1 package DBIx::Class::Relationship;
6 use base qw/DBIx::Class/;
8 __PACKAGE__->load_own_components(qw/
18 DBIx::Class::Relationship - Inter-table relationships
22 ## Creating relationships
23 MyDB::Schema::Actor->has_many('actorroles' => 'MyDB::Schema::ActorRole',
25 MyDB::Schema::Role->has_many('actorroles' => 'MyDB::Schema::ActorRole',
27 MyDB::Schema::ActorRole->belongs_to('role' => 'MyDB::Schema::Role');
28 MyDB::Schema::ActorRole->belongs_to('actor' => 'MyDB::Schema::Actor');
30 MyDB::Schema::Role->many_to_many('actors' => 'actorroles', 'actor');
31 MyDB::Schema::Actor->many_to_many('roles' => 'actorroles', 'role');
33 ## Using relationships
34 $schema->resultset('Actor')->find({ id => 1})->roles();
35 $schema->resultset('Role')->find({ id => 1 })->actorroles->search_related('actor', { Name => 'Fred' });
36 $schema->resultset('Actor')->add_to_roles({ Name => 'Sherlock Holmes'});
38 See L<DBIx::Class::Manual::Cookbook> for more.
42 This class provides methods to set up relationships between the tables
43 in your database model. Relationships are the most useful and powerful
44 technique that L<DBIx::Class> provides. To create efficient database queries,
45 create relationships between any and all tables that have something in
46 common, for example if you have a table Authors:
57 1 | 1 | Rulers of the universe
58 2 | 1 | Rulers of the galaxy
60 Then without relationships, the method of getting all books by Fred goes like
63 my $fred = $schema->resultset('Author')->find({ Name => 'Fred' });
64 my $fredsbooks = $schema->resultset('Book')->search({ Author => $fred->ID });
66 With a has_many relationship called "books" on Author (see below for details),
67 we can do this instead:
69 my $fredsbooks = $schema->resultset('Author')->find({ Name => 'Fred' })->books;
71 Each relationship sets up an accessor method on the
72 L<DBIx::Class::Manual::Glossary/"Row"> objects that represent the items
73 of your table. From L<DBIx::Class::Manual::Glossary/"ResultSet"> objects,
74 the relationships can be searched using the "search_related" method.
75 In list context, each returns a list of Row objects for the related class,
76 in scalar context, a new ResultSet representing the joined tables is
77 returned. Thus, the calls can be chained to produce complex queries.
78 Since the database is not actually queried until you attempt to retrieve
79 the data for an actual item, no time is wasted producing them.
81 my $cheapfredbooks = $schema->resultset('Author')->find({
83 })->books->search_related('prices', {
84 Price => { '<=' => '5.00' },
87 will produce a query something like:
89 SELECT * FROM Author me
90 LEFT JOIN Books books ON books.author = me.id
91 LEFT JOIN Prices prices ON prices.book = books.id
92 WHERE prices.Price <= 5.00
94 all without needing multiple fetches.
96 Only the helper methods for setting up standard relationship types
97 are documented here. For the basic, lower-level methods, and a description
98 of all the useful *_related methods that you get for free, see
99 L<DBIx::Class::Relationship::Base>.
103 All helper methods are called similar to the following template:
105 __PACKAGE__->$method_name('relname', 'Foreign::Class', $cond, $attrs);
107 Both C<$cond> and C<$attrs> are optional. Pass C<undef> for C<$cond> if
108 you want to use the default value for it, but still want to set C<$attrs>.
110 See L<DBIx::Class::Relationship::Base> for documentation on the
111 attrubutes that are allowed in the C<$attrs> argument.
118 =item Arguments: $accessor_name, $related_class, $our_fk_column|\%cond|\@cond?, \%attr?
122 Creates a relationship where the calling class stores the foreign
123 class's primary key in one (or more) of the calling class columns.
124 This relationship defaults to using C<$accessor_name> as the column
125 name in this class to resolve the join against the primary key from
126 C<$related_class>, unless C<$our_fk_column> specifies the foreign key column
127 in this class or C<cond> specifies a reference to a join condition hash.
133 This argument is the name of the method you can call on a
134 L<DBIx::Class::Row> object to retrieve the instance of the foreign
135 class matching this relationship. This is often called the
136 C<relation(ship) name>.
138 Use this accessor_name in L<DBIx::Class::ResultSet/join>
139 or L<DBIx::Class::ResultSet/prefetch> to join to the foreign table
140 indicated by this relationship.
144 This is the class name of the table referenced by the foreign key in
149 The column name on this class that contains the foreign key.
155 A hashref where the keys are C<foreign.$column_on_related_table> and
156 the values are C<self.$our_fk_column>. This is useful for
157 relations that are across multiple columns.
