Active Record Relation

Included Modules
MULTI_VALUE_METHODS = [:includes, :eager_load, :preload, :select, :group, :order, :joins, :left_outer_joins, :references, :extending, :unscope, :optimizer_hints, :annotate]
SINGLE_VALUE_METHODS = [:limit, :offset, :lock, :readonly, :reordering, :reverse_order, :distinct, :create_with, :skip_query_cache]
CLAUSE_METHODS = [:where, :having, :from]
INVALID_METHODS_FOR_DELETE_ALL = [:distinct, :group, :having]
[R] klass
[R] loaded
[R] loaded?
[R] model
[R] predicate_builder
[RW] skip_preloading_value
[R] table
Class Public methods
new(klass, table: klass.arel_table, predicate_builder: klass.predicate_builder, values: {})
   # File activerecord/lib/active_record/relation.rb
27 def initialize(klass, table: klass.arel_table, predicate_builder: klass.predicate_builder, values: {})
28   @klass  = klass
29   @table  = table
30   @values = values
31   @offsets = {}
32   @loaded = false
33   @predicate_builder = predicate_builder
34   @delegate_to_klass = false
35 end
Instance Public methods

Compares two relations for equality.

    # File activerecord/lib/active_record/relation.rb
681 def ==(other)
682   case other
683   when Associations::CollectionProxy, AssociationRelation
684     self == other.records
685   when Relation
686     other.to_sql == to_sql
687   when Array
688     records == other
689   end
690 end

Returns true if there are any records.

    # File activerecord/lib/active_record/relation.rb
277 def any?
278   return super if block_given?
279   !empty?
280 end

Returns true if relation is blank.

    # File activerecord/lib/active_record/relation.rb
697 def blank?
698   records.blank?
699 end
build(attributes = nil, &block)
Alias for: new
cache_key(timestamp_column = :updated_at)

Returns a stable cache key that can be used to identify this query. The cache key is built with a fingerprint of the SQL query.

Product.where("name like ?", "%Cosmic Encounter%").cache_key
# => "products/query-1850ab3d302391b85b8693e941286659"

If ActiveRecord::Base.collection_cache_versioning is turned off, as it was in Rails 6.0 and earlier, the cache key will also include a version.

ActiveRecord::Base.collection_cache_versioning = false
Product.where("name like ?", "%Cosmic Encounter%").cache_key
# => "products/query-1850ab3d302391b85b8693e941286659-1-20150714212553907087000"

You can also pass a custom timestamp column to fetch the timestamp of the last updated record.

Product.where("name like ?", "%Game%").cache_key(:last_reviewed_at)
    # File activerecord/lib/active_record/relation.rb
311 def cache_key(timestamp_column = :updated_at)
312   @cache_keys ||= {}
313   @cache_keys[timestamp_column] ||= klass.collection_cache_key(self, timestamp_column)
314 end
cache_version(timestamp_column = :updated_at)

Returns a cache version that can be used together with the cache key to form a recyclable caching scheme. The cache version is built with the number of records matching the query, and the timestamp of the last updated record. When a new record comes to match the query, or any of the existing records is updated or deleted, the cache version changes.

If the collection is loaded, the method will iterate through the records to generate the timestamp, otherwise it will trigger one SQL query like:

SELECT COUNT(*), MAX("products"."updated_at") FROM "products" WHERE (name like '%Cosmic Encounter%')
    # File activerecord/lib/active_record/relation.rb
338 def cache_version(timestamp_column = :updated_at)
339   if collection_cache_versioning
340     @cache_versions ||= {}
341     @cache_versions[timestamp_column] ||= compute_cache_version(timestamp_column)
342   end
343 end
create(attributes = nil, &block)

Tries to create a new record with the same scoped attributes defined in the relation. Returns the initialized object if validation fails.

Expects arguments in the same format as ActiveRecord::Base.create.


users = User.where(name: 'Oscar')
users.create # => #<User id: 3, name: "Oscar", ...>

users.create(name: 'fxn')
users.create # => #<User id: 4, name: "fxn", ...>

users.create { |user| = 'tenderlove' }
# => #<User id: 5, name: "tenderlove", ...>

users.create(name: nil) # validation on name
# => #<User id: nil, name: nil, ...>
    # File activerecord/lib/active_record/relation.rb
 95 def create(attributes = nil, &block)
 96   if attributes.is_a?(Array)
 97     attributes.collect { |attr| create(attr, &block) }
 98   else
 99     block = _deprecated_scope_block("create", &block)
100     scoping { klass.create(attributes, &block) }
101   end
102 end
create!(attributes = nil, &block)

Similar to create, but calls create! on the base class. Raises an exception if a validation error occurs.

