diff options
| -rw-r--r-- | activerecord/CHANGELOG.md | 6 | ||||
| -rw-r--r-- | activerecord/lib/active_record/core.rb | 8 | ||||
| -rw-r--r-- | activerecord/lib/active_record/relation/batches.rb | 38 | ||||
| -rw-r--r-- | activerecord/test/cases/batches_test.rb | 36 | ||||
| -rw-r--r-- | guides/source/configuring.md | 2 |
5 files changed, 81 insertions, 9 deletions
diff --git a/activerecord/CHANGELOG.md b/activerecord/CHANGELOG.md index cd4b297c8c..7438ac8a9d 100644 --- a/activerecord/CHANGELOG.md +++ b/activerecord/CHANGELOG.md @@ -1,3 +1,9 @@ +* Added a configuration option to have active record raise an ArgumentError + if the order or limit is ignored in a batch query, rather than logging a + warning message. + + *Scott Ringwelski* + * Honour the order of the joining model in a `has_many :through` association when eager loading. Example: diff --git a/activerecord/lib/active_record/core.rb b/activerecord/lib/active_record/core.rb index 24fd0aaecf..86ec8000fb 100644 --- a/activerecord/lib/active_record/core.rb +++ b/activerecord/lib/active_record/core.rb @@ -72,6 +72,14 @@ module ActiveRecord ## # :singleton-method: + # Specifies if an error should be raised on query limit or order being + # ignored when doing batch queries. Useful in applications where the + # limit or scope being ignored is error-worthy, rather than a warning. + mattr_accessor :error_on_ignored_order_or_limit, instance_writer: false + self.error_on_ignored_order_or_limit = false + + ## + # :singleton-method: # Specify whether or not to use timestamps for migration versions mattr_accessor :timestamped_migrations, instance_writer: false self.timestamped_migrations = true diff --git a/activerecord/lib/active_record/relation/batches.rb b/activerecord/lib/active_record/relation/batches.rb index 243ef0eae9..b99807adf3 100644 --- a/activerecord/lib/active_record/relation/batches.rb +++ b/activerecord/lib/active_record/relation/batches.rb @@ -2,6 +2,8 @@ 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" + # Looping through a collection of records from the database # (using the Scoping::Named::ClassMethods.all method, for example) # is very inefficient since it will try to instantiate all the objects at once. @@ -31,6 +33,9 @@ module ActiveRecord # * <tt>:batch_size</tt> - Specifies the size of the batch. Default to 1000. # * <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. + # # 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 @@ -48,13 +53,13 @@ module ActiveRecord # # NOTE: You can't set the limit either, that's used to control # the batch sizes. - def find_each(start: nil, finish: nil, batch_size: 1000) + 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) do |records| + find_in_batches(start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do |records| records.each { |record| yield record } end else - enum_for(:find_each, start: start, finish: finish, batch_size: batch_size) do + enum_for(:find_each, start: start, finish: finish, batch_size: batch_size, error_on_ignore: error_on_ignore) do relation = self apply_limits(relation, start, finish).size end @@ -83,6 +88,9 @@ module ActiveRecord # * <tt>:batch_size</tt> - Specifies the size of the batch. Default to 1000. # * <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. + # # 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 @@ -100,16 +108,16 @@ module ActiveRecord # # NOTE: You can't set the limit either, that's used to control # the batch sizes. - def find_in_batches(start: nil, finish: nil, batch_size: 1000) + def find_in_batches(start: nil, finish: nil, batch_size: 1000, error_on_ignore: nil) relation = self unless block_given? - return to_enum(:find_in_batches, start: start, finish: finish, batch_size: batch_size) do + 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 (total - 1).div(batch_size) + 1 end end - in_batches(of: batch_size, start: start, finish: finish, load: true) do |batch| + in_batches(of: batch_size, start: start, finish: finish, load: true, error_on_ignore: error_on_ignore) do |batch| yield batch.to_a end end @@ -140,6 +148,8 @@ module ActiveRecord # * <tt>:load</tt> - Specifies if the relation should be loaded. Default to false. # * <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. # # This is especially useful if you want to work with the # ActiveRecord::Relation object instead of the array of records, or if @@ -171,14 +181,14 @@ module ActiveRecord # # NOTE: You can't set the limit either, that's used to control the batch # sizes. - def in_batches(of: 1000, start: nil, finish: nil, load: false) + 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 logger && (arel.orders.present? || arel.taken.present?) - logger.warn("Scoped order and limit are ignored, it's forced to be batch order and batch size") + if arel.orders.present? || arel.taken.present? + act_on_order_or_limit_ignored(error_on_ignore) end relation = relation.reorder(batch_order).limit(of) @@ -219,5 +229,15 @@ module ActiveRecord def batch_order "#{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) + + if raise_error + raise ArgumentError.new(ORDER_OR_LIMIT_IGNORED_MESSAGE) + elsif logger + logger.warn(ORDER_OR_LIMIT_IGNORED_MESSAGE) + end + end end end diff --git a/activerecord/test/cases/batches_test.rb b/activerecord/test/cases/batches_test.rb index 84aac3e721..91ff5146fd 100644 --- a/activerecord/test/cases/batches_test.rb +++ b/activerecord/test/cases/batches_test.rb @@ -164,6 +164,42 @@ class EachTest < ActiveRecord::TestCase assert_equal posts(:welcome).id, posts.first.id end + def test_find_in_batches_should_error_on_ignore_the_order + assert_raise(ArgumentError) do + PostWithDefaultScope.find_in_batches(error_on_ignore: true){} + end + end + + def test_find_in_batches_should_not_error_if_config_overriden + # Set the config option which will be overriden + prev = ActiveRecord::Base.error_on_ignored_order_or_limit + ActiveRecord::Base.error_on_ignored_order_or_limit = true + assert_nothing_raised do + PostWithDefaultScope.find_in_batches(error_on_ignore: false){} + end + ensure + # Set back to default + ActiveRecord::Base.error_on_ignored_order_or_limit = prev + end + + def test_find_in_batches_should_error_on_config_specified_to_error + # Set the config option + prev = ActiveRecord::Base.error_on_ignored_order_or_limit + ActiveRecord::Base.error_on_ignored_order_or_limit = true + assert_raise(ArgumentError) do + PostWithDefaultScope.find_in_batches(){} + end + ensure + # Set back to default + ActiveRecord::Base.error_on_ignored_order_or_limit = prev + end + + def test_find_in_batches_should_not_error_by_default + assert_nothing_raised do + PostWithDefaultScope.find_in_batches(){} + end + end + def test_find_in_batches_should_not_ignore_the_default_scope_if_it_is_other_then_order special_posts_ids = SpecialPostWithDefaultScope.all.map(&:id).sort posts = [] diff --git a/guides/source/configuring.md b/guides/source/configuring.md index 41985c3661..1d3ff6e6e7 100644 --- a/guides/source/configuring.md +++ b/guides/source/configuring.md @@ -281,6 +281,8 @@ All these configuration options are delegated to the `I18n` library. * `config.active_record.schema_format` controls the format for dumping the database schema to a file. The options are `:ruby` (the default) for a database-independent version that depends on migrations, or `:sql` for a set of (potentially database-dependent) SQL statements. +* `config.active_record.error_on_ignored_order_or_limit` specifies if an error should be raised if the order or limit of a query is ignored during a batch query. The options are true (raise error) or false (warn). Default is false. + * `config.active_record.timestamped_migrations` controls whether migrations are numbered with serial integers or with timestamps. The default is true, to use timestamps, which are preferred if there are multiple developers working on the same application. * `config.active_record.lock_optimistically` controls whether Active Record will use optimistic locking and is true by default. |