About a month ago I ran into a pitfall in my git work flow. I had a topic branch where I was radically changing the schema and needed to flip back to another branch to fix some issue and found I had some new bizarre issues that had me scratching my head until I realized my db had changed from under me. Hmm.. I discussed the issue with some friends on irc and after a short while one of them came back with a quick solution he and his buddies at SevenWire whipped up and pushed out to github.

db_branch, is a simple plugin that will load an alternate database file for the current branch if it exists and some rake tasks to initialize the config file and databases.

Earlier today I worked on making it even better and added some tasks to clone an existing db and purge the branch dbs.

Git it on github, http://github.com/agile/db_branch/tree/master

1
2

git submodule add git://github.com/agile/db_branch.git vendor/plugins/db_branch
rake db:branch

So now my work flow when messing up the db, is something like..

1
2
3
4
5
6
7
8
9
10
11

git co -b messy_branch
rake db:branch
./script generate migration making_a_mess
# look at me, messing up my db, but it's ok because only this branch uses this db
rake db:migrate
git add db/migrate && git  ci "committing the business"

# and while I'm on that branch, I'll be using my awesomeapp_messy_branch db
git co master

#and I'm back to using my old awesomeapp db

nice!

Taking it to eleven..?

I think I’d like to figure out how to tie it into git’s hooks so that it just automatically creates a db branch and clones when a new branch is created, then purges when the branch is deleted. That’d be hot.