Expects arguments in the same format as create!.

    # File activerecord/lib/active_record/relation.rb
110 def create!(attributes = nil, &block)
111   if attributes.is_a?(Array)
112     attributes.collect { |attr| create!(attr, &block) }
113   else
114     block = _deprecated_scope_block("create!", &block)
115     scoping { klass.create!(attributes, &block) }
116   end
117 end
create_or_find_by(attributes, &block)

Attempts to create a record with the given attributes in a table that has a unique constraint on one or several of its columns. If a row already exists with one or several of these unique constraints, the exception such an insertion would normally raise is caught, and the existing record with those attributes is found using find_by!.

This is similar to find_or_create_by, but avoids the problem of stale reads between the SELECT and the INSERT, as that method needs to first query the table, then attempt to insert a row if none is found.

There are several drawbacks to create_or_find_by, though:

  • The underlying table must have the relevant columns defined with unique constraints.

  • A unique constraint violation may be triggered by only one, or at least less than all, of the given attributes. This means that the subsequent find_by! may fail to find a matching record, which will then raise an ActiveRecord::RecordNotFound exception, rather than a record with the given attributes.

  • While we avoid the race condition between SELECT -> INSERT from find_or_create_by, we actually have another race condition between INSERT -> SELECT, which can be triggered if a DELETE between those two statements is run by another client. But for most applications, that's a significantly less likely condition to hit.

  • It relies on exception handling to handle control flow, which may be marginally slower.

  • The primary key may auto-increment on each create, even if it fails. This can accelerate the problem of running out of integers, if the underlying table is still stuck on a primary key of type int (note: All Rails apps since 5.1+ have defaulted to bigint, which is not liable to this problem).

This method will return a record if all given attributes are covered by unique constraints (unless the INSERT -> DELETE -> SELECT race condition is triggered), but if creation was attempted and failed due to validation errors it won't be persisted, you get what create returns in such situation.

    # File activerecord/lib/active_record/relation.rb
209 def create_or_find_by(attributes, &block)
210   transaction(requires_new: true) { create(attributes, &block) }
211 rescue ActiveRecord::RecordNotUnique
212   find_by!(attributes)
213 end
create_or_find_by!(attributes, &block)

Like create_or_find_by, but calls create! so an exception is raised if the created record is invalid.

    # File activerecord/lib/active_record/relation.rb
218 def create_or_find_by!(attributes, &block)
219   transaction(requires_new: true) { create!(attributes, &block) }
220 rescue ActiveRecord::RecordNotUnique
221   find_by!(attributes)
222 end

Deletes the records without instantiating the records first, and hence not calling the #destroy method nor invoking callbacks. This is a single SQL DELETE statement that goes straight to the database, much more efficient than destroy_all. Be careful with relations though, in particular :dependent rules defined on associations are not honored. Returns the number of rows affected.

Post.where(person_id: 5).where(category: ['Something', 'Else']).delete_all

Both calls delete the affected posts all at once with a single DELETE statement. If you need to destroy dependent associations or call your before_* or after_destroy callbacks, use the destroy_all method instead.

If an invalid method is supplied, delete_all raises an ActiveRecordError:

# => ActiveRecord::ActiveRecordError: delete_all doesn't support distinct
    # File activerecord/lib/active_record/relation.rb
554 def delete_all
555   invalid_methods = do |method|
556     value = @values[method]
557     method == :distinct ? value : value&.any?
558   end
559   if invalid_methods.any?
560     raise"delete_all doesn't support #{invalid_methods.join(', ')}")
561   end
563   if eager_loading?
564     relation = apply_join_dependency
565     return relation.delete_all
566   end
568   stmt =
569   stmt.from(arel.join_sources.empty? ? table : arel.source)
570   stmt.key = arel_attribute(primary_key)
571   stmt.take(arel.limit)
572   stmt.offset(arel.offset)
573   stmt.order(*arel.orders)
574   stmt.wheres = arel.constraints
576   affected = @klass.connection.delete(stmt, "#{@klass} Destroy")
578   reset
579   affected
580 end

Finds and deletes all records matching the specified conditions. This is short-hand for relation.where(condition).delete_all. Returns the number of rows affected.

