diff options
Diffstat (limited to 'activerecord/lib')
7 files changed, 97 insertions, 53 deletions
diff --git a/activerecord/lib/active_record/associations/collection_proxy.rb b/activerecord/lib/active_record/associations/collection_proxy.rb index 5d1e7ffb73..17ccf5a86c 100644 --- a/activerecord/lib/active_record/associations/collection_proxy.rb +++ b/activerecord/lib/active_record/associations/collection_proxy.rb @@ -793,7 +793,7 @@ module ActiveRecord # Returns +true+ if the collection is empty. If the collection has been # loaded it is equivalent # to <tt>collection.size.zero?</tt>. If the collection has not been loaded, - # it is equivalent to <tt>collection.exists?</tt>. If the collection has + # it is equivalent to <tt>!collection.exists?</tt>. If the collection has # not already been loaded and you are going to fetch the records anyway it # is better to check <tt>collection.length.zero?</tt>. # diff --git a/activerecord/lib/active_record/connection_adapters/sqlite3_adapter.rb b/activerecord/lib/active_record/connection_adapters/sqlite3_adapter.rb index eb2268157b..3384012c49 100644 --- a/activerecord/lib/active_record/connection_adapters/sqlite3_adapter.rb +++ b/activerecord/lib/active_record/connection_adapters/sqlite3_adapter.rb @@ -159,7 +159,7 @@ module ActiveRecord end # Returns 62. SQLite supports index names up to 64 - # characters. The rest is used by rails internally to perform + # characters. The rest is used by Rails internally to perform # temporary rename operations def allowed_index_name_length index_name_length - 2 diff --git a/activerecord/lib/active_record/core.rb b/activerecord/lib/active_record/core.rb index de337b24d6..2c94463acd 100644 --- a/activerecord/lib/active_record/core.rb +++ b/activerecord/lib/active_record/core.rb @@ -72,11 +72,20 @@ module ActiveRecord ## # :singleton-method: - # Specifies if an error should be raised on query limit or order being + # Specifies if an error should be raised if the query has an order being # ignored when doing batch queries. Useful in applications where the - # limit or scope being ignored is error-worthy, rather than a warning. + # scope being ignored is error-worthy, rather than a warning. + mattr_accessor :error_on_ignored_order, instance_writer: false + self.error_on_ignored_order = false + mattr_accessor :error_on_ignored_order_or_limit, instance_writer: false - self.error_on_ignored_order_or_limit = false + def self.error_on_ignored_order_or_limit=(value) + ActiveSupport::Deprecation.warn(<<-MSG.squish) + The flag error_on_ignored_order_or_limit is deprecated. Limits are + now supported. Please use error_on_ignored_order= instead. + MSG + self.error_on_ignored_order = value + end ## # :singleton-method: diff --git a/activerecord/lib/active_record/model_schema.rb b/activerecord/lib/active_record/model_schema.rb index 7996c32bbc..114686c5d3 100644 --- a/activerecord/lib/active_record/model_schema.rb +++ b/activerecord/lib/active_record/model_schema.rb @@ -282,8 +282,12 @@ module ActiveRecord # # +attr_name+ The name of the attribute to retrieve the type for. Must be # a string - def type_for_attribute(attr_name) - attribute_types[attr_name] + def type_for_attribute(attr_name, &block) + if block + attribute_types.fetch(attr_name, &block) + else + attribute_types[attr_name] + end end # Returns a hash where the keys are column names and the values are diff --git a/activerecord/lib/active_record/railtie.rb b/activerecord/lib/active_record/railtie.rb index 98ea425d16..2c0ca62924 100644 --- a/activerecord/lib/active_record/railtie.rb +++ b/activerecord/lib/active_record/railtie.rb @@ -3,7 +3,7 @@ require "rails" require "active_model/railtie" # For now, action_controller must always be present with -# rails, so let's make sure that it gets required before +# Rails, so let's make sure that it gets required before # here. This is needed for correctly setting up the middleware. # In the future, this might become an optional require. require "action_controller/railtie" diff --git a/activerecord/lib/active_record/relation/batches.rb b/activerecord/lib/active_record/relation/batches.rb index 3639625722..20ed4526b0 100644 --- a/activerecord/lib/active_record/relation/batches.rb +++ b/activerecord/lib/active_record/relation/batches.rb @@ -1,8 +1,8 @@ -require "active_record/relation/batches/batch_enumerator" +require 'active_record/relation/batches/batch_enumerator' module ActiveRecord module Batches - ORDER_OR_LIMIT_IGNORED_MESSAGE = "Scoped order and limit are ignored, it's forced to be batch order and batch size." + ORDER_IGNORE_MESSAGE = "Scoped order is ignored, it's forced to be batch order." # Looping through a collection of records from the database # (using the Scoping::Named::ClassMethods.all method, for example) @@ -34,15 +34,19 @@ module ActiveRecord # * <tt>:start</tt> - Specifies the primary key value to start from, inclusive of the value. # * <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 - # the order and limit have to be ignored due to batching. + # an order is present in the relation. # - # This is especially useful if you want multiple workers dealing with - # the same processing queue. You can make worker 1 handle all the records - # between id 0 and 10,000 and worker 2 handle from 10,000 and beyond - # (by setting the +:start+ and +:finish+ option on each worker). + # 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. # - # # Let's process for a batch of 2000 records, skipping the first 2000 rows - # Person.find_each(start: 2000, batch_size: 2000) do |person| + # The options +start+ and +finish+ are especially useful if you want + # multiple workers dealing with the same processing queue. You can make + # worker 1 handle all the records between id 1 and 9999 and worker 2 + # handle from 10000 and beyond by setting the +:start+ and +:finish+ + # option on each worker. + # + # # Let's process from record 10_000 on. + # Person.find_each(start: 10_000) do |person| # person.party_all_night! # end # @@ -51,8 +55,8 @@ module ActiveRecord # work. This also means that this method only works when the primary key is # orderable (e.g. an integer or string). # - # NOTE: You can't set the limit either, that's used to control - # the batch sizes. + # 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) if block_given? find_in_batches(start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do |records| @@ -89,15 +93,19 @@ module ActiveRecord # * <tt>:start</tt> - Specifies the primary key value to start from, inclusive of the value. # * <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 - # the order and limit have to be ignored due to batching. + # an order is present in the relation. + # + # 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. # - # This is especially useful if you want multiple workers dealing with - # the same processing queue. You can make worker 1 handle all the records - # between id 0 and 10,000 and worker 2 handle from 10,000 and beyond - # (by setting the +:start+ and +:finish+ option on each worker). + # The options +start+ and +finish+ are especially useful if you want + # multiple workers dealing with the same processing queue. You can make + # worker 1 handle all the records between id 1 and 9999 and worker 2 + # handle from 10000 and beyond by setting the +:start+ and +:finish+ + # option on each worker. # - # # Let's process the next 2000 records - # Person.find_in_batches(start: 2000, batch_size: 2000) do |group| + # # Let's process from record 10_000 on. + # Person.find_in_batches(start: 10_000) do |group| # group.each { |person| person.party_all_night! } # end # @@ -106,8 +114,8 @@ module ActiveRecord # work. This also means that this method only works when the primary key is # orderable (e.g. an integer or string). # - # NOTE: You can't set the limit either, that's used to control - # the batch sizes. + # 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) relation = self unless block_given? @@ -149,17 +157,19 @@ module ActiveRecord # * <tt>:start</tt> - Specifies the primary key value to start from, inclusive of the value. # * <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 - # the order and limit have to be ignored due to batching. + # an order is present in the relation. + # + # 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. # - # This is especially useful if you want to work with the - # ActiveRecord::Relation object instead of the array of records, or if - # you