Empower PostgreSQL migrations in Rails app

Sponsored by Evil Martians

Gem Version Build Status

PGTrunk adds methods to ActiveRecord::Migration to create and manage various PostgreSQL objects (like views, functions, triggers, statistics, types etc.) in Rails.

This gem is greatly influenced by the Scenic, F(x) and ActiveRecord::PostgtresEnum projects but takes some steps further.

In addition to support of different objects, we are solving a problem of interdependency between them. For example, you can create a table, then a function using its type as an argument, then check constraint and index using the function:

create_table "users" do |t|
  t.text "first_name"
  t.text "last_name"

# depends on the `users` table
create_function "full_name(u users) text" do |f|
  f.volatility :immutable
  f.strict true
  f.parallel :safe
  f.body <<~SQL.strip
      SELECT COALESCE(u.first_name, '') + '.' + COALESCE(u.second_name, ''),

# both objects below depend on the `users` and `full_name(users)`
# so they couldn't be placed inside the `create_table` definition in the schema.

create_index "users", "full_name(users.*)", unique: true

# users.full_name is the PostgreSQL alternative syntax for the `full_name(users.*)`
create_check_constraint "users", "length(users.full_name) > 0", name: "full_name_present"

Notice, that we had to separate definitions of indexes and check constraints from tables, because there can be other objects (like functions or types) squeezing between them.

Another difference from aforementioned gems is that we explicitly register all objects created by migrations in the special table (pg_trunk). This let us distinct objects created by "regular" migration from temporary ones added manually and exclude the latter from the schema. We bind any object to a particular version of migration which added it. That's how only those objects that belong to the current branch are dumped into the schema.rb.

As of today we support creation, modification and dropping the following objects:

  • tables
  • indexes
  • foreign keys (including multi-column ones)
  • check constraints
  • views
  • materialized views
  • functions
  • procedures
  • triggers
  • custom statistics
  • enumerable types
  • composite types
  • domains types
  • rules
  • sequences

For tables and indexes we reuse the ActiveRecord's native methods. For check constraints and foreign keys we support both the native definitions inside the table and standalone methods (like create_foreign_key) with additional features. The other methods are implemented from scratch.

In the future other objects like aggregate functions, range types, operators, collations, and more will be supported.

From now and on we support all versions of PostgreSQL since v10.

The gem is targeted to support PostgreSQL-specific features, that's why we won't provide adapters to other databases like Scenic does.


The gem provides a lot of additional methods to create, rename, change a drop various objects. You can find the necessary details here.


Add this line to your application's Gemfile:

gem 'pg_trunk'

And then execute:

$ bundle install

Or install it yourself as:

$ gem install pg_trunk

Add the line somewhere in your ruby code:

require "pg_trunk"


After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and the created tag, and push the .gem file to


Bug reports and pull requests are welcome on GitHub at


The gem is available as open source under the terms of the MIT License.