1 package DBIx::Class::ResultSet;
5 use base qw/DBIx::Class/;
7 use DBIx::Class::ResultSetColumn;
8 use DBIx::Class::ResultClass::HashRefInflator;
9 use Scalar::Util qw/blessed weaken reftype/;
10 use DBIx::Class::_Util qw(
11 dbic_internal_try dump_value
12 fail_on_internal_wantarray fail_on_internal_call UNRESOLVABLE_CONDITION
17 # De-duplication in _merge_attr() is disabled, but left in for reference
18 # (the merger is used for other things that ought not to be de-duped)
19 *__HM_DEDUP = sub () { 0 };
29 # this is real - CDBICompat overrides it with insanity
30 # yes, prototype won't matter, but that's for now ;)
33 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class result_source/);
37 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
41 my $users_rs = $schema->resultset('User');
42 while( $user = $users_rs->next) {
43 print $user->username;
46 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
47 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
51 A ResultSet is an object which stores a set of conditions representing
52 a query. It is the backbone of DBIx::Class (i.e. the really
53 important/useful bit).
55 No SQL is executed on the database when a ResultSet is created, it
56 just stores all the conditions needed to create the query.
58 A basic ResultSet representing the data of an entire table is returned
59 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
60 L<Source|DBIx::Class::Manual::Glossary/ResultSource> name.
62 my $users_rs = $schema->resultset('User');
64 A new ResultSet is returned from calling L</search> on an existing
65 ResultSet. The new one will contain all the conditions of the
66 original, plus any new conditions added in the C<search> call.
68 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
69 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
72 The query that the ResultSet represents is B<only> executed against
73 the database when these methods are called:
74 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
76 If a resultset is used in a numeric context it returns the L</count>.
77 However, if it is used in a boolean context it is B<always> true. So if
78 you want to check if a resultset has any results, you must use C<if $rs
83 =head2 Chaining resultsets
85 Let's say you've got a query that needs to be run to return some data
86 to the user. But, you have an authorization system in place that
87 prevents certain users from seeing certain information. So, you want
88 to construct the basic query in one method, but add constraints to it in
93 my $request = $self->get_request; # Get a request object somehow.
94 my $schema = $self->result_source->schema;
96 my $cd_rs = $schema->resultset('CD')->search({
97 title => $request->param('title'),
98 year => $request->param('year'),
101 $cd_rs = $self->apply_security_policy( $cd_rs );
103 return $cd_rs->all();
106 sub apply_security_policy {
115 =head3 Resolving conditions and attributes
117 When a resultset is chained from another resultset (e.g.:
118 C<< my $new_rs = $old_rs->search(\%extra_cond, \%attrs) >>), conditions
119 and attributes with the same keys need resolving.
121 If any of L</columns>, L</select>, L</as> are present, they reset the
122 original selection, and start the selection "clean".
124 The L</join>, L</prefetch>, L</+columns>, L</+select>, L</+as> attributes
125 are merged into the existing ones from the original resultset.
127 The L</where> and L</having> attributes, and any search conditions, are
128 merged with an SQL C<AND> to the existing condition from the original
131 All other attributes are overridden by any new ones supplied in the
134 =head2 Multiple queries
136 Since a resultset just defines a query, you can do all sorts of
137 things with it with the same object.
139 # Don't hit the DB yet.
140 my $cd_rs = $schema->resultset('CD')->search({
141 title => 'something',
145 # Each of these hits the DB individually.
146 my $count = $cd_rs->count;
147 my $most_recent = $cd_rs->get_column('date_released')->max();
148 my @records = $cd_rs->all;
150 And it's not just limited to SELECT statements.
156 $cd_rs->create({ artist => 'Fred' });
158 Which is the same as:
160 $schema->resultset('CD')->create({
161 title => 'something',
166 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
168 =head2 Custom ResultSet classes
170 To add methods to your resultsets, you can subclass L<DBIx::Class::ResultSet>, similar to:
172 package MyApp::Schema::ResultSet::User;
177 use base 'DBIx::Class::ResultSet';
181 $self->search({ $self->current_source_alias . '.active' => 1 });
186 $self->search({ $self->current_source_alias . '.verified' => 0 });
189 sub created_n_days_ago {
190 my ($self, $days_ago) = @_;
192 $self->current_source_alias . '.create_date' => {
194 $self->result_source->schema->storage->datetime_parser->format_datetime(
195 DateTime->now( time_zone => 'UTC' )->subtract( days => $days_ago )
200 sub users_to_warn { shift->active->unverified->created_n_days_ago(7) }
204 See L<DBIx::Class::Schema/load_namespaces> on how DBIC can discover and
205 automatically attach L<Result|DBIx::Class::Manual::ResultClass>-specific
206 L<ResulSet|DBIx::Class::ResultSet> classes.
208 =head3 ResultSet subclassing with Moose and similar constructor-providers
210 Using L<Moose> or L<Moo> in your ResultSet classes is usually overkill, but
211 you may find it useful if your ResultSets contain a lot of business logic
212 (e.g. C<has xml_parser>, C<has json>, etc) or if you just prefer to organize
215 In order to write custom ResultSet classes with L<Moo> you need to use the
216 following template. The L<BUILDARGS|Moo/BUILDARGS> is necessary due to the
217 unusual signature of the L<constructor provided by DBIC
218 |DBIx::Class::ResultSet/new> C<< ->new($source, \%args) >>.
221 extends 'DBIx::Class::ResultSet';
222 sub BUILDARGS { $_[2] } # ::RS::new() expects my ($class, $rsrc, $args) = @_
228 If you want to build your custom ResultSet classes with L<Moose>, you need
229 a similar, though a little more elaborate template in order to interface the
230 inlining of the L<Moose>-provided
231 L<object constructor|Moose::Manual::Construction/WHERE'S THE CONSTRUCTOR?>,
234 package MyApp::Schema::ResultSet::User;
237 use MooseX::NonMoose;
238 extends 'DBIx::Class::ResultSet';
240 sub BUILDARGS { $_[2] } # ::RS::new() expects my ($class, $rsrc, $args) = @_
244 __PACKAGE__->meta->make_immutable;
248 The L<MooseX::NonMoose> is necessary so that the L<Moose> constructor does not
249 entirely overwrite the DBIC one (in contrast L<Moo> does this automatically).
250 Alternatively, you can skip L<MooseX::NonMoose> and get by with just L<Moose>
253 __PACKAGE__->meta->make_immutable(inline_constructor => 0);
261 =item Arguments: L<$source|DBIx::Class::ResultSource>, L<\%attrs?|/ATTRIBUTES>
263 =item Return Value: L<$resultset|/search>
267 The resultset constructor. Takes a source object (usually a
268 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
269 L</ATTRIBUTES> below). Does not perform any queries -- these are
270 executed as needed by the other methods.
272 Generally you never construct a resultset manually. Instead you get one
274 C<< $schema->L<resultset|DBIx::Class::Schema/resultset>('$source_name') >>
275 or C<< $another_resultset->L<search|/search>(...) >> (the later called in
278 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
284 If called on an object, proxies to L</new_result> instead, so
286 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
288 will return a CD object, not a ResultSet, and is equivalent to:
290 my $cd = $schema->resultset('CD')->new_result({ title => 'Spoon' });
292 Please also keep in mind that many internals call L</new_result> directly,
293 so overloading this method with the idea of intercepting new result object
294 creation B<will not work>. See also warning pertaining to L</create>.
304 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
305 return $class->new_result(@_);
308 my ($source, $attrs) = @_;
309 $source = $source->resolve
310 if $source->isa('DBIx::Class::ResultSourceHandle');
312 $attrs = { %{$attrs||{}} };
313 delete @{$attrs}{qw(_last_sqlmaker_alias_map _simple_passthrough_construction)};
315 if ($attrs->{page}) {
316 $attrs->{rows} ||= 10;
319 $attrs->{alias} ||= 'me';
322 result_source => $source,
323 cond => $attrs->{where},
328 # if there is a dark selector, this means we are already in a
329 # chain and the cleanup/sanification was taken care of by
331 $self->_normalize_selection($attrs)
332 unless $attrs->{_dark_selector};
335 $attrs->{result_class} || $source->result_class
345 =item Arguments: L<$cond|DBIx::Class::SQLMaker> | undef, L<\%attrs?|/ATTRIBUTES>
347 =item Return Value: $resultset (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
351 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
352 my $new_rs = $cd_rs->search({ year => 2005 });
354 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
355 # year = 2005 OR year = 2004
357 In list context, C<< ->all() >> is called implicitly on the resultset, thus
358 returning a list of L<result|DBIx::Class::Manual::ResultClass> objects instead.
359 To avoid that, use L</search_rs>.
361 If you need to pass in additional attributes but no additional condition,
362 call it as C<search(undef, \%attrs)>.
364 # "SELECT name, artistid FROM $artist_table"
365 my @all_artists = $schema->resultset('Artist')->search(undef, {
366 columns => [qw/name artistid/],
369 For a list of attributes that can be passed to C<search>, see
370 L</ATTRIBUTES>. For more examples of using this function, see
371 L<Searching|DBIx::Class::Manual::Cookbook/SEARCHING>. For a complete
372 documentation for the first argument, see L<SQL::Abstract/"WHERE CLAUSES">
373 and its extension L<DBIx::Class::SQLMaker>.
375 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
379 Note that L</search> does not process/deflate any of the values passed in the
380 L<SQL::Abstract>-compatible search condition structure. This is unlike other
381 condition-bound methods L</new_result>, L</create> and L</find>. The user must ensure
382 manually that any value passed to this method will stringify to something the
383 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
384 objects, for more info see:
385 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
391 my $rs = $self->search_rs( @_ );
394 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_WANTARRAY and my $sog = fail_on_internal_wantarray;
397 elsif (defined wantarray) {
401 # we can be called by a relationship helper, which in
402 # turn may be called in void context due to some braindead
403 # overload or whatever else the user decided to be clever
404 # at this particular day. Thus limit the exception to
405 # external code calls only
406 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense')
407 if (caller)[0] !~ /^\QDBIx::Class::/;
417 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
419 =item Return Value: L<$resultset|/search>
423 This method does the same exact thing as search() except it will
424 always return a resultset, even in list context.
