activerecord 5.2.6 → 6.1.3.2

data/lib/active_record/relation/batches.rb

@@ -37,6 +37,7 @@ module ActiveRecord
     # * <tt>:finish</tt> - Specifies the primary key value to end at, inclusive of the value.
     # * <tt>:error_on_ignore</tt> - Overrides the application config to specify if an error should be raised when
     #   an order is present in the relation.
+    # * <tt>:order</tt> - Specifies the primary key order (can be :asc or :desc). Defaults to :asc.
     #
     # Limits are honored, and if present there is no requirement for the batch
     # size: it can be less than, equal to, or greater than the limit.
@@ -57,22 +58,22 @@ module ActiveRecord
     #     person.party_all_night!
     #   end
     #
-    # NOTE: It's not possible to set the order. That is automatically set to
-    # ascending on the primary key ("id ASC") to make the batch ordering
-    # work. This also means that this method only works when the primary key is
+    # NOTE: Order can be ascending (:asc) or descending (:desc). It is automatically set to
+    # ascending on the primary key ("id ASC").
+    # This also means that this method only works when the primary key is
     # orderable (e.g. an integer or string).
     #
     # NOTE: By its nature, batch processing is subject to race conditions if
     # other processes are modifying the database.
-    def find_each(start: nil, finish: nil, batch_size: 1000, error_on_ignore: nil)
+    def find_each(start: nil, finish: nil, batch_size: 1000, error_on_ignore: nil, order: :asc)
       if block_given?
-        find_in_batches(start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do |records|
+        find_in_batches(start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore, order: order) do |records|
           records.each { |record| yield record }
         end
       else
-        enum_for(:find_each, start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do
+        enum_for(:find_each, start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore, order: order) do
           relation = self
-          apply_limits(relation, start, finish).size
+          apply_limits(relation, start, finish, order).size
         end
       end
     end
@@ -101,6 +102,7 @@ module ActiveRecord
     # * <tt>:finish</tt> - Specifies the primary key value to end at, inclusive of the value.
     # * <tt>:error_on_ignore</tt> - Overrides the application config to specify if an error should be raised when
     #   an order is present in the relation.
+    # * <tt>:order</tt> - Specifies the primary key order (can be :asc or :desc). Defaults to :asc.
     #
     # Limits are honored, and if present there is no requirement for the batch
     # size: it can be less than, equal to, or greater than the limit.
@@ -116,23 +118,23 @@ module ActiveRecord
     #     group.each { |person| person.party_all_night! }
     #   end
     #
-    # NOTE: It's not possible to set the order. That is automatically set to
-    # ascending on the primary key ("id ASC") to make the batch ordering
-    # work. This also means that this method only works when the primary key is
+    # NOTE: Order can be ascending (:asc) or descending (:desc). It is automatically set to
+    # ascending on the primary key ("id ASC").
+    # This also means that this method only works when the primary key is
     # orderable (e.g. an integer or string).
     #
     # NOTE: By its nature, batch processing is subject to race conditions if
     # other processes are modifying the database.
-    def find_in_batches(start: nil, finish: nil, batch_size: 1000, error_on_ignore: nil)
+    def find_in_batches(start: nil, finish: nil, batch_size: 1000, error_on_ignore: nil, order: :asc)
       relation = self
       unless block_given?
-        return to_enum(:find_in_batches, start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do
-          total = apply_limits(relation, start, finish).size
+        return to_enum(:find_in_batches, start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore, order: order) do
+          total = apply_limits(relation, start, finish, order).size
           (total - 1).div(batch_size) + 1
         end
       end
 
