emulate schema_migrations after rake db:schema:load

[jdowd]

When you run rake db:schema:load an entry for each db migration is added to the schema_migrations table.

I think it would be useful to have the same behavior wrt the data_migrations table. After all, there should be no need to migrate data right after a schema:load.

Without this behavior the data migrations will need a bunch of conditionals.

Example:
My db has students and enrollments tables.

Initially #call_sign was a column on students, but it was later moved to enrollments and dropped from students. At that point I added a data migration to move the data from students to enrollments.

But if I deploy to a new server that data_migration will fail with no column call_sign on table students

Is it possible for rails-data-migrations to modify the behavior of db:schema:load? Or to listen for it and respond to it?

[jdowd]

Maybe a better idea is to add a new task that adds each pending data migration to the data_migrations table, without actually attempting to execute them.

[jdowd]

something like this (which I added as a custom rake task in my app):

  task :mark_data_migrations_complete => :environment do
    migrations_path = RailsDataMigrations::Migrator.migrations_path
    source_versions = RailsDataMigrations::Migrator.migrations(migrations_path).collect(&:version)
    applied_versions = RailsDataMigrations::Migrator.get_all_versions

    (source_versions - applied_versions).sort.each do |version|
      RailsDataMigrations::LogEntry.create!(version: version.to_s)
    end

  end

[serggl]

@jdowd would you mind wrapping this into a PR? I think reset is a good name for this task

[jdowd]

yup, will do. Apologies in advance cuz it might take me a bit. I assume you'll want specs to match, right?