431 my $rsrc = $self->result_source;
432 my ($call_cond, $call_attrs);
434 # Special-case handling for (undef, undef) or (undef)
435 # Note that (foo => undef) is valid deprecated syntax
436 @_ = () if not scalar grep { defined $_ } @_;
442 # fish out attrs in the ($condref, $attr) case
443 elsif (@_ == 2 and ( ! defined $_[0] or length ref $_[0] ) ) {
444 ($call_cond, $call_attrs) = @_;
447 $self->throw_exception('Odd number of arguments to search')
451 carp_unique 'search( %condition ) is deprecated, use search( \%condition ) instead'
452 unless $rsrc->result_class->isa('DBIx::Class::CDBICompat');
454 for my $i (0 .. $#_) {
456 $self->throw_exception ('All keys in condition key/value pairs must be plain scalars')
457 if (! defined $_[$i] or length ref $_[$i] );
463 # see if we can keep the cache (no $rs changes)
465 my %safe = (alias => 1, cache => 1);
466 if ( ! grep { !$safe{$_} } keys %$call_attrs and (
469 ref $call_cond eq 'HASH' && ! keys %$call_cond
471 ref $call_cond eq 'ARRAY' && ! @$call_cond
473 $cache = $self->get_cache;
476 my $old_attrs = { %{$self->{attrs}} };
477 my ($old_having, $old_where) = delete @{$old_attrs}{qw(having where)};
479 my $new_attrs = { %$old_attrs };
481 # take care of call attrs (only if anything is changing)
482 if ($call_attrs and keys %$call_attrs) {
484 # copy for _normalize_selection
485 $call_attrs = { %$call_attrs };
487 my @selector_attrs = qw/select as columns cols +select +as +columns include_columns/;
489 # reset the current selector list if new selectors are supplied
490 delete @{$old_attrs}{(@selector_attrs, '_dark_selector')}
491 if grep { exists $call_attrs->{$_} } qw(columns cols select as);
493 # Normalize the new selector list (operates on the passed-in attr structure)
494 # Need to do it on every chain instead of only once on _resolved_attrs, in
495 # order to allow detection of empty vs partial 'as'
496 $call_attrs->{_dark_selector} = $old_attrs->{_dark_selector}
497 if $old_attrs->{_dark_selector};
498 $self->_normalize_selection ($call_attrs);
500 # start with blind overwriting merge, exclude selector attrs
501 $new_attrs = { %{$old_attrs}, %{$call_attrs} };
502 delete @{$new_attrs}{@selector_attrs};
504 for (@selector_attrs) {
505 $new_attrs->{$_} = $self->_merge_attr($old_attrs->{$_}, $call_attrs->{$_})
506 if ( exists $old_attrs->{$_} or exists $call_attrs->{$_} );
509 # older deprecated name, use only if {columns} is not there
510 if (my $c = delete $new_attrs->{cols}) {
511 carp_unique( "Resultset attribute 'cols' is deprecated, use 'columns' instead" );
512 if ($new_attrs->{columns}) {
513 carp "Resultset specifies both the 'columns' and the legacy 'cols' attributes - ignoring 'cols'";
516 $new_attrs->{columns} = $c;
521 # join/prefetch use their own crazy merging heuristics
522 foreach my $key (qw/join prefetch/) {
523 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key})
524 if exists $call_attrs->{$key};
527 # stack binds together
528 $new_attrs->{bind} = [ @{ $old_attrs->{bind} || [] }, @{ $call_attrs->{bind} || [] } ];
532 for ($old_where, $call_cond) {
534 $new_attrs->{where} = $self->_stack_cond (
535 $_, $new_attrs->{where}
540 if (defined $old_having) {
541 $new_attrs->{having} = $self->_stack_cond (
542 $old_having, $new_attrs->{having}
546 my $rs = (ref $self)->new($rsrc, $new_attrs);
548 $rs->set_cache($cache) if ($cache);
553 sub _normalize_selection {
554 my ($self, $attrs) = @_;
557 if ( exists $attrs->{include_columns} ) {
558 carp_unique( "Resultset attribute 'include_columns' is deprecated, use '+columns' instead" );
559 $attrs->{'+columns'} = $self->_merge_attr(
560 $attrs->{'+columns'}, delete $attrs->{include_columns}
564 # columns are always placed first, however
566 # Keep the X vs +X separation until _resolved_attrs time - this allows to
567 # delay the decision on whether to use a default select list ($rsrc->columns)
568 # allowing stuff like the remove_columns helper to work
570 # select/as +select/+as pairs need special handling - the amount of select/as
571 # elements in each pair does *not* have to be equal (think multicolumn
572 # selectors like distinct(foo, bar) ). If the selector is bare (no 'as'
573 # supplied at all) - try to infer the alias, either from the -as parameter
574 # of the selector spec, or use the parameter whole if it looks like a column
575 # name (ugly legacy heuristic). If all fails - leave the selector bare (which
576 # is ok as well), but make sure no more additions to the 'as' chain take place
577 for my $pref ('', '+') {
579 my ($sel, $as) = map {
580 my $key = "${pref}${_}";
582 my $val = [ ref $attrs->{$key} eq 'ARRAY'
584 : $attrs->{$key} || ()
586 delete $attrs->{$key};
590 if (! @$as and ! @$sel ) {
593 elsif (@$as and ! @$sel) {
594 $self->throw_exception(
595 "Unable to handle ${pref}as specification (@$as) without a corresponding ${pref}select"
599 # no as part supplied at all - try to deduce (unless explicit end of named selection is declared)
600 # if any @$as has been supplied we assume the user knows what (s)he is doing
601 # and blindly keep stacking up pieces
602 unless ($attrs->{_dark_selector}) {
605 if ( ref $_ eq 'HASH' and exists $_->{-as} ) {
606 push @$as, $_->{-as};
608 # assume any plain no-space, no-parenthesis string to be a column spec
609 # FIXME - this is retarded but is necessary to support shit like 'count(foo)'
610 elsif ( ! ref $_ and $_ =~ /^ [^\s\(\)]+ $/x) {
613 # if all else fails - raise a flag that no more aliasing will be allowed
615 $attrs->{_dark_selector} = {
618 local $Data::Dumper::Indent = 0;
627 elsif (@$as < @$sel) {
628 $self->throw_exception(
629 "Unable to handle an ${pref}as specification (@$as) with less elements than the corresponding ${pref}select"
632 elsif ($pref and $attrs->{_dark_selector}) {
633 $self->throw_exception(
634 "Unable to process named '+select', resultset contains an unnamed selector $attrs->{_dark_selector}{string}"
640 $attrs->{"${pref}select"} = $self->_merge_attr($attrs->{"${pref}select"}, $sel);
641 $attrs->{"${pref}as"} = $self->_merge_attr($attrs->{"${pref}as"}, $as);
646 my ($self, $left, $right) = @_;
649 (ref $_ eq 'ARRAY' and !@$_)
651 (ref $_ eq 'HASH' and ! keys %$_)
652 ) and $_ = undef for ($left, $right);
654 # either one of the two undef
655 if ( (defined $left) xor (defined $right) ) {
656 return defined $left ? $left : $right;
659 elsif ( ! defined $left ) {
663 return $self->result_source->schema->storage->_collapse_cond({ -and => [$left, $right] });
667 =head2 search_literal
669 B<CAVEAT>: C<search_literal> is provided for Class::DBI compatibility and
670 should only be used in that context. C<search_literal> is a convenience
671 method. It is equivalent to calling C<< $schema->search(\[]) >>, but if you
672 want to ensure columns are bound correctly, use L</search>.
674 See L<DBIx::Class::Manual::Cookbook/SEARCHING> and
675 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
676 require C<search_literal>.
680 =item Arguments: $sql_fragment, @standalone_bind_values
682 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
686 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
687 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
689 Pass a literal chunk of SQL to be added to the conditional part of the
692 Example of how to use C<search> instead of C<search_literal>
694 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
695 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
700 my ($self, $sql, @bind) = @_;
702 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
705 return $self->search(\[ $sql, map [ {} => $_ ], @bind ], ($attr || () ));
712 =item Arguments: \%columns_values | @pk_values, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
714 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
718 Finds and returns a single row based on supplied criteria. Takes either a
719 hashref with the same format as L</create> (including inference of foreign
720 keys from related objects), or a list of primary key values in the same
721 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
722 declaration on the L</result_source>.
724 In either case an attempt is made to combine conditions already existing on
725 the resultset with the condition passed to this method.
727 To aid with preparing the correct query for the storage you may supply the
728 C<key> attribute, which is the name of a
729 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
730 unique constraint corresponding to the
731 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
732 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
733 to construct a query that satisfies the named unique constraint fully (
734 non-NULL values for each column member of the constraint) an exception is
737 If no C<key> is specified, the search is carried over all unique constraints
738 which are fully defined by the available condition.
740 If no such constraint is found, C<find> currently defaults to a simple
741 C<< search->(\%column_values) >> which may or may not do what you expect.
742 Note that this fallback behavior may be deprecated in further versions. If
743 you need to search with arbitrary conditions - use L</search>. If the query
744 resulting from this fallback produces more than one row, a warning to the
745 effect is issued, though only the first row is constructed and returned as
748 In addition to C<key>, L</find> recognizes and applies standard
749 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
751 Note that if you have extra concerns about the correctness of the resulting
752 query you need to specify the C<key> attribute and supply the entire condition
753 as an argument to find (since it is not always possible to perform the
754 combination of the resultset condition with the supplied one, especially if
755 the resultset condition contains literal sql).