-      in_batches(of: batch_size, start: start, finish: finish, load: true, error_on_ignore: error_on_ignore) do |batch|
+      in_batches(of: batch_size, start: start, finish: finish, load: true, error_on_ignore: error_on_ignore, order: order) do |batch|
         yield batch.to_a
       end
     end
@@ -165,6 +167,7 @@ module ActiveRecord
     # * <tt>:finish</tt> - Specifies the primary key value to end at, inclusive of the value.
     # * <tt>:error_on_ignore</tt> - Overrides the application config to specify if an error should be raised when
     #   an order is present in the relation.
+    # * <tt>:order</tt> - Specifies the primary key order (can be :asc or :desc). Defaults to :asc.
     #
     # Limits are honored, and if present there is no requirement for the batch
     # size, it can be less than, equal, or greater than the limit.
@@ -191,19 +194,23 @@ module ActiveRecord
     #
     #   Person.in_batches.each_record(&:party_all_night!)
     #
-    # NOTE: It's not possible to set the order. That is automatically set to
-    # ascending on the primary key ("id ASC") to make the batch ordering
-    # consistent. Therefore the primary key must be orderable, e.g. an integer
-    # or a string.
+    # NOTE: Order can be ascending (:asc) or descending (:desc). It is automatically set to
+    # ascending on the primary key ("id ASC").
+    # This also means that this method only works when the primary key is
+    # orderable (e.g. an integer or string).
     #
     # NOTE: By its nature, batch processing is subject to race conditions if
     # other processes are modifying the database.
-    def in_batches(of: 1000, start: nil, finish: nil, load: false, error_on_ignore: nil)
+    def in_batches(of: 1000, start: nil, finish: nil, load: false, error_on_ignore: nil, order: :asc)
       relation = self
       unless block_given?
         return BatchEnumerator.new(of: of, start: start, finish: finish, relation: self)
       end
 
+      unless [:asc, :desc].include?(order)
+        raise ArgumentError, ":order must be :asc or :desc, got #{order.inspect}"
+      end
+
       if arel.orders.present?
         act_on_ignored_order(error_on_ignore)
       end
@@ -214,8 +221,8 @@ module ActiveRecord
         batch_limit = remaining if remaining < batch_limit
       end
 
-      relation = relation.reorder(batch_order).limit(batch_limit)
-      relation = apply_limits(relation, start, finish)
+      relation = relation.reorder(batch_order(order)).limit(batch_limit)
+      relation = apply_limits(relation, start, finish, order)
       relation.skip_query_cache! # Retaining the results in the query cache would undermine the point of batching
       batch_relation = relation
 
@@ -251,27 +258,29 @@ module ActiveRecord
           end
         end
 
-        attr = Relation::QueryAttribute.new(primary_key, primary_key_offset, klass.type_for_attribute(primary_key))
-        batch_relation = relation.where(arel_attribute(primary_key).gt(Arel::Nodes::BindParam.new(attr)))
+        batch_relation = relation.where(
+          predicate_builder[primary_key, primary_key_offset, order == :desc ? :lt : :gt]
+        )
       end
     end
 
     private
-
-      def apply_limits(relation, start, finish)
-        if start
-          attr = Relation::QueryAttribute.new(primary_key, start, klass.type_for_attribute(primary_key))
-          relation = relation.where(arel_attribute(primary_key).gteq(Arel::Nodes::BindParam.new(attr)))
-        end
-        if finish
-          attr = Relation::QueryAttribute.new(primary_key, finish, klass.type_for_attribute(primary_key))
-          relation = relation.where(arel_attribute(primary_key).lteq(Arel::Nodes::BindParam.new(attr)))
-        end
+      def apply_limits(relation, start, finish, order)
+        relation = apply_start_limit(relation, start, order) if start
+        relation = apply_finish_limit(relation, finish, order) if finish
         relation
       end
 
-      def batch_order
-        arel_attribute(primary_key).asc
+      def apply_start_limit(relation, start, order)
+        relation.where(predicate_builder[primary_key, start, order == :desc ? :lteq : :gteq])
+      end
+
+      def apply_finish_limit(relation, finish, order)
+        relation.where(predicate_builder[primary_key, finish, order == :desc ? :gteq : :lteq])
+      end
+
+      def batch_order(order)
+        table[primary_key].public_send(order)
       end
 
       def act_on_ignored_order(error_on_ignore)

