Seed Dump is a Rails plugin (compatible with Rails 4 through 8+) that adds a rake task named db:seed:dump.

It allows you to create seed data files from the existing data in your database.

You can also use Seed Dump from the Rails console. See below for usage examples.

Note: if you want to use Seed Dump with Rails 3 or earlier, use version 0.5.3.

Add it to your Gemfile with:

gem 'seed_dump'

Or install it by hand:

$ gem install seed_dump

Dump all data directly to db/seeds.rb:

  $ rake db:seed:dump

Result:

Product.create!([
  { category_id: 1, description: "Long Sleeve Shirt", name: "Long Sleeve Shirt" },
  { category_id: 3, description: "Plain White Tee Shirt", name: "Plain T-Shirt" }
])
User.create!([
  { password: "123456", username: "test_1" },
  { password: "234567", username: "test_2" }
])

Dump only data from the users table and dump a maximum of 1 record:

$ rake db:seed:dump MODELS=User LIMIT=1

Result:

User.create!([
  { password: "123456", username: "test_1" }
])

Append to db/seeds.rb instead of overwriting it:

rake db:seed:dump APPEND=true

Use another output file instead of db/seeds.rb:

rake db:seed:dump FILE=db/seeds/users.rb

Exclude name and age from the dump:

rake db:seed:dump EXCLUDE=name,age

There are more options that can be set— see below for all of them.

Output a dump of all User records:

irb(main):001:0> puts SeedDump.dump(User)
User.create!([
  { password: "123456", username: "test_1" },
  { password: "234567", username: "test_2" }
])

Write the dump to a file:

irb(main):002:0> SeedDump.dump(User, file: 'db/seeds.rb')

Append the dump to a file:

irb(main):003:0> SeedDump.dump(User, file: 'db/seeds.rb', append: true)

Exclude name and age from the dump:

irb(main):004:0> SeedDump.dump(User, exclude: [:name, :age])

Options are specified as a Hash for the second argument.

In the console, any relation of ActiveRecord rows can be dumped (not individual objects though):

irb(main):005:0> puts SeedDump.dump(User.where(is_admin: false))
User.create!([
  { password: "123456", username: "test_1", is_admin: false },
  { password: "234567", username: "test_2", is_admin: false }
])

Options are common to both the Rake task and the console, except where noted.

append: If set to true, append the data to the file instead of overwriting it. Default: false.

batch_size: Controls the number of records that are processed and written at a given time. Default: 1000. If you're running out of memory when dumping, try decreasing this. If things are dumping too slow, trying increasing this.

exclude: Attributes to be excluded from the dump. Pass a comma-separated list to the Rake task (e.g., EXCLUDE=name,age) and an array of symbols on the console (e.g., exclude: [:name, :age]). Default: [:id, :created_at, :updated_at, :created_on, :updated_on].

file: Write to the specified output file. The Rake task default is db/seeds.rb. The console returns the dump as a string by default if this option is omitted.

group_sti_by_class: If true, Single Table Inheritance (STI) records are grouped by their actual class (e.g., Dog, Cat) instead of the base class (e.g., Animal). This is necessary when STI subclasses have different enum definitions or other class-specific attributes that would be lost if dumped via the base class. Default: false. Example: rake db:seed:dump GROUP_STI_BY_CLASS=true or SeedDump.dump(Animal, group_sti_by_class: true). See the STI Handling section below for more details.

header: If true, adds a comment header to the output file showing the seed_dump command and options used. If a string, uses that string as the header comment. Default: false. Rake task only. Example: rake db:seed:dump HEADER=true or HEADER="Generated by seed_dump".

import: If true, output will be in the format needed by the activerecord-import gem, rather than the default format. You can also pass a Hash of options which will be passed through to the import call (e.g., IMPORT='{ "validate": false }' for Rake, or import: { validate: false } for console). Default: false.

include_all: If set to true, include all columns in the dump (including id, created_at, and updated_at). Equivalent to EXCLUDE="". Default: false. Rake task only. Example: rake db:seed:dump INCLUDE_ALL=true

insert_all: If true, output will use Rails 6+ insert_all for faster bulk inserts that bypass validations and callbacks. Default: false. Example: rake db:seed:dump INSERT_ALL=true or SeedDump.dump(User, insert_all: true).

limit: Dump no more than this amount of data per model. Default: no limit. Rake task only. In the console, just pass in an ActiveRecord::Relation with the appropriate limit (e.g., SeedDump.dump(User.limit(5))).

model_limits: Set different limits for specific models. Format: Model1:limit1,Model2:limit2. Use 0 to mean "no limit" for a specific model. This is useful when LIMIT would break foreign key relationships. Rake task only. Example: rake db:seed:dump LIMIT=10 MODEL_LIMITS="Teacher:0,Student:50" dumps all Teachers, 50 Students, and 10 of everything else.

model[s]: Restrict the dump to the specified comma-separated list of models. Default: all models that have data. If you are using a Rails engine you can dump a specific model by passing "EngineName::ModelName". Rake task only. Example: rake db:seed:dump MODELS="User, Position, Function"

models_exclude: Exclude the specified comma-separated list of models from the dump. Default: no models excluded. Rake task only. Example: rake db:seed:dump MODELS_EXCLUDE="User"

upsert_all: If true, output will use Rails 6+ upsert_all which preserves record IDs and handles conflicts by updating existing records. This is useful when you need to maintain foreign key relationships or want idempotent seed files. Automatically includes id in the dump. Default: false. Example: rake db:seed:dump UPSERT_ALL=true or SeedDump.dump(User, upsert_all: true).

Foreign Key Ordering: Models are automatically dumped in dependency order based on foreign key relationships. This ensures that parent records are created before child records that reference them.

STI Handling: Single Table Inheritance (STI) models are automatically deduplicated. By default, only the base class is dumped to avoid duplicate records (e.g., Animal.create! for both Dog and Cat records). However, if your STI subclasses have different enum definitions or other class-specific attributes, use the group_sti_by_class: true option to dump each subclass separately (e.g., Dog.create! and Cat.create!). This ensures that subclass-specific type casting and validations are properly applied when the seed data is loaded.

HABTM Handling: Has-and-belongs-to-many join tables are automatically detected and dumped without duplication.

If you're using ActiveRecord outside of Rails (e.g., with standalone-migrations), you can set up a custom Rake task:

# In your Rakefile
$LOAD_PATH.unshift(File.expand_path('app/models', __dir__))
Dir.glob(File.expand_path('app/models/*.rb', __dir__)).sort.each(&method(:require))
require 'seed_dump'

namespace :db do
  namespace :seed do
    desc "Dump records from the database into db/seeds.rb"
    task :dump => :environment do
      SeedDump.dump_using_environment(ENV)
    end
  end
end

This loads your models and creates a db:seed:dump task that works like the Rails version.