757 For example, to find a row by its primary key:
759 my $cd = $schema->resultset('CD')->find(5);
761 You can also find a row by a specific unique constraint:
763 my $cd = $schema->resultset('CD')->find(
765 artist => 'Massive Attack',
766 title => 'Mezzanine',
768 { key => 'cd_artist_title' }
771 See also L</find_or_create> and L</update_or_create>.
777 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
779 my $rsrc = $self->result_source;
782 if (exists $attrs->{key}) {
783 $constraint_name = defined $attrs->{key}
785 : $self->throw_exception("An undefined 'key' resultset attribute makes no sense")
789 # Parse out the condition from input
792 if (ref $_[0] eq 'HASH') {
793 $call_cond = { %{$_[0]} };
796 # if only values are supplied we need to default to 'primary'
797 $constraint_name = 'primary' unless defined $constraint_name;
799 my @c_cols = $rsrc->unique_constraint_columns($constraint_name);
801 $self->throw_exception(
802 "No constraint columns, maybe a malformed '$constraint_name' constraint?"
805 $self->throw_exception (
806 'find() expects either a column/value hashref, or a list of values '
807 . "corresponding to the columns of the specified unique constraint '$constraint_name'"
808 ) unless @c_cols == @_;
810 @{$call_cond}{@c_cols} = @_;
813 # process relationship data if any
814 for my $key (keys %$call_cond) {
816 length ref($call_cond->{$key})
818 my $relinfo = $rsrc->relationship_info($key)
820 # implicitly skip has_many's (likely MC)
821 (ref (my $val = delete $call_cond->{$key}) ne 'ARRAY' )
823 my ($rel_cond, $crosstable) = $rsrc->_resolve_condition(
824 $relinfo->{cond}, $val, $key, $key
827 $self->throw_exception("Complex condition via relationship '$key' is unsupported in find()")
828 if $crosstable or ref($rel_cond) ne 'HASH';
830 # supplement condition
831 # relationship conditions take precedence (?)
832 @{$call_cond}{keys %$rel_cond} = values %$rel_cond;
836 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
838 if (defined $constraint_name) {
839 $final_cond = $self->_qualify_cond_columns (
841 $self->result_source->_minimal_valueset_satisfying_constraint(
842 constraint_name => $constraint_name,
843 values => ($self->_merge_with_rscond($call_cond))[0],
850 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
851 # This means that we got here after a merger of relationship conditions
852 # in ::Relationship::Base::search_related (the row method), and furthermore
853 # the relationship is of the 'single' type. This means that the condition
854 # provided by the relationship (already attached to $self) is sufficient,
855 # as there can be only one row in the database that would satisfy the
859 my (@unique_queries, %seen_column_combinations, $ci, @fc_exceptions);
861 # no key was specified - fall down to heuristics mode:
862 # run through all unique queries registered on the resultset, and
863 # 'OR' all qualifying queries together
865 # always start from 'primary' if it exists at all
866 for my $c_name ( sort {
868 : $b eq 'primary' ? 1
870 } $rsrc->unique_constraint_names) {
872 next if $seen_column_combinations{
873 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
877 push @unique_queries, $self->_qualify_cond_columns(
878 $self->result_source->_minimal_valueset_satisfying_constraint(
879 constraint_name => $c_name,
880 values => ($self->_merge_with_rscond($call_cond))[0],
881 columns_info => ($ci ||= $self->result_source->columns_info),
887 push @fc_exceptions, $_ if $_ =~ /\bFilterColumn\b/;
892 @unique_queries ? \@unique_queries
893 : @fc_exceptions ? $self->throw_exception(join "; ", map { $_ =~ /(.*) at .+ line \d+$/s } @fc_exceptions )
894 : $self->_non_unique_find_fallback ($call_cond, $attrs)
898 # Run the query, passing the result_class since it should propagate for find
899 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
900 if ($rs->_resolved_attrs->{collapse}) {
902 carp "Query returned more than one row" if $rs->next;
910 # This is a stop-gap method as agreed during the discussion on find() cleanup:
911 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
913 # It is invoked when find() is called in legacy-mode with insufficiently-unique
914 # condition. It is provided for overrides until a saner way forward is devised
916 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
917 # the road. Please adjust your tests accordingly to catch this situation early
918 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
920 # The method will not be removed without an adequately complete replacement
921 # for strict-mode enforcement
922 sub _non_unique_find_fallback {
923 my ($self, $cond, $attrs) = @_;
925 return $self->_qualify_cond_columns(
927 exists $attrs->{alias}
929 : $self->{attrs}{alias}
934 sub _qualify_cond_columns {
935 my ($self, $cond, $alias) = @_;
937 my %aliased = %$cond;
938 for (keys %aliased) {
939 $aliased{"$alias.$_"} = delete $aliased{$_}
946 sub _build_unique_cond {
948 '_build_unique_cond is a private method, and moreover is about to go '
949 . 'away. Please contact the development team at %s if you believe you '
950 . 'have a genuine use for this method, in order to discuss alternatives.',
951 DBIx::Class::_ENV_::HELP_URL,
954 my ($self, $constraint_name, $cond, $croak_on_null) = @_;
956 $self->result_source->_minimal_valueset_satisfying_constraint(
957 constraint_name => $constraint_name,
959 carp_on_nulls => !$croak_on_null
963 =head2 search_related
967 =item Arguments: $rel_name, $cond?, L<\%attrs?|/ATTRIBUTES>
969 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
973 $new_rs = $cd_rs->search_related('artist', {
977 Searches the specified relationship, optionally specifying a condition and
978 attributes for matching records. See L</ATTRIBUTES> for more information.
980 In list context, C<< ->all() >> is called implicitly on the resultset, thus
981 returning a list of result objects instead. To avoid that, use L</search_related_rs>.
983 See also L</search_related_rs>.
988 return shift->related_resultset(shift)->search(@_);
991 =head2 search_related_rs
993 This method works exactly the same as search_related, except that
994 it guarantees a resultset, even in list context.
998 sub search_related_rs {
999 return shift->related_resultset(shift)->search_rs(@_);
1006 =item Arguments: none
1008 =item Return Value: L<$cursor|DBIx::Class::Cursor>
1012 Returns a storage-driven cursor to the given resultset. See
1013 L<DBIx::Class::Cursor> for more information.
1020 return $self->{cursor} ||= do {
1021 my $attrs = $self->_resolved_attrs;
1022 $self->result_source->storage->select(
1023 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
1032 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1034 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1038 my $cd = $schema->resultset('CD')->single({ year => 2001 });
1040 Inflates the first result without creating a cursor if the resultset has
1041 any records in it; if not returns C<undef>. Used by L</find> as a lean version
1044 While this method can take an optional search condition (just like L</search>)
1045 being a fast-code-path it does not recognize search attributes. If you need to
1046 add extra joins or similar, call L</search> and then chain-call L</single> on the
1047 L<DBIx::Class::ResultSet> returned.
1053 As of 0.08100, this method enforces the assumption that the preceding
1054 query returns only one row. If more than one row is returned, you will receive
1057 Query returned more than one row
1059 In this case, you should be using L</next> or L</find> instead, or if you really
1060 know what you are doing, use the L</rows> attribute to explicitly limit the size
1063 This method will also throw an exception if it is called on a resultset prefetching
1064 has_many, as such a prefetch implies fetching multiple rows from the database in
1065 order to assemble the resulting object.
1072 my ($self, $where) = @_;
1074 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
1077 my $attrs = { %{$self->_resolved_attrs} };
1079 $self->throw_exception(
1080 'single() can not be used on resultsets collapsing a has_many. Use find( \%cond ) or next() instead'
1081 ) if $attrs->{collapse};
1084 if (defined $attrs->{where}) {
1087 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
1088 $where, delete $attrs->{where} ]
1091 $attrs->{where} = $where;
1095 my $data = [ $self->result_source->storage->select_single(
1096 $attrs->{from}, $attrs->{select},
1097 $attrs->{where}, $attrs
1100 return undef unless @$data;
1101 $self->{_stashed_rows} = [ $data ];
1102 $self->_construct_results->[0];
1109 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1111 =item Return Value: L<$resultsetcolumn|DBIx::Class::ResultSetColumn>
1115 my $max_length = $rs->get_column('length')->max;
1117 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
1122 my ($self, $column) = @_;
1123 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
1131 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1133 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1137 # WHERE title LIKE '%blue%'
1138 $cd_rs = $rs->search_like({ title => '%blue%'});
1140 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
1141 that this is simply a convenience method retained for ex Class::DBI users.
1142 You most likely want to use L</search> with specific operators.
1144 For more information, see L<DBIx::Class::Manual::Cookbook>.
1146 This method is deprecated and will be removed in 0.09. Use L<search()|/search>
1147 instead. An example conversion is:
1149 ->search_like({ foo => 'bar' });
1153 ->search({ foo => { like => 'bar' } });
1160 'search_like() is deprecated and will be removed in DBIC version 0.09.'
1161 .' Instead use ->search({ x => { -like => "y%" } })'
1162 .' (note the outer pair of {}s - they are important!)'
1164 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
1165 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
1166 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
1167 return $class->search($query, { %$attrs });
1174 =item Arguments: $first, $last
1176 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1180 Returns a resultset or object list representing a subset of elements from the
1181 resultset slice is called on. Indexes are from 0, i.e., to get the first
1182 three records, call:
1184 my ($one, $two, $three) = $rs->slice(0, 2);
1189 my ($self, $min, $max) = @_;
1190 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
1191 $attrs->{offset} = $self->{attrs}{offset} || 0;
1192 $attrs->{offset} += $min;
1193 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
1194 return $self->search(undef, $attrs);
1201 =item Arguments: none
1203 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1207 Returns the next element in the resultset (C<undef> is there is none).
1209 Can be used to efficiently iterate over records in the resultset:
1211 my $rs = $schema->resultset('CD')->search;
1212 while (my $cd = $rs->next) {
1216 Note that you need to store the resultset object, and call C<next> on it.
1217 Calling C<< resultset('Table')->next >> repeatedly will always return the
1218 first record from the resultset.
