Add `:charset` and `:collation` options support for MySQL string and text columns

Example:

    create_table :foos do |t|
      t.string :string_utf8_bin, charset: 'utf8', collation: 'utf8_bin'
      t.text   :text_ascii,      charset: 'ascii'
    end

activerecord/CHANGELOG.md

+*   Add `:charset` and `:collation` options support for MySQL string and text columns.
+
+    Example:
+
+        create_table :foos do |t|
+          t.string :string_utf8_bin, charset: 'utf8', collation: 'utf8_bin'
+          t.text   :text_ascii,      charset: 'ascii'
+        end
+
+    *Ryuta Kamizono*
+
 *   Correctly dump `serial` and `bigserial`.
 
     *Ryuta Kamizono*

activerecord/lib/active_record/connection_adapters/abstract_mysql_adapter.rb

@@ -13,6 +13,10 @@ def primary_key(name, type = :primary_key, **options)
         end
       end
 
+      class ColumnDefinition < ActiveRecord::ConnectionAdapters::ColumnDefinition
+        attr_accessor :charset, :collation
+      end
+
       class TableDefinition < ActiveRecord::ConnectionAdapters::TableDefinition
         include ColumnMethods
 
@@ -23,8 +27,16 @@ def new_column_definition(name, type, options) # :nodoc:
             column.type = :integer
             column.auto_increment = true
           end
+          column.charset = options[:charset]
+          column.collation = options[:collation]
           column
         end
+
+        private
+
+        def create_column_definition(name, type)
+          ColumnDefinition.new(name, type)
+        end
       end
 
       class Table < ActiveRecord::ConnectionAdapters::Table
@@ -60,6 +72,23 @@ def visit_ChangeColumnDefinition(o)
           add_column_position!(change_column_sql, column_options(o.column))
         end
 
+        def column_options(o)
+          column_options = super
+          column_options[:charset] = o.charset
+          column_options[:collation] = o.collation
+          column_options
+        end
+
+        def add_column_options!(sql, options)
+          if options[:charset]
+            sql << " CHARACTER SET #{options[:charset]}"
+          end
+          if options[:collation]
+            sql << " COLLATE #{options[:collation]}"
+          end
+          super
+        end
+
         def add_column_position!(sql, options)
           if options[:first]
             sql << " FIRST"
@@ -99,9 +128,18 @@ def prepare_column_options(column)
         spec = super
         spec.delete(:precision) if /time/ === column.sql_type && column.precision == 0
         spec.delete(:limit)     if :boolean === column.type
+        if column.collation && table_name = column.instance_variable_get(:@table_name)
+          @collation_cache ||= {}
+          @collation_cache[table_name] ||= select_one("SHOW TABLE STATUS LIKE '#{table_name}'")["Collation"]
+          spec[:collation] = column.collation.inspect if column.collation != @collation_cache[table_name]
+        end
         spec
       end
 
+      def migration_keys
+        super + [:collation]
+      end
+
       class Column < ConnectionAdapters::Column # :nodoc:
         delegate :strict, :collation, :extra, to: :sql_type_metadata, allow_nil: true
 

activerecord/lib/active_record/connection_adapters/column.rb

@@ -28,6 +28,7 @@ def initialize(name, default, sql_type_metadata = nil, null = true, default_func
         @null = null
         @default = default
         @default_function = default_function
+        @table_name = nil
       end
 
       def has_default?

activerecord/lib/active_record/schema_dumper.rb

@@ -105,7 +105,10 @@ def tables(stream)
       end
 
       def table(table, stream)
-        columns = @connection.columns(table)
+        columns = @connection.columns(table).map do |column|
+          column.instance_variable_set(:@table_name, table)
+          column
+        end
         begin
           tbl = StringIO.new
 

activerecord/test/cases/adapters/mysql/charset_collation_test.rb

+require "cases/helper"
+require 'support/schema_dumping_helper'
+
+class CharsetCollationTest < ActiveRecord::TestCase
+  include SchemaDumpingHelper
+  self.use_transactional_fixtures = false
+
+  setup do
+    @connection = ActiveRecord::Base.connection
+    @connection.create_table :charset_collations, force: true do |t|
+      t.string :string_ascii_bin, charset: 'ascii', collation: 'ascii_bin'
+      t.text :text_ucs2_unicode_ci, charset: 'ucs2', collation: 'ucs2_unicode_ci'
+    end
+  end
+
+  teardown do
+    @connection.drop_table :charset_collations, if_exists: true
+  end
+
+  test "string column with charset and collation" do
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'string_ascii_bin' }
+    assert_equal :string, column.type
+    assert_equal 'ascii_bin', column.collation
+  end
+
+  test "text column with charset and collation" do
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'text_ucs2_unicode_ci' }
+    assert_equal :text, column.type
+    assert_equal 'ucs2_unicode_ci', column.collation
+  end
+
+  test "add column with charset and collation" do
+    @connection.add_column :charset_collations, :title, :string, charset: 'utf8', collation: 'utf8_bin'
+
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'title' }
+    assert_equal :string, column.type
+    assert_equal 'utf8_bin', column.collation
+  end
+
+  test "change column with charset and collation" do
+    @connection.add_column :charset_collations, :description, :string, charset: 'utf8', collation: 'utf8_unicode_ci'
+    @connection.change_column :charset_collations, :description, :text, charset: 'utf8', collation: 'utf8_general_ci'
+
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'description' }
+    assert_equal :text, column.type
+    assert_equal 'utf8_general_ci', column.collation
+  end
+
+  test "schema dump includes collation" do
+    output = dump_table_schema("charset_collations")
+    assert_match %r{t.string\s+"string_ascii_bin",\s+limit: 255,\s+collation: "ascii_bin"$}, output
+    assert_match %r{t.text\s+"text_ucs2_unicode_ci",\s+limit: 65535,\s+collation: "ucs2_unicode_ci"$}, output
+  end
+end