162 # in a Book class (where Author has many Books)
163 My::DBIC::Schema::Book->belongs_to(
165 'My::DBIC::Schema::Author',
170 My::DBIC::Schema::Book->belongs_to(
172 'My::DBIC::Schema::Author',
173 { 'foreign.author_id' => 'self.author_id' }
176 # OR (similar result but uglier accessor name)
177 My::DBIC::Schema::Book->belongs_to(
179 'My::DBIC::Schema::Author'
183 my $author_obj = $book->author; # get author object
184 $book->author( $new_author_obj ); # set author object
185 $book->author_id(); # get the plain id
187 # To retrieve the plain id if you used the ugly version:
188 $book->get_column('author_id');
191 If the relationship is optional -- i.e. the column containing the foreign key
192 can be NULL -- then the belongs_to relationship does the right thing. Thus, in
193 the example above C<$obj-E<gt>author> would return C<undef>. However in this
194 case you would probably want to set the C<join_type> attribute so that a C<LEFT
195 JOIN> is done, which makes complex resultsets involving C<join> or C<prefetch>
196 operations work correctly. The modified declaration is shown below:
198 # in a Book class (where Author has_many Books)
199 __PACKAGE__->belongs_to(
201 'My::DBIC::Schema::Author',
203 { join_type => 'left' }
207 Cascading deletes are off by default on a C<belongs_to>
208 relationship. To turn them on, pass C<< cascade_delete => 1 >>
209 in the $attr hashref.
211 By default, DBIC will return undef and avoid querying the database if a
212 C<belongs_to> accessor is called when any part of the foreign key IS NULL. To
213 disable this behavior, pass C<< undef_on_null_fk => 0 >> in the C<$attr>
216 NOTE: If you are used to L<Class::DBI> relationships, this is the equivalent
219 See L<DBIx::Class::Relationship::Base> for documentation on relationship
220 methods and valid relationship attributes. Also see L<DBIx::Class::ResultSet>
221 for a L<list of standard resultset attributes|DBIx::Class::ResultSet/ATTRIBUTES>
222 which can be assigned to relationships as well.
228 =item Arguments: $accessor_name, $related_class, $their_fk_column|\%cond|\@cond?, \%attr?
232 Creates a one-to-many relationship, where the corresponding elements
233 of the foreign class store the calling class's primary key in one (or
234 more) of the foreign class columns. This relationship defaults to using
235 the end of this classes namespace as the foreign key in C<$related_class>
236 to resolve the join, unless C<$their_fk_column> specifies the foreign
237 key column in C<$related_class> or C<cond> specifies a reference to a
244 This argument is the name of the method you can call on a
245 L<DBIx::Class::Row> object to retrieve a resultset of the related
246 class restricted to the ones related to the row object. In list
247 context it returns the row objects. This is often called the
248 C<relation(ship) name>.
250 Use this accessor_name in L<DBIx::Class::ResultSet/join>
251 or L<DBIx::Class::ResultSet/prefetch> to join to the foreign table
252 indicated by this relationship.
256 This is the class name of the table which contains a foreign key
257 column containing PK values of this class.
259 =item their_fk_column
261 The column name on the related class that contains the foreign key.
267 A hashref where the keys are C<foreign.$their_fk_column> and
268 the values are C<self.$matching_column>. This is useful for
269 relations that are across multiple columns.
273 An arrayref containing an SQL::Abstract-like condition. For example a
274 link table where two columns link back to the same table. This is an
277 My::Schema::Item->has_many('rels', 'My::Schema::Relationships',
278 [ { 'foreign.LItemID' => 'self.ID' },
279 { 'foreign.RItemID' => 'self.ID'} ]);
283 # in an Author class (where Author has_many Books)
284 # assuming related class is storing our PK in "author_id"
285 My::DBIC::Schema::Author->has_many(
287 'My::DBIC::Schema::Book',
292 My::DBIC::Schema::Author->has_many(
294 'My::DBIC::Schema::Book',
295 { 'foreign.author_id' => 'self.id' },
298 # OR (similar result, assuming related_class is storing our PK, in "author")
299 # (the "author" is guessed at from "Author" in the class namespace)
300 My::DBIC::Schema::Author->has_many(
302 'My::DBIC::Schema::Book',
307 # resultset of Books belonging to author
308 my $booklist = $author->books;
310 # resultset of Books belonging to author, restricted by author name
311 my $booklist = $author->books({
312 name => { LIKE => '%macaroni%' },
313 { prefetch => [qw/book/],
316 # array of Book objects belonging to author
317 my @book_objs = $author->books;
319 # force resultset even in list context
320 my $books_rs = $author->books;
321 ( $books_rs ) = $obj->books_rs;
323 # create a new book for this author, the relation fields are auto-filled
324 $author->create_related('books', \%col_data);
325 # alternative method for the above
326 $author->add_to_books(\%col_data);
329 Three methods are created when you create a has_many relationship. The first
330 method is the expected accessor method, C<$accessor_name()>. The second is
331 almost exactly the same as the accessor method but "_rs" is added to the end of
332 the method name. This method works just like the normal accessor, except that
