init

README

@@ -0,0 +1,33 @@
+= YamlDb
+
+YamlDb is a database-independent format for dumping and restoring data.  It complements the the database-independent schema format found in db/schema.rb.  The data is saved into db/data.yml.
+
+This can be used as a replacement for mysqldump or pg_dump, but only for the databases typically used by Rails apps.  Users, permissions, schemas, triggers, and other advanced database features are not supported - by design.
+
+Any database that has an ActiveRecord adapter should work.
+
+== Usage
+
+rake db:data:dump   ->   Dump contents of Rails database to db/data.yml
+rake db:data:load   ->   Load contents of db/data.yml into the database
+
+Further, there are tasks db:dump and db:load which do the entire database (the equivalent of running db:schema:dump followed by db:data:load).
+
+== Examples
+
+One common use would be to switch your data from one database backend to another.  For example, let's say you wanted to switch from SQLite to MySQL.  You might execute the following steps:
+
+1. rake db:dump
+
+2. Edit config/database.yml and change your adapter to mysql, set up database params
+
+3. mysqladmin create [database name]
+
+4. rake db:load
+
+== Credits
+
+Created by Orion Henry and Adam Wiggins.  Major updates by Ricardo Chimal, Jr.  Patches contributed by Michael Irwin.
+
+Send questions, feedback, or patches to the Heroku mailing list: http://groups.google.com/group/heroku
+

Rakefile

@@ -0,0 +1,10 @@
+require 'rake'
+require 'spec/rake/spectask'
+
+desc "Run all specs"
+Spec::Rake::SpecTask.new('spec') do |t|
+	t.spec_files = FileList['spec/*_spec.rb']
+end
+
+task :default => :spec
+

about.yml

@@ -0,0 +1,5 @@
+author: Orion Henry and Adam Wiggins of Heroku
+summary: Dumps and loads a database-independent data dump format in db/data.yml.
+homepage: http://opensource.heroku.com/
+license: MIT
+rails_version: 1.2+

init.rb

@@ -0,0 +1 @@
+require 'yaml_db'

lib/yaml_db.rb

@@ -0,0 +1,168 @@
+require 'rubygems'
+require 'yaml'
+require 'active_record'
+
+
+module YamlDb
+	def self.dump(filename)
+		disable_logger
+		YamlDb::Dump.dump(File.new(filename, "w"))
+		reenable_logger
+	end
+
+	def self.load(filename)
+		disable_logger
+		YamlDb::Load.load(File.new(filename, "r"))
+		reenable_logger
+	end
+
+	def self.disable_logger
+		@@old_logger = ActiveRecord::Base.logger
+		ActiveRecord::Base.logger = nil
+	end
+
+	def self.reenable_logger
+		ActiveRecord::Base.logger = @@old_logger
+	end
+end
+
+
+module YamlDb::Utils
+	def self.chunk_records(records)
+		yaml = [ records ].to_yaml
+		yaml.sub!("--- \n", "")
+		yaml.sub!('- - -', '  - -')
+		yaml
+	end
+
+	def self.unhash(hash, keys)
+		keys.map { |key| hash[key] }
+	end
+
+	def self.unhash_records(records, keys)
+		records.each_with_index do |record, index|
+			records[index] = unhash(record, keys)	
+		end
+
+		records
+	end
+
+	def self.convert_booleans(records, columns)
+		records.each do |record|
+			columns.each do |column|
+				next if is_boolean(record[column])
+				record[column] = (record[column] == 't' or record[column] == '1')
+			end
+		end
+		records
+	end
+
+	def self.boolean_columns(table)
+		columns = ActiveRecord::Base.connection.columns(table).reject { |c| c.type != :boolean }
+		columns.map { |c| c.name }
+	end
+
+	def self.is_boolean(value)
+		value.kind_of?(TrueClass) or value.kind_of?(FalseClass)
+	end
+end
+
+
+module YamlDb::Dump
+	def self.dump(io)
+		ActiveRecord::Base.connection.tables.each do |table|
+			dump_table(io, table)
+		end
+	end
+
+	def self.dump_table(io, table)
+		return if table_record_count(table).zero?
+
+		dump_table_columns(io, table)
+		dump_table_records(io, table)
+	end
+
+	def self.dump_table_columns(io, table)
+		io.write("\n")
+		io.write({ table => { 'columns' => table_column_names(table) } }.to_yaml)
+	end
+
+	def self.dump_table_records(io, table)
+		table_record_header(io)	
+
+		column_names = table_column_names(table)
+
+		each_table_page(table) do |records|
+			rows = YamlDb::Utils.unhash_records(records, column_names)
+			io.write(YamlDb::Utils.chunk_records(records))
+		end
+	end
+
+	def self.table_record_header(io)
+		io.write("  records: \n")
+	end
+
+	def self.table_column_names(table)
+		ActiveRecord::Base.connection.columns(table).map { |c| c.name }
+	end
+
+	def self.each_table_page(table, records_per_page=1000)
+		total_count = table_record_count(table)
+		pages = (total_count.to_f / records_per_page).ceil - 1
+		id = table_column_names(table).first
+		boolean_columns = YamlDb::Utils.boolean_columns(table)
+
+		(0..pages).to_a.each do |page|
+			sql_limit = "LIMIT #{records_per_page} OFFSET #{records_per_page*page}"
+			records = ActiveRecord::Base.connection.select_all("SELECT * FROM #{table} ORDER BY #{id} #{sql_limit}")
+			records = YamlDb::Utils.convert_booleans(records, boolean_columns)
+			yield records
+		end
+	end
+
+	def self.table_record_count(table)
+		ActiveRecord::Base.connection.select_one("SELECT COUNT(*) FROM #{table}").values.first.to_i
+	end
+end
+
+
+module YamlDb::Load
+	def self.load(io)
+		ActiveRecord::Base.connection.transaction do
+			YAML.load_documents(io) do |ydoc|
+				ydoc.keys.each do |table_name|
+					next if ydoc[table_name].nil?
+					load_table(table_name, ydoc[table_name])
+				end
+			end
+		end
+	end
+
+	def self.truncate_table(table)
+		begin
+			ActiveRecord::Base.connection.execute("TRUNCATE #{table}")
+		rescue Exception
+			ActiveRecord::Base.connection.execute("DELETE FROM #{table}")
+		end
+	end
+
+	def self.load_table(table, data)
+		column_names = data['columns']
+		truncate_table(table)
+		load_records(table, column_names, data['records'])
+		reset_pk_sequence!(table)
+	end
+
+	def self.load_records(table, column_names, records)
+		quoted_column_names = column_names.map { |column| ActiveRecord::Base.connection.quote_column_name(column) }.join(',')
+		records.each do |record|
+			ActiveRecord::Base.connection.execute("INSERT INTO #{table} (#{quoted_column_names}) VALUES (#{record.map { |r| ActiveRecord::Base.connection.quote(r) }.join(',')})")
+		end
+	end
+
+	def self.reset_pk_sequence!(table_name)
+		if ActiveRecord::Base.connection.kind_of?(ActiveRecord::ConnectionAdapters::PostgreSQLAdapter)
+			ActiveRecord::Base.connection.reset_pk_sequence!(table_name)
+		end
+	end
+end