want multiple workers dealing with the same processing queue. You can - # make worker 1 handle all the records between id 0 and 10,000 and worker 2 - # handle from 10,000 and beyond (by setting the +:start+ and +:finish+ - # option on each worker). + # The options +start+ and +finish+ are especially useful if you want + # multiple workers dealing with the same processing queue. You can make + # worker 1 handle all the records between id 1 and 9999 and worker 2 + # handle from 10000 and beyond by setting the +:start+ and +:finish+ + # option on each worker. # - # # Let's process the next 2000 records - # Person.in_batches(of: 2000, start: 2000).update_all(awesome: true) + # # Let's process from record 10_000 on. + # Person.in_batches(start: 10_000).update_all(awesome: true) # # An example of calling where query method on the relation: # @@ -179,19 +189,25 @@ module ActiveRecord # consistent. Therefore the primary key must be orderable, e.g an integer # or a string. # - # NOTE: You can't set the limit either, that's used to control the batch - # sizes. + # 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) relation = self unless block_given? return BatchEnumerator.new(of: of, start: start, finish: finish, relation: self) end - if arel.orders.present? || arel.taken.present? - act_on_order_or_limit_ignored(error_on_ignore) + if arel.orders.present? + act_on_ignored_order(error_on_ignore) + end + + batch_limit = of + if limit_value + remaining = limit_value + batch_limit = remaining if remaining < batch_limit end - relation = relation.reorder(batch_order).limit(of) + relation = relation.reorder(batch_order).limit(batch_limit) relation = apply_limits(relation, start, finish) batch_relation = relation @@ -199,11 +215,11 @@ module ActiveRecord if load records = batch_relation.records ids = records.map(&:id) - yielded_relation = self.where(primary_key => ids) + yielded_relation = where(primary_key => ids) yielded_relation.load_records(records) else ids = batch_relation.pluck(primary_key) - yielded_relation = self.where(primary_key => ids) + yielded_relation = where(primary_key => ids) end break if ids.empty? @@ -213,7 +229,20 @@ module ActiveRecord yield yielded_relation - break if ids.length < of + break if ids.length < batch_limit + + if limit_value + remaining -= ids.length + + if remaining == 0 + # Saves a useless iteration when the limit is a multiple of the + # batch size. + break + elsif remaining < batch_limit + relation = relation.limit(remaining) + end + end + batch_relation = relation.where(arel_attribute(primary_key).gt(primary_key_offset)) end end @@ -230,13 +259,13 @@ module ActiveRecord "#{quoted_table_name}.#{quoted_primary_key} ASC" end - def act_on_order_or_limit_ignored(error_on_ignore) - raise_error = (error_on_ignore.nil? ? self.klass.error_on_ignored_order_or_limit : error_on_ignore) + def act_on_ignored_order(error_on_ignore) + raise_error = (error_on_ignore.nil? ? self.klass.error_on_ignored_order : error_on_ignore) if raise_error - raise ArgumentError.new(ORDER_OR_LIMIT_IGNORED_MESSAGE) + raise ArgumentError.new(ORDER_IGNORE_MESSAGE) elsif logger - logger.warn(ORDER_OR_LIMIT_IGNORED_MESSAGE) + logger.warn(ORDER_IGNORE_MESSAGE) end end end diff --git a/activerecord/lib/active_record/relation/calculations.rb b/activerecord/lib/active_record/relation/calculations.rb index d6d92b8607..a97b71815a 100644 --- a/activerecord/lib/active_record/relation/calculations.rb +++ b/activerecord/lib/active_record/relation/calculations.rb @@ -312,8 +312,10 @@ module ActiveRecord Hash[calculated_data.map do |row| key = group_columns.map { |aliaz, col_name| - column = calculated_data.column_types.fetch(aliaz) do - type_for(col_name) + column = type_for(col_name) do + calculated_data.column_types.fetch(aliaz) do + Type::Value.new + end end type_cast_calculated_value(row[aliaz], column) } @@ -346,9 +348,9 @@ module ActiveRecord @klass.connection.table_alias_for(table_name) end - def type_for(field) + def type_for(field, &block) field_name = field.respond_to?(:name) ? field.name.to_s : field.to_s.split('.').last - @klass.type_for_attribute(field_name) + @klass.type_for_attribute(field_name, &block) end def type_cast_calculated_value(value, type, operation = nil) |