If no record is found, returns 0 as zero rows were affected.

Person.delete_by(id: 13)
Person.delete_by(name: 'Spartacus', rating: 4)
Person.delete_by("published_at < ?", 2.weeks.ago)
    # File activerecord/lib/active_record/relation.rb
604 def delete_by(*args)
605   where(*args).delete_all
606 end

Destroys the records by instantiating each record and calling its #destroy method. Each object's callbacks are executed (including :dependent association options). Returns the collection of objects that were destroyed; each will be frozen, to reflect that no changes should be made (since they can't be persisted).

Note: Instantiation, callback execution, and deletion of each record can be time consuming when you're removing many records at once. It generates at least one SQL DELETE query per record (or possibly more, to enforce your callbacks). If you want to delete many rows quickly, without concern for their associations or callbacks, use delete_all instead.


Person.where(age: 0..18).destroy_all
    # File activerecord/lib/active_record/relation.rb
532 def destroy_all
533   records.each(&:destroy).tap { reset }
534 end

Finds and destroys all records matching the specified conditions. This is short-hand for relation.where(condition).destroy_all. Returns the collection of objects that were destroyed.

If no record is found, returns empty array.

Person.destroy_by(id: 13)
Person.destroy_by(name: 'Spartacus', rating: 4)
Person.destroy_by("published_at < ?", 2.weeks.ago)
    # File activerecord/lib/active_record/relation.rb
591 def destroy_by(*args)
592   where(*args).destroy_all
593 end

Returns true if relation needs eager loading.

    # File activerecord/lib/active_record/relation.rb
666 def eager_loading?
667   @should_eager_load ||=
668     eager_load_values.any? ||
669     includes_values.any? && (joined_includes_values.any? || references_eager_loaded_tables?)
670 end

Returns true if there are no records.

    # File activerecord/lib/active_record/relation.rb
265 def empty?
266   return @records.empty? if loaded?
267   !exists?
268 end

Serializes the relation objects Array.

    # File activerecord/lib/active_record/relation.rb
255 def encode_with(coder)
256   coder.represent_seq(nil, records)
257 end

Runs EXPLAIN on the query or queries triggered by this relation and returns the result as a string. The string is formatted imitating the ones printed by the database shell.

Note that this method actually runs the queries, since the results of some are needed by the next ones when eager loading is going on.

Please see further details in the Active Record Query Interface guide.

    # File activerecord/lib/active_record/relation.rb
239 def explain
240   exec_explain(collecting_queries_for_explain { exec_queries })
241 end
find_or_create_by(attributes, &block)

Finds the first record with the given attributes, or creates a record with the attributes if one is not found:

# Find the first user named "Penélope" or create a new one.
User.find_or_create_by(first_name: 'Penélope')
# => #<User id: 1, first_name: "Penélope", last_name: nil>

# Find the first user named "Penélope" or create a new one.
# We already have one so the existing record will be returned.
User.find_or_create_by(first_name: 'Penélope')
# => #<User id: 1, first_name: "Penélope", last_name: nil>

# Find the first user named "Scarlett" or create a new one with
# a particular last name.
User.create_with(last_name: 'Johansson').find_or_create_by(first_name: 'Scarlett')
# => #<User id: 2, first_name: "Scarlett", last_name: "Johansson">

This method accepts a block, which is passed down to create. The last example above can be alternatively written this way:

# Find the first user named "Scarlett" or create a new one with a
# different last name.
User.find_or_create_by(first_name: 'Scarlett') do |user|
  user.last_name = 'Johansson'
# => #<User id: 2, first_name: "Scarlett", last_name: "Johansson">

This method always returns a record, but if creation was attempted and failed due to validation errors it won't be persisted, you get what create returns in such situation.

Please note this method is not atomic, it runs first a SELECT, and if there are no results an INSERT is attempted. If there are other threads or processes there is a race condition between both calls and it could be the case that you end up with two similar records.

If this might be a problem for your application, please see create_or_find_by.

    # File activerecord/lib/active_record/relation.rb
168 def find_or_create_by(attributes, &block)
169   find_by(attributes) || create(attributes, &block)
170 end
find_or_create_by!(attributes, &block)

Like find_or_create_by, but calls create! so an exception is raised if the created record is invalid.