333 it always returns a resultset, even in list context. The third method,
334 named C<< add_to_$relname >>, will also be added to your Row items; this
335 allows you to insert new related items, using the same mechanism as in
336 L<DBIx::Class::Relationship::Base/"create_related">.
338 If you delete an object in a class with a C<has_many> relationship, all
339 the related objects will be deleted as well. To turn this behaviour off,
340 pass C<< cascade_delete => 0 >> in the C<$attr> hashref.
342 The cascaded operations are performed after the requested delete or
343 update, so if your database has a constraint on the relationship, it
344 will have deleted/updated the related records or raised an exception
345 before DBIx::Class gets to perform the cascaded operation.
347 If you copy an object in a class with a C<has_many> relationship, all
348 the related objects will be copied as well. To turn this behaviour off,
349 pass C<< cascade_copy => 0 >> in the C<$attr> hashref. The behaviour
350 defaults to C<< cascade_copy => 1 >>.
352 See L<DBIx::Class::Relationship::Base> for documentation on relationship
353 methods and valid relationship attributes. Also see L<DBIx::Class::ResultSet>
354 for a L<list of standard resultset attributes|DBIx::Class::ResultSet/ATTRIBUTES>
355 which can be assigned to relationships as well.
361 =item Arguments: $accessor_name, $related_class, $their_fk_column|\%cond|\@cond?, \%attr?
365 Creates an optional one-to-one relationship with a class. This relationship
366 defaults to using C<$accessor_name> as the foreign key in C<$related_class> to
367 resolve the join, unless C<$their_fk_column> specifies the foreign key
368 column in C<$related_class> or C<cond> specifies a reference to a join
375 This argument is the name of the method you can call on a
376 L<DBIx::Class::Row> object to retrieve the instance of the foreign
377 class matching this relationship. This is often called the
378 C<relation(ship) name>.
380 Use this accessor_name in L<DBIx::Class::ResultSet/join>
381 or L<DBIx::Class::ResultSet/prefetch> to join to the foreign table
382 indicated by this relationship.
386 This is the class name of the table which contains a foreign key
387 column containing PK values of this class.
389 =item their_fk_column
391 The column name on the related class that contains the foreign key.
397 A hashref where the keys are C<foreign.$their_fk_column> and
398 the values are C<self.$matching_column>. This is useful for
399 relations that are across multiple columns.
403 # Author may have an entry in the pseudonym table
404 My::DBIC::Schema::Author->might_have(
406 'My::DBIC::Schema::Pseudonym',
410 # OR (same result, assuming the related_class stores our PK)
411 My::DBIC::Schema::Author->might_have(
413 'My::DBIC::Schema::Pseudonym',
417 My::DBIC::Schema::Author->might_have(
419 'My::DBIC::Schema::Pseudonym',
420 { 'foreign.author_id' => 'self.id' },
424 my $pname = $author->pseudonym; # to get the Pseudonym object
426 If you update or delete an object in a class with a C<might_have>
427 relationship, the related object will be updated or deleted as well. To
428 turn off this behavior, add C<< cascade_delete => 0 >> to the C<$attr>
431 The cascaded operations are performed after the requested delete or
432 update, so if your database has a constraint on the relationship, it
433 will have deleted/updated the related records or raised an exception
434 before DBIx::Class gets to perform the cascaded operation.
436 See L<DBIx::Class::Relationship::Base> for documentation on relationship
437 methods and valid relationship attributes. Also see L<DBIx::Class::ResultSet>
438 for a L<list of standard resultset attributes|DBIx::Class::ResultSet/ATTRIBUTES>
439 which can be assigned to relationships as well.
445 =item Arguments: $accessor_name, $related_class, $their_fk_column|\%cond|\@cond?, \%attr?
449 Creates a one-to-one relationship with a class. This relationship
450 defaults to using C<$accessor_name> as the foreign key in C<$related_class> to
451 resolve the join, unless C<$their_fk_column> specifies the foreign key
452 column in C<$related_class> or C<cond> specifies a reference to a join
459 This argument is the name of the method you can call on a
460 L<DBIx::Class::Row> object to retrieve the instance of the foreign
461 class matching this relationship. This is often called the
462 C<relation(ship) name>.
464 Use this accessor_name in L<DBIx::Class::ResultSet/join>
465 or L<DBIx::Class::ResultSet/prefetch> to join to the foreign table
466 indicated by this relationship.