1225 if (my $cache = $self->get_cache) {
1226 $self->{all_cache_position} ||= 0;
1227 return $cache->[$self->{all_cache_position}++];
1230 if ($self->{attrs}{cache}) {
1231 delete $self->{pager};
1232 $self->{all_cache_position} = 1;
1233 return ($self->all)[0];
1236 return shift(@{$self->{_stashed_results}}) if @{ $self->{_stashed_results}||[] };
1238 $self->{_stashed_results} = $self->_construct_results
1241 return shift @{$self->{_stashed_results}};
1244 # Constructs as many results as it can in one pass while respecting
1245 # cursor laziness. Several modes of operation:
1247 # * Always builds everything present in @{$self->{_stashed_rows}}
1248 # * If called with $fetch_all true - pulls everything off the cursor and
1249 # builds all result structures (or objects) in one pass
1250 # * If $self->_resolved_attrs->{collapse} is true, checks the order_by
1251 # and if the resultset is ordered properly by the left side:
1252 # * Fetches stuff off the cursor until the "master object" changes,
1253 # and saves the last extra row (if any) in @{$self->{_stashed_rows}}
1255 # * Just fetches, and collapses/constructs everything as if $fetch_all
1256 # was requested (there is no other way to collapse except for an
1258 # * If no collapse is requested - just get the next row, construct and
1260 sub _construct_results {
1261 my ($self, $fetch_all) = @_;
1263 my $rsrc = $self->result_source;
1264 my $attrs = $self->_resolved_attrs;
1269 ! $attrs->{order_by}
1273 my @pcols = $rsrc->primary_columns
1275 # default order for collapsing unless the user asked for something
1276 $attrs->{order_by} = [ map { join '.', $attrs->{alias}, $_} @pcols ];
1277 $attrs->{_ordered_for_collapse} = 1;
1278 $attrs->{_order_is_artificial} = 1;
1281 # this will be used as both initial raw-row collector AND as a RV of
1282 # _construct_results. Not regrowing the array twice matters a lot...
1283 # a surprising amount actually
1284 my $rows = delete $self->{_stashed_rows};
1286 my $cursor; # we may not need one at all
1288 my $did_fetch_all = $fetch_all;
1291 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1292 $rows = [ ($rows ? @$rows : ()), $self->cursor->all ];
1294 elsif( $attrs->{collapse} ) {
1296 # a cursor will need to be closed over in case of collapse
1297 $cursor = $self->cursor;
1299 $attrs->{_ordered_for_collapse} = (
1305 ->_extract_colinfo_of_stable_main_source_order_by_portion($attrs)
1307 ) unless defined $attrs->{_ordered_for_collapse};
1309 if (! $attrs->{_ordered_for_collapse}) {
1312 # instead of looping over ->next, use ->all in stealth mode
1313 # *without* calling a ->reset afterwards
1314 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1315 if (! $cursor->{_done}) {
1316 $rows = [ ($rows ? @$rows : ()), $cursor->all ];
1317 $cursor->{_done} = 1;
1322 if (! $did_fetch_all and ! @{$rows||[]} ) {
1323 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1324 $cursor ||= $self->cursor;
1325 if (scalar (my @r = $cursor->next) ) {
1330 return undef unless @{$rows||[]};
1332 # sanity check - people are too clever for their own good
1333 if ($attrs->{collapse} and my $aliastypes = $attrs->{_last_sqlmaker_alias_map} ) {
1335 my $multiplied_selectors;
1336 for my $sel_alias ( grep { $_ ne $attrs->{alias} } keys %{ $aliastypes->{selecting} } ) {
1338 $aliastypes->{multiplying}{$sel_alias}
1340 $aliastypes->{premultiplied}{$sel_alias}
1342 $multiplied_selectors->{$_} = 1 for values %{$aliastypes->{selecting}{$sel_alias}{-seen_columns}}
1346 for my $i (0 .. $#{$attrs->{as}} ) {
1347 my $sel = $attrs->{select}[$i];
1349 if (ref $sel eq 'SCALAR') {
1352 elsif( ref $sel eq 'REF' and ref $$sel eq 'ARRAY' ) {
1356 $self->throw_exception(
1357 'Result collapse not possible - selection from a has_many source redirected to the main object'
1358 ) if ($multiplied_selectors->{$sel} and $attrs->{as}[$i] !~ /\./);
1362 # hotspot - skip the setter
1363 my $res_class = $self->_result_class;
1365 my $inflator_cref = $self->{_result_inflator}{cref} ||= do {
1366 $res_class->can ('inflate_result')
1367 or $self->throw_exception("Inflator $res_class does not provide an inflate_result() method");
1370 my $infmap = $attrs->{as};
1372 $self->{_result_inflator}{is_core_row} = ( (
1375 ( \&DBIx::Class::Row::inflate_result || die "No ::Row::inflate_result() - can't happen" )
1376 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_core_row};
1378 $self->{_result_inflator}{is_hri} = ( (
1379 ! $self->{_result_inflator}{is_core_row}
1381 $inflator_cref == \&DBIx::Class::ResultClass::HashRefInflator::inflate_result
1382 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_hri};
1385 if ($attrs->{_simple_passthrough_construction}) {
1386 # construct a much simpler array->hash folder for the one-table HRI cases right here
1387 if ($self->{_result_inflator}{is_hri}) {
1388 for my $r (@$rows) {
1389 $r = { map { $infmap->[$_] => $r->[$_] } 0..$#$infmap };
1392 # FIXME SUBOPTIMAL this is a very very very hot spot
1393 # while rather optimal we can *still* do much better, by
1394 # building a smarter Row::inflate_result(), and
1395 # switch to feeding it data via a much leaner interface
1397 # crude unscientific benchmarking indicated the shortcut eval is not worth it for
1398 # this particular resultset size
1399 elsif ( $self->{_result_inflator}{is_core_row} and @$rows < 60 ) {
1400 for my $r (@$rows) {
1401 $r = $inflator_cref->($res_class, $rsrc, { map { $infmap->[$_] => $r->[$_] } (0..$#$infmap) } );
1406 ( $self->{_result_inflator}{is_core_row}
1407 ? '$_ = $inflator_cref->($res_class, $rsrc, { %s }) for @$rows'
1408 # a custom inflator may be a multiplier/reductor - put it in direct list ctx
1409 : '@$rows = map { $inflator_cref->($res_class, $rsrc, { %s } ) } @$rows'
1411 ( join (', ', map { "\$infmap->[$_] => \$_->[$_]" } 0..$#$infmap ) )
1417 $self->{_result_inflator}{is_hri} ? 'hri'
1418 : $self->{_result_inflator}{is_core_row} ? 'classic_pruning'
1419 : 'classic_nonpruning'
1422 unless( $self->{_row_parser}{$parser_type}{cref} ) {
1424 # $args and $attrs to _mk_row_parser are separated to delineate what is
1425 # core collapser stuff and what is dbic $rs specific
1426 $self->{_row_parser}{$parser_type}{src} = $rsrc->_mk_row_parser({
1427 inflate_map => $infmap,
1428 collapse => $attrs->{collapse},
1429 premultiplied => $attrs->{_main_source_premultiplied},
1430 hri_style => $self->{_result_inflator}{is_hri},
1431 prune_null_branches => $self->{_result_inflator}{is_hri} || $self->{_result_inflator}{is_core_row},
1434 $self->{_row_parser}{$parser_type}{cref} = do {
1435 package # hide form PAUSE
1436 DBIx::Class::__GENERATED_ROW_PARSER__;
1438 eval $self->{_row_parser}{$parser_type}{src};
1442 # this needs to close over the *current* cursor, hence why it is not cached above
1443 my $next_cref = ($did_fetch_all or ! $attrs->{collapse})
1446 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1447 my @r = $cursor->next or return;
1452 $self->{_row_parser}{$parser_type}{cref}->(
1455 ( $self->{_stashed_rows} = [] ),
1456 ( my $null_violations = {} ),
1459 $self->throw_exception(
1460 'Collapse aborted - the following columns are declared (or defaulted to) '
1461 . 'non-nullable within DBIC but NULLs were retrieved from storage: '
1462 . join( ', ', map { "'$infmap->[$_]'" } sort { $a <=> $b } keys %$null_violations )
1463 . ' within data row ' . dump_value({
1466 ( ! defined $self->{_stashed_rows}[0][$_] or length $self->{_stashed_rows}[0][$_] < 50 )
1467 ? $self->{_stashed_rows}[0][$_]
1468 : substr( $self->{_stashed_rows}[0][$_], 0, 50 ) . '...'
1469 } 0 .. $#{$self->{_stashed_rows}[0]}
1471 ) if keys %$null_violations;
1473 # simple in-place substitution, does not regrow $rows
1474 if ($self->{_result_inflator}{is_core_row}) {
1475 $_ = $inflator_cref->($res_class, $rsrc, @$_) for @$rows
1477 # Special-case multi-object HRI - there is no $inflator_cref pass at all
1478 elsif ( ! $self->{_result_inflator}{is_hri} ) {
1479 # the inflator may be a multiplier/reductor - put it in list ctx
1480 @$rows = map { $inflator_cref->($res_class, $rsrc, @$_) } @$rows;
1484 # The @$rows check seems odd at first - why wouldn't we want to warn
1485 # regardless? The issue is things like find() etc, where the user
1486 # *knows* only one result will come back. In these cases the ->all
1487 # is not a pessimization, but rather something we actually want
1489 'Unable to properly collapse has_many results in iterator mode due '
1490 . 'to order criteria - performed an eager cursor slurp underneath. '
1491 . 'Consider using ->all() instead'
1492 ) if ( ! $fetch_all and @$rows > 1 );
1497 =head2 result_source
1501 =item Arguments: L<$result_source?|DBIx::Class::ResultSource>
1503 =item Return Value: L<$result_source|DBIx::Class::ResultSource>
1507 An accessor for the primary ResultSource object from which this ResultSet
1514 =item Arguments: $result_class?