data/lib/active_record/relation/batches/batch_enumerator.rb

@@ -41,19 +41,35 @@ module ActiveRecord
         end
       end
 
-      # Delegates #delete_all, #update_all, #destroy_all methods to each batch.
+      # Deletes records in batches. Returns the total number of rows affected.
       #
-      #   People.in_batches.delete_all
-      #   People.where('age < 10').in_batches.destroy_all
-      #   People.in_batches.update_all('age = age + 1')
-      [:delete_all, :update_all, :destroy_all].each do |method|
-        define_method(method) do |*args, &block|
-          @relation.to_enum(:in_batches, of: @of, start: @start, finish: @finish, load: false).each do |relation|
-            relation.send(method, *args, &block)
-          end
+      #   Person.in_batches.delete_all
+      #
+      # See Relation#delete_all for details of how each batch is deleted.
+      def delete_all
+        sum(&:delete_all)
+      end
+
+      # Updates records in batches. Returns the total number of rows affected.
+      #
+      #   Person.in_batches.update_all("age = age + 1")
+      #
+      # See Relation#update_all for details of how each batch is updated.
+      def update_all(updates)
+        sum do |relation|
+          relation.update_all(updates)
         end
       end
 
+      # Destroys records in batches.
+      #
+      #   Person.where("age < 10").in_batches.destroy_all
+      #
+      # See Relation#destroy_all for details of how each batch is destroyed.
+      def destroy_all
+        each(&:destroy_all)
+      end
+
       # Yields an ActiveRecord::Relation object for each batch of records.
       #
       #   Person.in_batches.each do |relation|

data/lib/active_record/relation/calculations.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require "active_support/core_ext/enumerable"
+
 module ActiveRecord
   module Calculations
     # Count the records.
@@ -41,15 +43,13 @@ module ActiveRecord
     def count(column_name = nil)
       if block_given?
         unless column_name.nil?
-          ActiveSupport::Deprecation.warn \
-            "When `count' is called with a block, it ignores other arguments. " \
-            "This behavior is now deprecated and will result in an ArgumentError in Rails 6.0."
+          raise ArgumentError, "Column name argument is not supported when a block is passed."
         end
 
-        return super()
+        super()
+      else
+        calculate(:count, column_name)
       end
-
-      calculate(:count, column_name)
     end
 
     # Calculates the average value on a given column. Returns +nil+ if there's
@@ -86,15 +86,13 @@ module ActiveRecord
     def sum(column_name = nil)
       if block_given?
         unless column_name.nil?
-          ActiveSupport::Deprecation.warn \
-            "When `sum' is called with a block, it ignores other arguments. " \
-            "This behavior is now deprecated and will result in an ArgumentError in Rails 6.0."
+          raise ArgumentError, "Column name argument is not supported when a block is passed."
         end
 
-        return super()
+        super()
+      else
+        calculate(:sum, column_name)
       end
-
-      calculate(:sum, column_name)
     end
 
     # This calculates aggregate values in the given column. Methods for #count, #sum, #average,
@@ -138,7 +136,7 @@ module ActiveRecord
             relation.select_values = [ klass.primary_key || table[Arel.star] ]
           end
           # PostgreSQL: ORDER BY expressions must appear in SELECT list when using DISTINCT
-          relation.order_values = []
+          relation.order_values = [] if group_values.empty?
         end
 
         relation.calculate(operation, column_name)
@@ -176,14 +174,14 @@ module ActiveRecord
     #   # SELECT people.id FROM people WHERE people.age = 21 LIMIT 5
     #   # => [2, 3]
     #
-    #   Person.pluck('DATEDIFF(updated_at, created_at)')
+    #   Person.pluck(Arel.sql('DATEDIFF(updated_at, created_at)'))
     #   # SELECT DATEDIFF(updated_at, created_at) FROM people
     #   # => ['0', '27761', '173']
     #
     # See also #ids.
     #
     def pluck(*column_names)
-      if loaded? && (column_names.map(&:to_s) - @klass.attribute_names - @klass.attribute_aliases.keys).empty?
+      if loaded? && all_attributes?(column_names)
         return records.pluck(*column_names)
       end
 
