Question

I am going to migrate from rails 3 app, that used AR and arel to datamapper. I love chains of scopes like Person.where(…).where(…).somescope.paginate(…).order(…). how to migrate from this arel approach to datamapper.

Answer 1

Named scopes in DataMapper are simply class methods that you define on your model classes. Inside those class methods, you typically call #all with some conditions to get back a DataMapper::Collection. In order for your class methods to be chainable, you must make sure to return a DataMapper::Collection.

For completeness’ sake, here are the installation instructions …

gem install dm-core
gem install dm-migrations
gem install dm-sqlite-adapter
gem install dm-chunked_query

And the code to get you going (put this into test.rb for maximum reproducability)

require 'rubygems'
require 'dm-core'
require 'dm-migrations'
require 'dm-chunked_query'

class Person
  include DataMapper::Resource

  property :id,        Serial
  property :name,      String
  property :hobby,     String
  property :country,   String
  property :continent, String

  def self.european
    all(:continent => 'Europe')
  end

  def self.hackers
    all(:hobby => 'Hacking')
  end

  def self.named(name)
    all(:name => name)
  end
end

DataMapper::Logger.new($stdout, :debug)
DataMapper.setup(:default, 'sqlite::memory:')

DataMapper.finalize.auto_migrate!

1.upto(10) do |i|
  Person.create(:name => "Alex", :hobby => 'Hacking', :country => "Country-#{i}", :continent => 'Europe')
end

# you could even skip the explicit call to #all
# i just left it in there because it reads nicely
Person.all.european.hackers.named('Alex').chunks(5).each_with_index do |chunk, idx|
  puts "Rendering page #{idx + 1}"
  chunk.each do |person|
    puts "Someone named #{person.name} who lives in #{person.country}"
  end
end

__END__

ruby-1.9.2-p180@datamapper mungo:dm-rails snusnu$ ruby test.rb
 ~ (0.000102) SELECT sqlite_version(*)
 ~ (0.000132) DROP TABLE IF EXISTS "people"
 ~ (0.000013) PRAGMA table_info("people")
 ~ (0.000315) CREATE TABLE "people" ("id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT, "name" VARCHAR(50), "hobby" VARCHAR(50), "country" VARCHAR(50), "continent" VARCHAR(50))
 ~ (0.000049) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-1', 'Europe')
 ~ (0.000056) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-2', 'Europe')
 ~ (0.000044) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-3', 'Europe')
 ~ (0.000043) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-4', 'Europe')
 ~ (0.000037) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-5', 'Europe')
 ~ (0.000038) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-6', 'Europe')
 ~ (0.000035) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-7', 'Europe')
 ~ (0.000035) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-8', 'Europe')
 ~ (0.000036) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-9', 'Europe')
 ~ (0.000039) INSERT INTO "people" ("name", "hobby", "country", "continent") VALUES ('Alex', 'Hacking', 'Country-10', 'Europe')
 ~ (0.000069) SELECT "id", "name", "hobby", "country", "continent" FROM "people" WHERE ("continent" = 'Europe' AND "hobby" = 'Hacking' AND "name" = 'Alex') ORDER BY "id"
Rendering page 1
Someone named Alex who lives in Country-1
Someone named Alex who lives in Country-2
Someone named Alex who lives in Country-3
Someone named Alex who lives in Country-4
Someone named Alex who lives in Country-5
Rendering page 2
Someone named Alex who lives in Country-6
Someone named Alex who lives in Country-7
Someone named Alex who lives in Country-8
Someone named Alex who lives in Country-9
Someone named Alex who lives in Country-10

See https://github.com/postmodern/dm-chunked_query for more information on the lowlevel approach to pagination (batch processing) that provides the #chunks method used in this example.