1516 =item Return Value: $result_class
1520 An accessor for the class to use when creating result objects. Defaults to
1521 C<< result_source->result_class >> - which in most cases is the name of the
1522 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1524 Note that changing the result_class will also remove any components
1525 that were originally loaded in the source class via
1526 L<load_components|Class::C3::Componentised/load_components( @comps )>.
1527 Any overloaded methods in the original source class will not run.
1532 my ($self, $result_class) = @_;
1533 if ($result_class) {
1535 # don't fire this for an object
1536 $self->ensure_class_loaded($result_class)
1537 unless ref($result_class);
1539 if ($self->get_cache) {
1540 carp_unique('Changing the result_class of a ResultSet instance with cached results is a noop - the cache contents will not be altered');
1542 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1543 elsif ($self->{cursor} && $self->{cursor}{_pos}) {
1544 $self->throw_exception('Changing the result_class of a ResultSet instance with an active cursor is not supported');
1547 $self->_result_class($result_class);
1549 delete $self->{_result_inflator};
1551 $self->_result_class;
1558 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1560 =item Return Value: $count
1564 Performs an SQL C<COUNT> with the same query as the resultset was built
1565 with to find the number of elements. Passing arguments is equivalent to
1566 C<< $rs->search ($cond, \%attrs)->count >>
1572 return $self->search(@_)->count if @_ and defined $_[0];
1573 return scalar @{ $self->get_cache } if $self->get_cache;
1575 my $attrs = { %{ $self->_resolved_attrs } };
1577 # this is a little optimization - it is faster to do the limit
1578 # adjustments in software, instead of a subquery
1579 my ($rows, $offset) = delete @{$attrs}{qw/rows offset/};
1582 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1583 $crs = $self->_count_subq_rs ($attrs);
1586 $crs = $self->_count_rs ($attrs);
1588 my $count = $crs->next;
1590 $count -= $offset if $offset;
1591 $count = $rows if $rows and $rows < $count;
1592 $count = 0 if ($count < 0);
1601 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1603 =item Return Value: L<$count_rs|DBIx::Class::ResultSetColumn>
1607 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1608 This can be very handy for subqueries:
1610 ->search( { amount => $some_rs->count_rs->as_query } )
1612 As with regular resultsets the SQL query will be executed only after
1613 the resultset is accessed via L</next> or L</all>. That would return
1614 the same single value obtainable via L</count>.
1620 return $self->search(@_)->count_rs if @_;
1622 # this may look like a lack of abstraction (count() does about the same)
1623 # but in fact an _rs *must* use a subquery for the limits, as the
1624 # software based limiting can not be ported if this $rs is to be used
1625 # in a subquery itself (i.e. ->as_query)
1626 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1627 return $self->_count_subq_rs($self->{_attrs});
1630 return $self->_count_rs($self->{_attrs});
1635 # returns a ResultSetColumn object tied to the count query
1638 my ($self, $attrs) = @_;
1640 my $rsrc = $self->result_source;
1642 my $tmp_attrs = { %$attrs };
1643 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1644 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1646 # overwrite the selector (supplied by the storage)
1647 $rsrc->resultset_class->new($rsrc, {
1649 select => $rsrc->storage->_count_select ($rsrc, $attrs),
1651 })->get_column ('count');
1655 # same as above but uses a subquery
1657 sub _count_subq_rs {
1658 my ($self, $attrs) = @_;
1660 my $rsrc = $self->result_source;
1662 my $sub_attrs = { %$attrs };
1663 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1664 delete @{$sub_attrs}{qw/collapse columns as select order_by for/};
1666 # if we multi-prefetch we group_by something unique, as this is what we would
1667 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1668 if ( $attrs->{collapse} ) {
1669 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } @{
1670 $rsrc->_identifying_column_set || $self->throw_exception(
1671 'Unable to construct a unique group_by criteria properly collapsing the '
1672 . 'has_many prefetch before count()'
1677 # Calculate subquery selector
1678 if (my $g = $sub_attrs->{group_by}) {
1680 my $sql_maker = $rsrc->storage->sql_maker;
1682 # necessary as the group_by may refer to aliased functions
1684 for my $sel (@{$attrs->{select}}) {
1685 $sel_index->{$sel->{-as}} = $sel
1686 if (ref $sel eq 'HASH' and $sel->{-as});
1689 # anything from the original select mentioned on the group-by needs to make it to the inner selector
1690 # also look for named aggregates referred in the having clause
1691 # having often contains scalarrefs - thus parse it out entirely
1693 if ($attrs->{having}) {
1694 local $sql_maker->{having_bind};
1695 local $sql_maker->{quote_char} = $sql_maker->{quote_char};
1696 local $sql_maker->{name_sep} = $sql_maker->{name_sep};
1697 unless (defined $sql_maker->{quote_char} and length $sql_maker->{quote_char}) {
1698 $sql_maker->{quote_char} = [ "\x00", "\xFF" ];
1699 # if we don't unset it we screw up retarded but unfortunately working
1700 # 'MAX(foo.bar)' => { '>', 3 }
1701 $sql_maker->{name_sep} = '';
1704 my ($lquote, $rquote, $sep) = map { quotemeta $_ } ($sql_maker->_quote_chars, $sql_maker->name_sep);
1706 my $having_sql = $sql_maker->_parse_rs_attrs ({ having => $attrs->{having} });
1709 # search for both a proper quoted qualified string, for a naive unquoted scalarref
1710 # and if all fails for an utterly naive quoted scalar-with-function
1711 while ($having_sql =~ /
1712 $rquote $sep $lquote (.+?) $rquote
1714 [\s,] \w+ \. (\w+) [\s,]
1716 [\s,] $lquote (.+?) $rquote [\s,]
1718 my $part = $1 || $2 || $3; # one of them matched if we got here
1719 unless ($seen_having{$part}++) {
1726 my $colpiece = $sel_index->{$_} || $_;
1728 # unqualify join-based group_by's. Arcane but possible query
1729 # also horrible horrible hack to alias a column (not a func.)
1730 # (probably need to introduce SQLA syntax)
1731 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1734 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1736 push @{$sub_attrs->{select}}, $colpiece;
1740 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1741 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1744 return $rsrc->resultset_class
1745 ->new ($rsrc, $sub_attrs)
1747 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1748 ->get_column ('count');
1752 =head2 count_literal
1754 B<CAVEAT>: C<count_literal> is provided for Class::DBI compatibility and
1755 should only be used in that context. See L</search_literal> for further info.
1759 =item Arguments: $sql_fragment, @standalone_bind_values
1761 =item Return Value: $count
1765 Counts the results in a literal query. Equivalent to calling L</search_literal>
1766 with the passed arguments, then L</count>.
1770 sub count_literal { shift->search_literal(@_)->count; }
1776 =item Arguments: none
1778 =item Return Value: L<@result_objs|DBIx::Class::Manual::ResultClass>
1782 Returns all elements in the resultset.
1789 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1792 delete @{$self}{qw/_stashed_rows _stashed_results/};
1794 if (my $c = $self->get_cache) {
1798 $self->cursor->reset;
1800 my $objs = $self->_construct_results('fetch_all') || [];
1802 $self->set_cache($objs) if $self->{attrs}{cache};
1811 =item Arguments: none
1813 =item Return Value: $self
1817 Resets the resultset's cursor, so you can iterate through the elements again.
1818 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1826 delete @{$self}{qw/_stashed_rows _stashed_results/};
1827 $self->{all_cache_position} = 0;
1828 $self->cursor->reset;
1836 =item Arguments: none
1838 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1842 L<Resets|/reset> the resultset (causing a fresh query to storage) and returns
1843 an object for the first result (or C<undef> if the resultset is empty).
1848 return $_[0]->reset->next;
1854 # Determines whether and what type of subquery is required for the $rs operation.
1855 # If grouping is necessary either supplies its own, or verifies the current one
1856 # After all is done delegates to the proper storage method.
1858 sub _rs_update_delete {
1859 my ($self, $op, $values) = @_;
1861 my $rsrc = $self->result_source;
1862 my $storage = $rsrc->schema->storage;
1864 my $attrs = { %{$self->_resolved_attrs} };
1866 my $join_classifications;
1867 my ($existing_group_by) = delete @{$attrs}{qw(group_by _grouped_by_distinct)};
1869 # do we need a subquery for any reason?
1871 defined $existing_group_by
1873 # if {from} is unparseable wrap a subq
1874 ref($attrs->{from}) ne 'ARRAY'
1876 # limits call for a subq
1877 $self->_has_resolved_attr(qw/rows offset/)
1880 # simplify the joinmap, so we can further decide if a subq is necessary
1881 if (!$needs_subq and @{$attrs->{from}} > 1) {
1883 ($attrs->{from}, $join_classifications) =
1884 $storage->_prune_unused_joins ($attrs);
1886 # any non-pruneable non-local restricting joins imply subq
1887 $needs_subq = grep { $_ ne $attrs->{alias} } keys %{ $join_classifications->{restricting} || {} };
1890 # check if the head is composite (by now all joins are thrown out unless $needs_subq)
1892 (ref $attrs->{from}[0]) ne 'HASH'
1894 ref $attrs->{from}[0]{ $attrs->{from}[0]{-alias} }
1898 # do we need anything like a subquery?
1899 if (! $needs_subq) {
1900 # Most databases do not allow aliasing of tables in UPDATE/DELETE. Thus
1901 # a condition containing 'me' or other table prefixes will not work
1902 # at all. Tell SQLMaker to dequalify idents via a gross hack.