@@ -191,12 +189,41 @@ module ActiveRecord
         relation = apply_join_dependency
         relation.pluck(*column_names)
       else
-        klass.enforce_raw_sql_whitelist(column_names)
+        klass.disallow_raw_sql!(column_names)
+        columns = arel_columns(column_names)
         relation = spawn
-        relation.select_values = column_names
-        result = skip_query_cache_if_necessary { klass.connection.select_all(relation.arel, nil) }
-        result.cast_values(klass.attribute_types)
+        relation.select_values = columns
+        result = skip_query_cache_if_necessary do
+          if where_clause.contradiction?
+            ActiveRecord::Result.new([], [])
+          else
+            klass.connection.select_all(relation.arel, nil)
+          end
+        end
+        type_cast_pluck_values(result, columns)
+      end
+    end
+
+    # Pick the value(s) from the named column(s) in the current relation.
+    # This is short-hand for <tt>relation.limit(1).pluck(*column_names).first</tt>, and is primarily useful
+    # when you have a relation that's already narrowed down to a single row.
+    #
+    # Just like #pluck, #pick will only load the actual value, not the entire record object, so it's also
+    # more efficient. The value is, again like with pluck, typecast by the column type.
+    #
+    #   Person.where(id: 1).pick(:name)
+    #   # SELECT people.name FROM people WHERE id = 1 LIMIT 1
+    #   # => 'David'
+    #
+    #   Person.where(id: 1).pick(:name, :email_address)
+    #   # SELECT people.name, people.email_address FROM people WHERE id = 1 LIMIT 1
+    #   # => [ 'David', 'david@loudthinking.com' ]
+    def pick(*column_names)
+      if loaded? && all_attributes?(column_names)
+        return records.pick(*column_names)
       end
+
+      limit(1).pluck(*column_names).first
     end
 
     # Pluck all the ID's for the relation using the table's primary key
@@ -208,6 +235,10 @@ module ActiveRecord
     end
 
     private
+      def all_attributes?(column_names)
+        (column_names.map(&:to_s) - @klass.attribute_names - @klass.attribute_aliases.keys).empty?
+      end
+
       def has_include?(column_name)
         eager_loading? || (includes_values.present? && column_name && column_name != :all)
       end
@@ -246,20 +277,16 @@ module ActiveRecord
       def aggregate_column(column_name)
         return column_name if Arel::Expressions === column_name
 
-        if @klass.has_attribute?(column_name) || @klass.attribute_alias?(column_name)
-          @klass.arel_attribute(column_name)
-        else
-          Arel.sql(column_name == :all ? "*" : column_name.to_s)
+        arel_column(column_name.to_s) do |name|
+          Arel.sql(column_name == :all ? "*" : name)
         end
       end
 
       def operation_over_aggregate_column(column, operation, distinct)
-        operation == "count" ? column.count(distinct) : column.send(operation)
+        operation == "count" ? column.count(distinct) : column.public_send(operation)
       end
 
       def execute_simple_calculation(operation, column_name, distinct) #:nodoc:
-        column_alias = column_name
-
         if operation == "count" && (column_name == :all && distinct || has_limit_or_offset?)
           # Shortcut when limit is zero.
           return 0 if limit_value == 0
@@ -270,56 +297,55 @@ module ActiveRecord
           relation = unscope(:order).distinct!(false)
 
           column = aggregate_column(column_name)
-
           select_value = operation_over_aggregate_column(column, operation, distinct)
-          if operation == "sum" && distinct
-            select_value.distinct = true
-          end
+          select_value.distinct = true if operation == "sum" && distinct
 
-          column_alias = select_value.alias
-          column_alias ||= @klass.connection.column_name_for_operation(operation, select_value)
           relation.select_values = [select_value]
 
           query_builder = relation.arel
         end
 
-        result = skip_query_cache_if_necessary { @klass.connection.select_all(query_builder, nil) }
-        row    = result.first
-        value  = row && row.values.first
-        type   = result.column_types.fetch(column_alias) do
-          type_for(column_name)
-        end
+        result = skip_query_cache_if_necessary { @klass.connection.select_all(query_builder) }
 