activerecord/test/cases/adapters/mysql2/charset_collation_test.rb

+require "cases/helper"
+require 'support/schema_dumping_helper'
+
+class CharsetCollationTest < ActiveRecord::TestCase
+  include SchemaDumpingHelper
+  self.use_transactional_fixtures = false
+
+  setup do
+    @connection = ActiveRecord::Base.connection
+    @connection.create_table :charset_collations, force: true do |t|
+      t.string :string_ascii_bin, charset: 'ascii', collation: 'ascii_bin'
+      t.text :text_ucs2_unicode_ci, charset: 'ucs2', collation: 'ucs2_unicode_ci'
+    end
+  end
+
+  teardown do
+    @connection.drop_table :charset_collations, if_exists: true
+  end
+
+  test "string column with charset and collation" do
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'string_ascii_bin' }
+    assert_equal :string, column.type
+    assert_equal 'ascii_bin', column.collation
+  end
+
+  test "text column with charset and collation" do
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'text_ucs2_unicode_ci' }
+    assert_equal :text, column.type
+    assert_equal 'ucs2_unicode_ci', column.collation
+  end
+
+  test "add column with charset and collation" do
+    @connection.add_column :charset_collations, :title, :string, charset: 'utf8', collation: 'utf8_bin'
+
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'title' }
+    assert_equal :string, column.type
+    assert_equal 'utf8_bin', column.collation
+  end
+
+  test "change column with charset and collation" do
+    @connection.add_column :charset_collations, :description, :string, charset: 'utf8', collation: 'utf8_unicode_ci'
+    @connection.change_column :charset_collations, :description, :text, charset: 'utf8', collation: 'utf8_general_ci'
+
+    column = @connection.columns(:charset_collations).find { |c| c.name == 'description' }
+    assert_equal :text, column.type
+    assert_equal 'utf8_general_ci', column.collation
+  end
+
+  test "schema dump includes collation" do
+    output = dump_table_schema("charset_collations")
+    assert_match %r{t.string\s+"string_ascii_bin",\s+limit: 255,\s+collation: "ascii_bin"$}, output
+    assert_match %r{t.text\s+"text_ucs2_unicode_ci",\s+limit: 65535,\s+collation: "ucs2_unicode_ci"$}, output
+  end
+end

activerecord/test/schema/mysql2_specific_schema.rb

@@ -24,6 +24,11 @@
   add_index :key_tests, :pizza, :using => :btree, :name => 'index_key_tests_on_pizza'
   add_index :key_tests, :snacks, :name => 'index_key_tests_on_snack'
 
+  create_table :collation_tests, id: false, force: true do |t|
+    t.string :string_cs_column, limit: 1, collation: 'utf8_bin'
+    t.string :string_ci_column, limit: 1, collation: 'utf8_general_ci'
+  end
+
   ActiveRecord::Base.connection.execute <<-SQL
 DROP PROCEDURE IF EXISTS ten;
 SQL
@@ -33,15 +38,6 @@
 BEGIN
 	select 10;
 END
-SQL
-
-  ActiveRecord::Base.connection.drop_table "collation_tests", if_exists: true
-
-  ActiveRecord::Base.connection.execute <<-SQL
-CREATE TABLE collation_tests (
-  string_cs_column VARCHAR(1) COLLATE utf8_bin,
-  string_ci_column VARCHAR(1) COLLATE utf8_general_ci
-) CHARACTER SET utf8 COLLATE utf8_general_ci
 SQL
 
   ActiveRecord::Base.connection.drop_table "enum_tests", if_exists: true

activerecord/test/schema/mysql_specific_schema.rb

@@ -24,6 +24,11 @@
   add_index :key_tests, :pizza, :using => :btree, :name => 'index_key_tests_on_pizza'
   add_index :key_tests, :snacks, :name => 'index_key_tests_on_snack'
 
+  create_table :collation_tests, id: false, force: true do |t|
+    t.string :string_cs_column, limit: 1, collation: 'utf8_bin'
+    t.string :string_ci_column, limit: 1, collation: 'utf8_general_ci'
+  end
+
   ActiveRecord::Base.connection.execute <<-SQL
 DROP PROCEDURE IF EXISTS ten;
 SQL
@@ -44,15 +49,6 @@
 BEGIN
 	select * from topics limit 1;
 END
-SQL
-
-  ActiveRecord::Base.connection.drop_table "collation_tests", if_exists: true
-
-  ActiveRecord::Base.connection.execute <<-SQL
-CREATE TABLE collation_tests (
-  string_cs_column VARCHAR(1) COLLATE utf8_bin,
-  string_ci_column VARCHAR(1) COLLATE utf8_general_ci
-) CHARACTER SET utf8 COLLATE utf8_general_ci
 SQL
 
   ActiveRecord::Base.connection.drop_table "enum_tests", if_exists: true