1904 my $sqla = $rsrc->storage->sql_maker;
1905 local $sqla->{_dequalify_idents} = 1;
1906 \[ $sqla->_recurse_where($self->{cond}) ];
1910 # we got this far - means it is time to wrap a subquery
1911 my $idcols = $rsrc->_identifying_column_set || $self->throw_exception(
1913 "Unable to perform complex resultset %s() without an identifying set of columns on source '%s'",
1919 # make a new $rs selecting only the PKs (that's all we really need for the subq)
1920 delete $attrs->{$_} for qw/select as collapse/;
1921 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } @$idcols ];
1923 # this will be consumed by the pruner waaaaay down the stack
1924 $attrs->{_force_prune_multiplying_joins} = 1;
1926 my $subrs = (ref $self)->new($rsrc, $attrs);
1928 if (@$idcols == 1) {
1929 $cond = { $idcols->[0] => { -in => $subrs->as_query } };
1931 elsif ($storage->_use_multicolumn_in) {
1932 # no syntax for calling this properly yet
1933 # !!! EXPERIMENTAL API !!! WILL CHANGE !!!
1934 $cond = $storage->sql_maker->_where_op_multicolumn_in (
1935 $idcols, # how do I convey a list of idents...? can binds reside on lhs?
1940 # if all else fails - get all primary keys and operate over a ORed set
1941 # wrap in a transaction for consistency
1942 # this is where the group_by/multiplication starts to matter
1946 # we do not need to check pre-multipliers, since if the premulti is there, its
1947 # parent (who is multi) will be there too
1948 keys %{ $join_classifications->{multiplying} || {} }
1950 # make sure if there is a supplied group_by it matches the columns compiled above
1951 # perfectly. Anything else can not be sanely executed on most databases so croak
1952 # right then and there
1953 if ($existing_group_by) {
1954 my @current_group_by = map
1955 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1960 join ("\x00", sort @current_group_by)
1962 join ("\x00", sort @{$attrs->{columns}} )
1964 $self->throw_exception (
1965 "You have just attempted a $op operation on a resultset which does group_by"
1966 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1967 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1968 . ' kind of queries. Please retry the operation with a modified group_by or'
1969 . ' without using one at all.'
1974 $subrs = $subrs->search({}, { group_by => $attrs->{columns} });
1977 $guard = $storage->txn_scope_guard;
1979 for my $row ($subrs->cursor->all) {
1981 { $idcols->[$_] => $row->[$_] }
1988 my $res = $cond ? $storage->$op (
1990 $op eq 'update' ? $values : (),
1994 $guard->commit if $guard;
2003 =item Arguments: \%values
2005 =item Return Value: $underlying_storage_rv
2009 Sets the specified columns in the resultset to the supplied values in a
2010 single query. Note that this will not run any accessor/set_column/update
2011 triggers, nor will it update any result object instances derived from this
2012 resultset (this includes the contents of the L<resultset cache|/set_cache>
2013 if any). See L</update_all> if you need to execute any on-update
2014 triggers or cascades defined either by you or a
2015 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2017 The return value is a pass through of what the underlying
2018 storage backend returned, and may vary. See L<DBI/execute> for the most
2023 Note that L</update> does not process/deflate any of the values passed in.
2024 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
2025 ensure manually that any value passed to this method will stringify to
2026 something the RDBMS knows how to deal with. A notable example is the
2027 handling of L<DateTime> objects, for more info see:
2028 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
2033 my ($self, $values) = @_;
2034 $self->throw_exception('Values for update must be a hash')
2035 unless ref $values eq 'HASH';
2037 return $self->_rs_update_delete ('update', $values);
2044 =item Arguments: \%values
2046 =item Return Value: 1
2050 Fetches all objects and updates them one at a time via
2051 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
2052 triggers, while L</update> will not.
2057 my ($self, $values) = @_;
2058 $self->throw_exception('Values for update_all must be a hash')
2059 unless ref $values eq 'HASH';
2061 my $guard = $self->result_source->schema->txn_scope_guard;
2062 $_->update({%$values}) for $self->all; # shallow copy - update will mangle it
2071 =item Arguments: none
2073 =item Return Value: $underlying_storage_rv
2077 Deletes the rows matching this resultset in a single query. Note that this
2078 will not run any delete triggers, nor will it alter the
2079 L<in_storage|DBIx::Class::Row/in_storage> status of any result object instances
2080 derived from this resultset (this includes the contents of the
2081 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
2082 execute any on-delete triggers or cascades defined either by you or a
2083 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2085 The return value is a pass through of what the underlying storage backend
2086 returned, and may vary. See L<DBI/execute> for the most common case.
2092 $self->throw_exception('delete does not accept any arguments')
2095 return $self->_rs_update_delete ('delete');
2102 =item Arguments: none
2104 =item Return Value: 1
2108 Fetches all objects and deletes them one at a time via
2109 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
2110 triggers, while L</delete> will not.
2116 $self->throw_exception('delete_all does not accept any arguments')
2119 my $guard = $self->result_source->schema->txn_scope_guard;
2120 $_->delete for $self->all;
2129 =item Arguments: [ \@column_list, \@row_values+ ] | [ \%col_data+ ]
2131 =item Return Value: L<\@result_objects|DBIx::Class::Manual::ResultClass> (scalar context) | L<@result_objects|DBIx::Class::Manual::ResultClass> (list context)
2135 Accepts either an arrayref of hashrefs or alternatively an arrayref of
2142 The context of this method call has an important effect on what is
2143 submitted to storage. In void context data is fed directly to fastpath
2144 insertion routines provided by the underlying storage (most often
2145 L<DBI/execute_for_fetch>), bypassing the L<new|DBIx::Class::Row/new> and
2146 L<insert|DBIx::Class::Row/insert> calls on the
2147 L<Result|DBIx::Class::Manual::ResultClass> class, including any
2148 augmentation of these methods provided by components. For example if you
2149 are using something like L<DBIx::Class::UUIDColumns> to create primary
2150 keys for you, you will find that your PKs are empty. In this case you
2151 will have to explicitly force scalar or list context in order to create
2156 In non-void (scalar or list) context, this method is simply a wrapper
2157 for L</create>. Depending on list or scalar context either a list of
2158 L<Result|DBIx::Class::Manual::ResultClass> objects or an arrayref
2159 containing these objects is returned.
2161 When supplying data in "arrayref of arrayrefs" invocation style, the
2162 first element should be a list of column names and each subsequent
2163 element should be a data value in the earlier specified column order.
2166 $schema->resultset("Artist")->populate([
2167 [ qw( artistid name ) ],
2168 [ 100, 'A Formally Unknown Singer' ],
2169 [ 101, 'A singer that jumped the shark two albums ago' ],
2170 [ 102, 'An actually cool singer' ],
2173 For the arrayref of hashrefs style each hashref should be a structure
2174 suitable for passing to L</create>. Multi-create is also permitted with
2177 $schema->resultset("Artist")->populate([
2178 { artistid => 4, name => 'Manufactured Crap', cds => [
2179 { title => 'My First CD', year => 2006 },
2180 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2183 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
2184 { title => 'My parents sold me to a record company', year => 2005 },
2185 { title => 'Why Am I So Ugly?', year => 2006 },
2186 { title => 'I Got Surgery and am now Popular', year => 2007 }
2191 If you attempt a void-context multi-create as in the example above (each
2192 Artist also has the related list of CDs), and B<do not> supply the
2193 necessary autoinc foreign key information, this method will proxy to the
2194 less efficient L</create>, and then throw the Result objects away. In this
2195 case there are obviously no benefits to using this method over L</create>.