-        type_cast_calculated_value(value, type, operation)
+        type_cast_calculated_value(result.cast_values.first, operation) do |value|
+          type = column.try(:type_caster) ||
+            lookup_cast_type_from_join_dependencies(column_name.to_s) || Type.default_value
+          type.deserialize(value)
+        end
       end
 
       def execute_grouped_calculation(operation, column_name, distinct) #:nodoc:
-        group_attrs = group_values
+        group_fields = group_values
+        group_fields = group_fields.uniq if group_fields.size > 1
+
+        unless group_fields == group_values
+          ActiveSupport::Deprecation.warn(<<-MSG.squish)
+            `#{operation}` with group by duplicated fields does no longer affect to result in Rails 6.2.
+            To migrate to Rails 6.2's behavior, use `uniq!(:group)` to deduplicate group fields
+            (`#{klass.name&.tableize || klass.table_name}.uniq!(:group).#{operation}(#{column_name.inspect})`).
+          MSG
+          group_fields = group_values
+        end
 
-        if group_attrs.first.respond_to?(:to_sym)
-          association  = @klass._reflect_on_association(group_attrs.first)
-          associated   = group_attrs.size == 1 && association && association.belongs_to? # only count belongs_to associations
-          group_fields = Array(associated ? association.foreign_key : group_attrs)
-        else
-          group_fields = group_attrs
+        if group_fields.size == 1 && group_fields.first.respond_to?(:to_sym)
+          association  = klass._reflect_on_association(group_fields.first)
+          associated   = association && association.belongs_to? # only count belongs_to associations
+          group_fields = Array(association.foreign_key) if associated
         end
         group_fields = arel_columns(group_fields)
 
-        group_aliases = group_fields.map { |field| column_alias_for(field) }
+        group_aliases = group_fields.map { |field|
+          field = connection.visitor.compile(field) if Arel.arel_node?(field)
+          column_alias_for(field.to_s.downcase)
+        }
         group_columns = group_aliases.zip(group_fields)
 
-        if operation == "count" && column_name == :all
-          aggregate_alias = "count_all"
-        else
-          aggregate_alias = column_alias_for([operation, column_name].join(" "))
-        end
+        column = aggregate_column(column_name)
+        column_alias = column_alias_for("#{operation} #{column_name.to_s.downcase}")
+        select_value = operation_over_aggregate_column(column, operation, distinct)
+        select_value.as(column_alias)
 
-        select_values = [
-          operation_over_aggregate_column(
-            aggregate_column(column_name),
-            operation,
-            distinct).as(aggregate_alias)
-        ]
+        select_values = [select_value]
         select_values += self.select_values unless having_clause.empty?
 
         select_values.concat group_columns.map { |aliaz, field|
@@ -339,43 +365,50 @@ module ActiveRecord
         if association
           key_ids     = calculated_data.collect { |row| row[group_aliases.first] }
           key_records = association.klass.base_class.where(association.klass.base_class.primary_key => key_ids)
-          key_records = Hash[key_records.map { |r| [r.id, r] }]
+          key_records = key_records.index_by(&:id)
         end
 
-        Hash[calculated_data.map do |row|
-          key = group_columns.map { |aliaz, col_name|
-            type = type_for(col_name) do
-              calculated_data.column_types.fetch(aliaz, Type.default_value)
-            end
-            type_cast_calculated_value(row[aliaz], type)
-          }
+        key_types = group_columns.each_with_object({}) do |(aliaz, col_name), types|
+          types[aliaz] = type_for(col_name) do
+            calculated_data.column_types.fetch(aliaz, Type.default_value)
+          end
+        end
+
+        hash_rows = calculated_data.cast_values(key_types).map! do |row|
+          calculated_data.columns.each_with_object({}).with_index do |(col_name, hash), i|
+            hash[col_name] = row[i]
+          end
+        end
+
+        type = nil
+        hash_rows.each_with_object({}) do |row, result|
+          key = group_aliases.map { |aliaz| row[aliaz] }
           key = key.first if key.size == 1
           key = key_records[key] if associated
 