470 This is the class name of the table which contains a foreign key
471 column containing PK values of this class.
473 =item their_fk_column
475 The column name on the related class that contains the foreign key.
481 A hashref where the keys are C<foreign.$their_fk_column> and
482 the values are C<self.$matching_column>. This is useful for
483 relations that are across multiple columns.
487 # Every book has exactly one ISBN
488 My::DBIC::Schema::Book->has_one(
490 'My::DBIC::Schema::ISBN',
494 # OR (same result, assuming related_class stores our PK)
495 My::DBIC::Schema::Book->has_one(
497 'My::DBIC::Schema::ISBN',
501 My::DBIC::Schema::Book->has_one(
503 'My::DBIC::Schema::ISBN',
504 { 'foreign.book_id' => 'self.id' },
508 my $isbn_obj = $book->isbn; # to get the ISBN object
510 Creates a one-to-one relationship with another class. This is just
511 like C<might_have>, except the implication is that the other object is
512 always present. The only difference between C<has_one> and
513 C<might_have> is that C<has_one> uses an (ordinary) inner join,
514 whereas C<might_have> defaults to a left join.
516 The has_one relationship should be used when a row in the table has exactly one
517 related row in another table. If the related row might not exist in the foreign
518 table, use the L<DBIx::Class::Relationship/might_have> relationship.
520 In the above example, each Book in the database is associated with exactly one
523 See L<DBIx::Class::Relationship::Base> for documentation on relationship
524 methods and valid relationship attributes. Also see L<DBIx::Class::ResultSet>
525 for a L<list of standard resultset attributes|DBIx::Class::ResultSet/ATTRIBUTES>
526 which can be assigned to relationships as well.
532 =item Arguments: $accessor_name, $link_rel_name, $foreign_rel_name, \%attr?
536 C<many_to_many> is not strictly a relationship in its own right. Instead, it is
537 a bridge between two resultsets which provide the same kind of convenience
538 accessors as true relationships provide. Although the accessor will return a
539 resultset or collection of objects just like has_many does, you cannot call
540 C<related_resultset> and similar methods which operate on true relationships.
546 This argument is the name of the method you can call on a
547 L<DBIx::Class::Row> object to retrieve the rows matching this
550 On a many_to_many, unlike other relationships, this cannot be used in
551 L<DBIx::Class::ResultSet/search> to join tables. Use the relations
552 bridged across instead.
556 This is the accessor_name from the has_many relationship we are
559 =item foreign_rel_name
561 This is the accessor_name of the belongs_to relationship in the link
562 table that we are bridging across (which gives us the table we are
567 To create a many_to_many relationship from Actor to Role:
569 My::DBIC::Schema::Actor->has_many( actor_roles =>
570 'My::DBIC::Schema::ActorRoles',
572 My::DBIC::Schema::ActorRoles->belongs_to( role =>
573 'My::DBIC::Schema::Role' );
574 My::DBIC::Schema::ActorRoles->belongs_to( actor =>
575 'My::DBIC::Schema::Actor' );
577 My::DBIC::Schema::Actor->many_to_many( roles => 'actor_roles',
580 And, for the reverse relationship, from Role to Actor:
582 My::DBIC::Schema::Role->has_many( actor_roles =>
583 'My::DBIC::Schema::ActorRoles',
586 My::DBIC::Schema::Role->many_to_many( actors => 'actor_roles', 'actor' );
588 To add a role for your actor, and fill in the year of the role in the
591 $actor->add_to_roles($role, { year => 1995 });
593 In the above example, ActorRoles is the link table class, and Role is the
594 foreign class. The C<$link_rel_name> parameter is the name of the accessor for
595 the has_many relationship from this table to the link table, and the
596 C<$foreign_rel_name> parameter is the accessor for the belongs_to relationship
597 from the link table to the foreign table.
599 To use many_to_many, existing relationships from the original table to the link
600 table, and from the link table to the end table must already exist, these
601 relation names are then used in the many_to_many call.
603 In the above example, the Actor class will have 3 many_to_many accessor methods
604 set: C<roles>, C<add_to_roles>, C<set_roles>, and similarly named accessors
605 will be created for the Role class for the C<actors> many_to_many
608 See L<DBIx::Class::Relationship::Base> for documentation on relationship
609 methods and valid relationship attributes. Also see L<DBIx::Class::ResultSet>
610 for a L<list of standard resultset attributes|DBIx::Class::ResultSet/ATTRIBUTES>
611 which can be assigned to relationships as well.
623 You may distribute this code under the same terms as Perl itself.