2202 # this is naive and just a quick check
2203 # the types will need to be checked more thoroughly when the
2204 # multi-source populate gets added
2206 ref $_[0] eq 'ARRAY'
2208 ( @{$_[0]} or return )
2210 ( ref $_[0][0] eq 'HASH' or ref $_[0][0] eq 'ARRAY' )
2213 ) or $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
2215 # FIXME - no cref handling
2216 # At this point assume either hashes or arrays
2218 if(defined wantarray) {
2219 my (@results, $guard);
2221 if (ref $data->[0] eq 'ARRAY') {
2222 # column names only, nothing to do
2223 return if @$data == 1;
2225 $guard = $self->result_source->schema->storage->txn_scope_guard
2229 { my $vals = $_; $self->new_result({ map { $data->[0][$_] => $vals->[$_] } 0..$#{$data->[0]} })->insert }
2230 @{$data}[1 .. $#$data]
2235 $guard = $self->result_source->schema->storage->txn_scope_guard
2238 @results = map { $self->new_result($_)->insert } @$data;
2241 $guard->commit if $guard;
2242 return wantarray ? @results : \@results;
2245 # we have to deal with *possibly incomplete* related data
2246 # this means we have to walk the data structure twice
2247 # whether we want this or not
2248 # jnap, I hate you ;)
2249 my $rsrc = $self->result_source;
2250 my $rel_info = { map { $_ => $rsrc->relationship_info($_) } $rsrc->relationships };
2252 my ($colinfo, $colnames, $slices_with_rels);
2256 for my $i (0 .. $#$data) {
2258 my $current_slice_seen_rel_infos;
2260 ### Determine/Supplement collists
2261 ### BEWARE - This is a hot piece of code, a lot of weird idioms were used
2262 if( ref $data->[$i] eq 'ARRAY' ) {
2264 # positional(!) explicit column list
2266 # column names only, nothing to do
2267 return if @$data == 1;
2269 $colinfo->{$data->[0][$_]} = { pos => $_, name => $data->[0][$_] } and push @$colnames, $data->[0][$_]
2270 for 0 .. $#{$data->[0]};
2277 for (values %$colinfo) {
2278 if ($_->{is_rel} ||= (
2279 $rel_info->{$_->{name}}
2282 ref $data->[$i][$_->{pos}] eq 'ARRAY'
2284 ref $data->[$i][$_->{pos}] eq 'HASH'
2286 ( defined blessed $data->[$i][$_->{pos}] and $data->[$i][$_->{pos}]->isa('DBIx::Class::Row') )
2292 # moar sanity check... sigh
2293 for ( ref $data->[$i][$_->{pos}] eq 'ARRAY' ? @{$data->[$i][$_->{pos}]} : $data->[$i][$_->{pos}] ) {
2294 if ( defined blessed $_ and $_->isa('DBIx::Class::Row' ) ) {
2295 carp_unique("Fast-path populate() with supplied related objects is not possible - falling back to regular create()");
2296 return my $throwaway = $self->populate(@_);
2300 push @$current_slice_seen_rel_infos, $rel_info->{$_->{name}};
2305 if ($current_slice_seen_rel_infos) {
2306 push @$slices_with_rels, { map { $colnames->[$_] => $data->[$i][$_] } 0 .. $#$colnames };
2308 # this is needed further down to decide whether or not to fallback to create()
2309 $colinfo->{$colnames->[$_]}{seen_null} ||= ! defined $data->[$i][$_]
2310 for 0 .. $#$colnames;
2313 elsif( ref $data->[$i] eq 'HASH' ) {
2315 for ( sort keys %{$data->[$i]} ) {
2317 $colinfo->{$_} ||= do {
2319 $self->throw_exception("Column '$_' must be present in supplied explicit column list")
2320 if $data_start; # it will be 0 on AoH, 1 on AoA
2322 push @$colnames, $_;
2325 { pos => $#$colnames, name => $_ }
2328 if ($colinfo->{$_}{is_rel} ||= (
2332 ref $data->[$i]{$_} eq 'ARRAY'
2334 ref $data->[$i]{$_} eq 'HASH'
2336 ( defined blessed $data->[$i]{$_} and $data->[$i]{$_}->isa('DBIx::Class::Row') )
2342 # moar sanity check... sigh
2343 for ( ref $data->[$i]{$_} eq 'ARRAY' ? @{$data->[$i]{$_}} : $data->[$i]{$_} ) {
2344 if ( defined blessed $_ and $_->isa('DBIx::Class::Row' ) ) {
2345 carp_unique("Fast-path populate() with supplied related objects is not possible - falling back to regular create()");
2346 return my $throwaway = $self->populate(@_);
2350 push @$current_slice_seen_rel_infos, $rel_info->{$_};
2354 if ($current_slice_seen_rel_infos) {
2355 push @$slices_with_rels, $data->[$i];
2357 # this is needed further down to decide whether or not to fallback to create()
2358 $colinfo->{$_}{seen_null} ||= ! defined $data->[$i]{$_}
2359 for keys %{$data->[$i]};
2363 $self->throw_exception('Unexpected populate() data structure member type: ' . ref $data->[$i] );
2367 { $_->{attrs}{is_depends_on} }
2368 @{ $current_slice_seen_rel_infos || [] }
2370 carp_unique("Fast-path populate() of belongs_to relationship data is not possible - falling back to regular create()");
2371 return my $throwaway = $self->populate(@_);
2375 if( $slices_with_rels ) {
2377 # need to exclude the rel "columns"
2378 $colnames = [ grep { ! $colinfo->{$_}{is_rel} } @$colnames ];
2380 # extra sanity check - ensure the main source is in fact identifiable
2381 # the localizing of nullability is insane, but oh well... the use-case is legit
2382 my $ci = $rsrc->columns_info($colnames);
2384 $ci->{$_} = { %{$ci->{$_}}, is_nullable => 0 }
2385 for grep { ! $colinfo->{$_}{seen_null} } keys %$ci;
2387 unless( $rsrc->_identifying_column_set($ci) ) {
2388 carp_unique("Fast-path populate() of non-uniquely identifiable rows with related data is not possible - falling back to regular create()");
2389 return my $throwaway = $self->populate(@_);
2393 ### inherit the data locked in the conditions of the resultset
2394 my ($rs_data) = $self->_merge_with_rscond({});
2395 delete @{$rs_data}{@$colnames}; # passed-in stuff takes precedence
2397 # if anything left - decompose rs_data
2399 if (keys %$rs_data) {
2400 push @$rs_data_vals, $rs_data->{$_}
2401 for sort keys %$rs_data;
2406 $guard = $rsrc->schema->storage->txn_scope_guard
2407 if $slices_with_rels;
2409 ### main source data
2410 # FIXME - need to switch entirely to a coderef-based thing,
2411 # so that large sets aren't copied several times... I think
2412 $rsrc->storage->_insert_bulk(
2414 [ @$colnames, sort keys %$rs_data ],
2416 ref $data->[$_] eq 'ARRAY'
2418 $slices_with_rels ? [ @{$data->[$_]}[0..$#$colnames], @{$rs_data_vals||[]} ] # the collist changed
2419 : $rs_data_vals ? [ @{$data->[$_]}, @$rs_data_vals ]
2422 : [ @{$data->[$_]}{@$colnames}, @{$rs_data_vals||[]} ]
2423 } $data_start .. $#$data ],
2426 ### do the children relationships
2427 if ( $slices_with_rels ) {
2428 my @rels = grep { $colinfo->{$_}{is_rel} } keys %$colinfo
2429 or die 'wtf... please report a bug with DBIC_TRACE=1 output (stacktrace)';
2431 for my $sl (@$slices_with_rels) {
2433 my ($main_proto, $main_proto_rs);
2434 for my $rel (@rels) {
2435 next unless defined $sl->{$rel};
2439 (map { $_ => $sl->{$_} } @$colnames),
2442 unless (defined $colinfo->{$rel}{rs}) {
2444 $colinfo->{$rel}{rs} = $rsrc->related_source($rel)->resultset;
2446 $colinfo->{$rel}{fk_map} = { reverse %{ $rsrc->_resolve_relationship_condition(
2448 self_alias => "\xFE", # irrelevant
2449 foreign_alias => "\xFF", # irrelevant
2450 )->{identity_map} || {} } };
2454 $colinfo->{$rel}{rs}->search({ map # only so that we inherit them values properly, no actual search
2457 ( $main_proto_rs ||= $rsrc->resultset->search($main_proto) )
2458 ->get_column( $colinfo->{$rel}{fk_map}{$_} )
2462 keys %{$colinfo->{$rel}{fk_map}}
2463 })->populate( ref $sl->{$rel} eq 'ARRAY' ? $sl->{$rel} : [ $sl->{$rel} ] );
2470 $guard->commit if $guard;
2477 =item Arguments: none
2479 =item Return Value: L<$pager|Data::Page>
2483 Returns a L<Data::Page> object for the current resultset. Only makes
2484 sense for queries with a C<page> attribute.
2486 To get the full count of entries for a paged resultset, call
2487 C<total_entries> on the L<Data::Page> object.
2494 return $self->{pager} if $self->{pager};
2496 my $attrs = $self->{attrs};
2497 if (!defined $attrs->{page}) {
2498 $self->throw_exception("Can't create pager for non-paged rs");
2500 elsif ($attrs->{page} <= 0) {
2501 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
2503 $attrs->{rows} ||= 10;
2505 # throw away the paging flags and re-run the count (possibly
2506 # with a subselect) to get the real total count
2507 my $count_attrs = { %$attrs };
2508 delete @{$count_attrs}{qw/rows offset page pager/};
2510 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
2512 require DBIx::Class::ResultSet::Pager;
2513 return $self->{pager} = DBIx::Class::ResultSet::Pager->new(
2514 sub { $total_rs->count }, #lazy-get the total
2516 $self->{attrs}{page},
2524 =item Arguments: $page_number
2526 =item Return Value: L<$resultset|/search>
2530 Returns a resultset for the $page_number page of the resultset on which page
2531 is called, where each page contains a number of rows equal to the 'rows'
2532 attribute set on the resultset (10 by default).
2537 my ($self, $page) = @_;
2538 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2545 =item Arguments: \%col_data
2547 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2551 Creates a new result object in the resultset's result class and returns
2552 it. The row is not inserted into the database at this point, call
2553 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2554 will tell you whether the result object has been inserted or not.
2556 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2561 my ($self, $values) = @_;
2563 $self->throw_exception( "new_result takes only one argument - a hashref of values" )
2566 $self->throw_exception( "Result object instantiation requires a hashref as argument" )
2567 unless (ref $values eq 'HASH');
2569 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2571 my $new = $self->result_class->new({
2573 ( @$cols_from_relations
2574 ? (-cols_from_relations => $cols_from_relations)
2577 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2581 reftype($new) eq 'HASH'
2587 carp_unique (sprintf (
2588 "%s->new returned a blessed empty hashref - a strong indicator something is wrong with its inheritance chain",
2589 $self->result_class,
2596 # _merge_with_rscond
2598 # Takes a simple hash of K/V data and returns its copy merged with the
2599 # condition already present on the resultset. Additionally returns an
2600 # arrayref of value/condition names, which were inferred from related
2601 # objects (this is needed for in-memory related objects)
2602 sub _merge_with_rscond {
2603 my ($self, $data) = @_;
2605 my ($implied_data, @cols_from_relations);
2607 my $alias = $self->{attrs}{alias};
2609 if (! defined $self->{cond}) {
2610 # just massage $data below
2612 elsif ($self->{cond} eq UNRESOLVABLE_CONDITION) {
2613 $implied_data = $self->{attrs}{related_objects}; # nothing might have been inserted yet
2614 @cols_from_relations = keys %{ $implied_data || {} };
2617 my $eqs = $self->result_source->schema->storage->_extract_fixed_condition_columns($self->{cond}, 'consider_nulls');
2618 $implied_data = { map {
2619 ( ($eqs->{$_}||'') eq UNRESOLVABLE_CONDITION ) ? () : ( $_ => $eqs->{$_} )
2625 { %{ $self->_remove_alias($_, $alias) } }
2626 # precedence must be given to passed values over values inherited from
2627 # the cond, so the order here is important.
2628 ( $implied_data||(), $data)
2630 \@cols_from_relations
2634 # _has_resolved_attr
2636 # determines if the resultset defines at least one
2637 # of the attributes supplied
2639 # used to determine if a subquery is necessary
2641 # supports some virtual attributes:
2643 # This will scan for any joins being present on the resultset.
