using Test Driven Development, with rspec. Everything is configured to run automatically with Guard.
You'll need to configure spec/spec_helper.rb with the details of a local SQL server, as some of the tests assert on actual sql manipulations.
Further to this, the branching strategy is gitflow (https://github.com/nvie/gitflow), so please ensure you do your work in feature branches first.
- Clone the repo
- bundle install
- Create a feature branch
- Write some tests
- Write some code
- Run your tests
- Finish your feature branch
- Submit a pull request to me
- 1.2.0 Enabled SSMS GO statement support
- 1.1.5 Graceful failure when going up
- 1.1.4 Big bug fixes
- 1.1.3 More informative feedback
- 1.1.2 Enabled connection override by environment variables
- 1.1.0 Moved TDS requirement to be used when required
- 1.0.6 Introduced apply
- 1.0.5 Refactors and code improvements
- 1.0.4 Updates to gemspec
- 1.0.3 Readme updates
- 1.0.2 Small bug fixes
- 1.0.1 Small bug fixes
- 1.0.0 Initial Dev Release
Copyright (c) 2014 Karl Stoney
Licensed under the MIT license.