3

I have been using Michael Bayer's (@zzzeek) Alembic Migrations library to handle my application postgres database needs for over a year and really like it. It's a great tool and has been a great asset.

I have a team of several developers and we create a lot of branches in our Git repository. Therefore, we end up with a lot of merge scripts as we merge our Develop branch into our Feature branches and sometimes when we merge our Feature branches back. This sometimes results in merge scripts that reference merge scripts.

As this process goes on, the process of running through a migration scripts gets worse and worse. We run our integration tests through CodeShip, which runs the Alembic update before each integration test and can see the performance degrade as merge scripts start to build up.

Does anyone know how I can solve this problem? If there was a way to flatten my merge scripts or some other means to start at new base version?

Keep in mind that scripting the entire database and starting the scripts from there is a working options, something I've done, but is very error prone with other branches.

Corey Coogan
  • 1,569
  • 2
  • 17
  • 31
  • It probably depends a lot on the scale of modifications made to the SQLAlchemy models in your other branches, but you might be able to drop the alembic table from your database, delete all of the migration scripts, restamp the database and generate a base migration again. Other git branches would have to rebase off the branch in which you did this, though, before making any new migration scripts. – Aaron D Apr 10 '16 at 01:56

0 Answers0