spec/base.rb

@@ -0,0 +1,7 @@
+require 'rubygems'
+require 'spec'
+
+$LOAD_PATH.unshift(File.dirname(__FILE__) + '/../lib')
+require 'yaml_db'
+
+

spec/yaml_dump_spec.rb

@@ -0,0 +1,89 @@
+require File.dirname(__FILE__) + '/base'
+
+describe YamlDb::Dump do
+	before do
+		File.stub!(:new).with('dump.yml', 'w').and_return(StringIO.new)
+
+		ActiveRecord::Base = mock('ActiveRecord::Base', :null_object => true)
+		ActiveRecord::Base.connection = mock('connection')
+		ActiveRecord::Base.connection.stub!(:tables).and_return([ 'mytable' ])
+		ActiveRecord::Base.connection.stub!(:columns).with('mytable').and_return([ mock('a',:name => 'a'), mock('b', :name => 'b') ])
+		ActiveRecord::Base.connection.stub!(:select_one).and_return({"count"=>"2"})
+		ActiveRecord::Base.connection.stub!(:select_all).and_return([ { 'a' => 1, 'b' => 2 }, { 'a' => 3, 'b' => 4 } ])
+	end
+
+	before(:each) do
+		@io = StringIO.new
+	end
+
+	it "should return a formatted string" do
+		YamlDb::Dump.table_record_header(@io)
+		@io.rewind
+		@io.read.should == "  records: \n"
+	end
+
+	it "should return a list of column names" do
+		YamlDb::Dump.table_column_names('mytable').should == [ 'a', 'b' ]
+	end
+
+	it "should return the total number of records in a table" do
+		YamlDb::Dump.table_record_count('mytable').should == 2
+	end
+
+	it "should return a yaml string that contains a table header and column names" do
+		YamlDb::Dump.stub!(:table_column_names).with('mytable').and_return([ 'a', 'b' ])
+		YamlDb::Dump.dump_table_columns(@io, 'mytable')
+		@io.rewind
+		@io.read.should == <<EOYAML
+
+--- 
+mytable: 
+  columns: 
+  - a
+  - b
+EOYAML
+	end
+
+	it "should return all records from the database and return them when there is only 1 page" do
+		YamlDb::Dump.each_table_page('mytable') do |records|
+			records.should == [ { 'a' => 1, 'b' => 2 }, { 'a' => 3, 'b' => 4 } ]
+		end
+	end
+
+	it "should paginate records from the database and return them" do
+		ActiveRecord::Base.connection.stub!(:select_all).and_return([ { 'a' => 1, 'b' => 2 } ], [ { 'a' => 3, 'b' => 4 } ])
+
+		records = [ ]
+		YamlDb::Dump.each_table_page('mytable', 1) do |page|
+			page.size.should == 1
+			records.concat(page)
+		end
+
+		records.should == [ { 'a' => 1, 'b' => 2 }, { 'a' => 3, 'b' => 4 } ]
+	end
+
+	it "should return dump the records for a table in yaml to a given io stream" do
+		YamlDb::Dump.dump_table_records(@io, 'mytable')
+		@io.rewind
+		@io.read.should == <<EOYAML
+  records: 
+  - - 1
+    - 2
+  - - 3
+    - 4
+EOYAML
+	end
+
+	it "should dump a table's contents to yaml" do
+		YamlDb::Dump.should_receive(:dump_table_columns)
+		YamlDb::Dump.should_receive(:dump_table_records)
+		YamlDb::Dump.dump_table(@io, 'mytable')
+	end
+
+	it "should not dump a table's contents when the record count is zero" do
+		YamlDb::Dump.stub!(:table_record_count).with('mytable').and_return(0)
+		YamlDb::Dump.should_not_receive(:dump_table_columns)
+		YamlDb::Dump.should_not_receive(:dump_table_records)
+		YamlDb::Dump.dump_table(@io, 'mytable')
+	end
+end