    # File activerecord/lib/active_record/relation.rb
175 def find_or_create_by!(attributes, &block)
176   find_by(attributes) || create!(attributes, &block)
177 end
find_or_initialize_by(attributes, &block)

Like find_or_create_by, but calls new instead of create.

    # File activerecord/lib/active_record/relation.rb
226 def find_or_initialize_by(attributes, &block)
227   find_by(attributes) || new(attributes, &block)
228 end
   # File activerecord/lib/active_record/relation.rb
37 def initialize_copy(other)
38   @values = @values.dup
39   reset
40 end
    # File activerecord/lib/active_record/relation.rb
705 def inspect
706   subject = loaded? ? records : self
707   entries = subject.take([limit_value, 11].compact.min).map!(&:inspect)
709   entries[10] = "..." if entries.size == 11
711   "#<#{} [#{entries.join(', ')}]>"
712 end

Joins that are also marked for preloading. In which case we should just eager load them. Note that this is a naive implementation because we could have strings and symbols which represent the same association, but that aren't matched by this. Also, we could have nested hashes which partially match, e.g. { a: :b } & { a: [:b, :c] }

    # File activerecord/lib/active_record/relation.rb
676 def joined_includes_values
677   includes_values & joins_values
678 end

Causes the records to be loaded from the database if they have not been loaded already. You can use this if for some reason you need to explicitly load some records before actually using them. The return value is the relation itself, not the records.

Post.where(published: true).load # => #<ActiveRecord::Relation>
    # File activerecord/lib/active_record/relation.rb
614 def load(&block)
615   exec_queries(&block) unless loaded?
617   self
618 end

Returns true if there is more than one record.

    # File activerecord/lib/active_record/relation.rb
289 def many?
290   return super if block_given?
291   limit_value ? records.many? : size > 1
292 end
new(attributes = nil, &block)

Initializes new record from relation while maintaining the current scope.

Expects arguments in the same format as

users = User.where(name: 'DHH')
user = # => #<User id: nil, name: "DHH", created_at: nil, updated_at: nil>

You can also pass a block to new with the new record as argument:

user = { |user| = 'Oscar' } # => Oscar
Also aliased as: build
   # File activerecord/lib/active_record/relation.rb
69 def new(attributes = nil, &block)
70   block = _deprecated_scope_block("new", &block)
71   scoping {, &block) }
72 end

Returns true if there are no records.

    # File activerecord/lib/active_record/relation.rb
271 def none?
272   return super if block_given?
273   empty?
274 end

Returns true if there is exactly one record.

    # File activerecord/lib/active_record/relation.rb
283 def one?
284   return super if block_given?
285   limit_value ? : size == 1
286 end
    # File activerecord/lib/active_record/relation.rb
692 def pretty_print(q)
693   q.pp(records)
694 end

Forces reloading of relation.

    # File activerecord/lib/active_record/relation.rb
621 def reload
622   reset
623   load
624 end
    # File activerecord/lib/active_record/relation.rb
626 def reset
627   @delegate_to_klass = false
628   @_deprecated_scope_source = nil
629   @to_sql = @arel = @loaded = @should_eager_load = nil
630   @records = [].freeze
631   @offsets = {}
632   self
633 end
    # File activerecord/lib/active_record/relation.rb
661 def scope_for_create
662   where_values_hash.merge!(create_with_value.stringify_keys)
663 end

Scope all queries to the current scope.

Comment.where(post_id: 1).scoping do
# => SELECT "comments".* FROM "comments" WHERE "comments"."post_id" = 1 ORDER BY "comments"."id" ASC LIMIT 1

Please check unscoped if you want to remove all previous scopes (including the default_scope) during the execution of a block.

    # File activerecord/lib/active_record/relation.rb
397 def scoping
398   already_in_scope? ? yield : _scoping(self) { yield }
399 end

Returns size of the records.

    # File activerecord/lib/active_record/relation.rb
260 def size
261   loaded? ? @records.length : count(:all)
262 end
Alias for: to_ary

Converts relation objects to Array.

Also aliased as: to_a
    # File activerecord/lib/active_record/relation.rb
244 def to_ary
245   records.dup
246 end

Returns sql statement for the relation.