-          type = calculated_data.column_types.fetch(aggregate_alias) { type_for(column_name) }
-          [key, type_cast_calculated_value(row[aggregate_alias], type, operation)]
-        end]
+          result[key] = type_cast_calculated_value(row[column_alias], operation) do |value|
+            type ||= column.try(:type_caster) ||
+              lookup_cast_type_from_join_dependencies(column_name.to_s) || Type.default_value
+            type.deserialize(value)
+          end
+        end
       end
 
-      # Converts the given keys to the value that the database adapter returns as
+      # Converts the given field to the value that the database adapter returns as
       # a usable column name:
       #
       #   column_alias_for("users.id")                 # => "users_id"
       #   column_alias_for("sum(id)")                  # => "sum_id"
       #   column_alias_for("count(distinct users.id)") # => "count_distinct_users_id"
       #   column_alias_for("count(*)")                 # => "count_all"
-      def column_alias_for(keys)
-        if keys.respond_to? :name
-          keys = "#{keys.relation.name}.#{keys.name}"
-        end
-
-        table_name = keys.to_s.downcase
-        table_name.gsub!(/\*/, "all")
-        table_name.gsub!(/\W+/, " ")
-        table_name.strip!
-        table_name.gsub!(/ +/, "_")
-
-        @klass.connection.table_alias_for(table_name)
+      def column_alias_for(field)
+        column_alias = +field
+        column_alias.gsub!(/\*/, "all")
+        column_alias.gsub!(/\W+/, " ")
+        column_alias.strip!
+        column_alias.gsub!(/ +/, "_")
+
+        connection.table_alias_for(column_alias)
       end
 
       def type_for(field, &block)
@@ -383,12 +416,41 @@ module ActiveRecord
         @klass.type_for_attribute(field_name, &block)
       end
 
-      def type_cast_calculated_value(value, type, operation = nil)
+      def lookup_cast_type_from_join_dependencies(name, join_dependencies = build_join_dependencies)
+        each_join_dependencies(join_dependencies) do |join|
+          type = join.base_klass.attribute_types.fetch(name, nil)
+          return type if type
+        end
+        nil
+      end
+
+      def type_cast_pluck_values(result, columns)
+        cast_types = if result.columns.size != columns.size
+          klass.attribute_types
+        else
+          join_dependencies = nil
+          columns.map.with_index do |column, i|
+            column.try(:type_caster) ||
+              klass.attribute_types.fetch(name = result.columns[i]) do
+                join_dependencies ||= build_join_dependencies
+                lookup_cast_type_from_join_dependencies(name, join_dependencies) ||
+                  result.column_types[name] || Type.default_value
+              end
+          end
+        end
+        result.cast_values(cast_types)
+      end
+
+      def type_cast_calculated_value(value, operation)
         case operation
-        when "count"   then value.to_i
-        when "sum"     then type.deserialize(value || 0)
-        when "average" then value && value.respond_to?(:to_d) ? value.to_d : value
-        else type.deserialize(value)
+        when "count"
+          value.to_i
+        when "sum"
+          yield value || 0
+        when "average"
+          value&.respond_to?(:to_d) ? value.to_d : value
+        else # "minimum", "maximum"
+          yield value
         end
       end
 
@@ -403,16 +465,17 @@ module ActiveRecord
 
       def build_count_subquery(relation, column_name, distinct)
         if column_name == :all
+          column_alias = Arel.star
           relation.select_values = [ Arel.sql(FinderMethods::ONE_AS_ONE) ] unless distinct
         else
           column_alias = Arel.sql("count_column")
           relation.select_values = [ aggregate_column(column_name).as(column_alias) ]
         end
 
-        subquery = relation.arel.as(Arel.sql("subquery_for_count"))
-        select_value = operation_over_aggregate_column(column_alias || Arel.star, "count", false)
+        subquery_alias = Arel.sql("subquery_for_count")
+        select_value = operation_over_aggregate_column(column_alias, "count", false)
 
-        Arel::SelectManager.new(subquery).project(select_value)
+        relation.build_subquery(subquery_alias, select_value)
       end
   end
 end