User.where(name: 'Oscar').to_sql
# => SELECT "users".* FROM "users"  WHERE "users"."name" = 'Oscar'
    # File activerecord/lib/active_record/relation.rb
639 def to_sql
640   @to_sql ||= begin
641     if eager_loading?
642       apply_join_dependency do |relation, join_dependency|
643         relation = join_dependency.apply_column_aliases(relation)
644         relation.to_sql
645       end
646     else
647       conn = klass.connection
648       conn.unprepared_statement { conn.to_sql(arel) }
649     end
650   end
651 end
touch_all(*names, time: nil)

Touches all records in the current relation without instantiating records first with the updated_at/updated_on attributes set to the current time or the time specified. This method can be passed attribute names and an optional time argument. If attribute names are passed, they are updated along with updated_at/updated_on attributes. If no time argument is passed, the current time is used as default.


# Touch all records
# => "UPDATE \"people\" SET \"updated_at\" = '2018-01-04 22:55:23.132670'"

# Touch multiple records with a custom attribute
# => "UPDATE \"people\" SET \"updated_at\" = '2018-01-04 22:55:23.132670', \"created_at\" = '2018-01-04 22:55:23.132670'"

# Touch multiple records with a specified time
Person.all.touch_all(time:, 5, 16, 0, 0, 0))
# => "UPDATE \"people\" SET \"updated_at\" = '2020-05-16 00:00:00'"

# Touch records with scope
Person.where(name: 'David').touch_all
# => "UPDATE \"people\" SET \"updated_at\" = '2018-01-04 22:55:23.132670' WHERE \"people\".\"name\" = 'David'"
    # File activerecord/lib/active_record/relation.rb
512 def touch_all(*names, time: nil)
513   update_all klass.touch_attributes_with_time(*names, time: time)
514 end

Updates all records in the current relation with details given. This method constructs a single SQL UPDATE statement and sends it straight to the database. It does not instantiate the involved models and it does not trigger Active Record callbacks or validations. However, values passed to update_all will still go through Active Record's normal type casting and serialization.

Note: As Active Record callbacks are not triggered, this method will not automatically update updated_at/updated_on columns.


  • updates - A string, array, or hash representing the SET part of an SQL statement.


# Update all customers with the given attributes
Customer.update_all wants_email: true

# Update all books with 'Rails' in their title
Book.where('title LIKE ?', '%Rails%').update_all(author: 'David')

# Update all books that match conditions, but limit it to 5 ordered by date
Book.where('title LIKE ?', '%Rails%').order(:created_at).limit(5).update_all(author: 'David')

# Update all invoices and set the number column to its id value.
Invoice.update_all('number = id')
    # File activerecord/lib/active_record/relation.rb
432 def update_all(updates)
433   raise ArgumentError, "Empty list of attributes to change" if updates.blank?
435   if eager_loading?
436     relation = apply_join_dependency
437     return relation.update_all(updates)
438   end
440   stmt =
441   stmt.table(arel.join_sources.empty? ? table : arel.source)
442   stmt.key = arel_attribute(primary_key)
443   stmt.take(arel.limit)
444   stmt.offset(arel.offset)
445   stmt.order(*arel.orders)
446   stmt.wheres = arel.constraints
448   if updates.is_a?(Hash)
449     if klass.locking_enabled? &&
450         !updates.key?(klass.locking_column) &&
451         !updates.key?(klass.locking_column.to_sym)
452       attr = arel_attribute(klass.locking_column)
453       updates[] = _increment_attribute(attr)
454     end
455     stmt.set _substitute_values(updates)
456   else
457     stmt.set Arel.sql(klass.sanitize_sql_for_assignment(updates,
458   end
460   @klass.connection.update stmt, "#{@klass} Update All"
461 end
    # File activerecord/lib/active_record/relation.rb
701 def values
702   @values.dup
703 end
where_values_hash(relation_table_name = klass.table_name)

Returns a hash of where conditions.

User.where(name: 'Oscar').where_values_hash
# => {name: "Oscar"}
    # File activerecord/lib/active_record/relation.rb
657 def where_values_hash(relation_table_name = klass.table_name)
658   where_clause.to_h(relation_table_name)
659 end
Instance Protected methods
    # File activerecord/lib/active_record/relation.rb
742 def load_records(records)
743   @records = records.freeze
744   @loaded = true
745 end