2644 # It is not a mere key-search but a deep inspection of {from}
2647 sub _has_resolved_attr {
2648 my ($self, @attr_names) = @_;
2650 my $attrs = $self->_resolved_attrs;
2654 for my $n (@attr_names) {
2655 if (grep { $n eq $_ } (qw/-join/) ) {
2656 $extra_checks{$n}++;
2660 my $attr = $attrs->{$n};
2662 next if not defined $attr;
2664 if (ref $attr eq 'HASH') {
2665 return 1 if keys %$attr;
2667 elsif (ref $attr eq 'ARRAY') {
2675 # a resolved join is expressed as a multi-level from
2677 $extra_checks{-join}
2679 ref $attrs->{from} eq 'ARRAY'
2681 @{$attrs->{from}} > 1
2689 # Remove the specified alias from the specified query hash. A copy is made so
2690 # the original query is not modified.
2693 my ($self, $query, $alias) = @_;
2695 my %orig = %{ $query || {} };
2698 foreach my $key (keys %orig) {
2700 $unaliased{$key} = $orig{$key};
2703 $unaliased{$1} = $orig{$key}
2704 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2714 =item Arguments: none
2716 =item Return Value: \[ $sql, L<@bind_values|/DBIC BIND VALUES> ]
2720 Returns the SQL query and bind vars associated with the invocant.
2722 This is generally used as the RHS for a subquery.
2729 my $attrs = { %{ $self->_resolved_attrs } };
2731 my $aq = $self->result_source->storage->_select_args_to_query (
2732 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
2742 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2744 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2748 my $artist = $schema->resultset('Artist')->find_or_new(
2749 { artist => 'fred' }, { key => 'artists' });
2751 $cd->cd_to_producer->find_or_new({ producer => $producer },
2752 { key => 'primary' });
2754 Find an existing record from this resultset using L</find>. if none exists,
2755 instantiate a new result object and return it. The object will not be saved
2756 into your storage until you call L<DBIx::Class::Row/insert> on it.
2758 You most likely want this method when looking for existing rows using a unique
2759 constraint that is not the primary key, or looking for related rows.
2761 If you want objects to be saved immediately, use L</find_or_create> instead.
2763 B<Note>: Make sure to read the documentation of L</find> and understand the
2764 significance of the C<key> attribute, as its lack may skew your search, and
2765 subsequently result in spurious new objects.
2767 B<Note>: Take care when using C<find_or_new> with a table having
2768 columns with default values that you intend to be automatically
2769 supplied by the database (e.g. an auto_increment primary key column).
2770 In normal usage, the value of such columns should NOT be included at
2771 all in the call to C<find_or_new>, even when set to C<undef>.
2777 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2778 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2779 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2782 return $self->new_result($hash);
2789 =item Arguments: \%col_data
2791 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2795 Attempt to create a single new row or a row with multiple related rows
2796 in the table represented by the resultset (and related tables). This
2797 will not check for duplicate rows before inserting, use
2798 L</find_or_create> to do that.
2800 To create one row for this resultset, pass a hashref of key/value
2801 pairs representing the columns of the table and the values you wish to
2802 store. If the appropriate relationships are set up, foreign key fields
2803 can also be passed an object representing the foreign row, and the
2804 value will be set to its primary key.
2806 To create related objects, pass a hashref of related-object column values
2807 B<keyed on the relationship name>. If the relationship is of type C<multi>
2808 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2809 The process will correctly identify columns holding foreign keys, and will
2810 transparently populate them from the keys of the corresponding relation.
2811 This can be applied recursively, and will work correctly for a structure
2812 with an arbitrary depth and width, as long as the relationships actually
2813 exists and the correct column data has been supplied.
2815 Instead of hashrefs of plain related data (key/value pairs), you may
2816 also pass new or inserted objects. New objects (not inserted yet, see
2817 L</new_result>), will be inserted into their appropriate tables.
2819 Effectively a shortcut for C<< ->new_result(\%col_data)->insert >>.
2821 Example of creating a new row.
2823 $person_rs->create({
2824 name=>"Some Person",
2825 email=>"somebody@someplace.com"
2828 Example of creating a new row and also creating rows in a related C<has_many>
2829 or C<has_one> resultset. Note Arrayref.
2832 { artistid => 4, name => 'Manufactured Crap', cds => [
2833 { title => 'My First CD', year => 2006 },
2834 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2839 Example of creating a new row and also creating a row in a related
2840 C<belongs_to> resultset. Note Hashref.
2843 title=>"Music for Silly Walks",
2846 name=>"Silly Musician",
2854 When subclassing ResultSet never attempt to override this method. Since
2855 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2856 lot of the internals simply never call it, so your override will be
2857 bypassed more often than not. Override either L<DBIx::Class::Row/new>
2858 or L<DBIx::Class::Row/insert> depending on how early in the
2859 L</create> process you need to intervene. See also warning pertaining to
2867 #my ($self, $col_data) = @_;
2868 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
2869 return shift->new_result(shift)->insert;
2872 =head2 find_or_create
2876 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2878 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2882 $cd->cd_to_producer->find_or_create({ producer => $producer },
2883 { key => 'primary' });
2885 Tries to find a record based on its primary key or unique constraints; if none
2886 is found, creates one and returns that instead.
2888 my $cd = $schema->resultset('CD')->find_or_create({
2890 artist => 'Massive Attack',
2891 title => 'Mezzanine',
2895 Also takes an optional C<key> attribute, to search by a specific key or unique
2896 constraint. For example:
2898 my $cd = $schema->resultset('CD')->find_or_create(
2900 artist => 'Massive Attack',
2901 title => 'Mezzanine',
2903 { key => 'cd_artist_title' }
2906 B<Note>: Make sure to read the documentation of L</find> and understand the
2907 significance of the C<key> attribute, as its lack may skew your search, and
2908 subsequently result in spurious row creation.
2910 B<Note>: Because find_or_create() reads from the database and then
2911 possibly inserts based on the result, this method is subject to a race
2912 condition. Another process could create a record in the table after
2913 the find has completed and before the create has started. To avoid
2914 this problem, use find_or_create() inside a transaction.
2916 B<Note>: Take care when using C<find_or_create> with a table having
2917 columns with default values that you intend to be automatically
2918 supplied by the database (e.g. an auto_increment primary key column).
2919 In normal usage, the value of such columns should NOT be included at
2920 all in the call to C<find_or_create>, even when set to C<undef>.
2922 See also L</find> and L</update_or_create>. For information on how to declare
2923 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2925 If you need to know if an existing row was found or a new one created use
2926 L</find_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2927 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2930 my $cd = $schema->resultset('CD')->find_or_new({
2932 artist => 'Massive Attack',
2933 title => 'Mezzanine',
2937 if( !$cd->in_storage ) {
2944 sub find_or_create {
2946 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2947 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2948 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2951 return $self->new_result($hash)->insert;
2954 =head2 update_or_create
2958 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2960 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2964 $resultset->update_or_create({ col => $val, ... });
2966 Like L</find_or_create>, but if a row is found it is immediately updated via
2967 C<< $found_row->update (\%col_data) >>.
2970 Takes an optional C<key> attribute to search on a specific unique constraint.
2973 # In your application
2974 my $cd = $schema->resultset('CD')->update_or_create(
2976 artist => 'Massive Attack',
2977 title => 'Mezzanine',
2980 { key => 'cd_artist_title' }
2983 $cd->cd_to_producer->update_or_create({
2984 producer => $producer,
2990 B<Note>: Make sure to read the documentation of L</find> and understand the
2991 significance of the C<key> attribute, as its lack may skew your search, and
2992 subsequently result in spurious row creation.
2994 B<Note>: Take care when using C<update_or_create> with a table having
2995 columns with default values that you intend to be automatically
2996 supplied by the database (e.g. an auto_increment primary key column).
2997 In normal usage, the value of such columns should NOT be included at
2998 all in the call to C<update_or_create>, even when set to C<undef>.
3000 See also L</find> and L</find_or_create>. For information on how to declare
3001 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
3003 If you need to know if an existing row was updated or a new one created use
3004 L</update_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
3005 to call L<DBIx::Class::Row/insert> to save the newly created row to the
3010 sub update_or_create {
3012 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
3013 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
3015 my $row = $self->find($cond, $attrs);
3017 $row->update($cond);
3021 return $self->new_result($cond)->insert;
3024 =head2 update_or_new
3028 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
3030 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
3034 $resultset->update_or_new({ col => $val, ... });
3036 Like L</find_or_new> but if a row is found it is immediately updated via
3037 C<< $found_row->update (\%col_data) >>.
3041 # In your application
3042 my $cd = $schema->resultset('CD')->update_or_new(
3044 artist => 'Massive Attack',
3045 title => 'Mezzanine',
3048 { key => 'cd_artist_title' }
3051 if ($cd->in_storage) {
3052 # the cd was updated
3055 # the cd is not yet in the database, let's insert it
3059 B<Note>: Make sure to read the documentation of L</find> and understand the
3060 significance of the C<key> attribute, as its lack may skew your search, and
3061 subsequently result in spurious new objects.
3063 B<Note>: Take care when using C<update_or_new> with a table having
3064 columns with default values that you intend to be automatically
3065 supplied by the database (e.g. an auto_increment primary key column).
3066 In normal usage, the value of such columns should NOT be included at
3067 all in the call to C<update_or_new>, even when set to C<undef>.
3069 See also L</find>, L</find_or_create> and L</find_or_new>.
3075 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
3076 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
3078 my $row = $self->find( $cond, $attrs );
3079 if ( defined $row ) {
3080 $row->update($cond);
3084 return $self->new_result($cond);
3091 =item Arguments: none
3093 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass> | undef
3097 Gets the contents of the cache for the resultset, if the cache is set.
3099 The cache is populated either by using the L</prefetch> attribute to
3100 L</search> or by calling L</set_cache>.
3112 =item Arguments: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3114 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3118 Sets the contents of the cache for the resultset. Expects an arrayref
3119 of objects of the same class as those produced by the resultset. Note that
3120 if the cache is set, the resultset will return the cached objects rather
3121 than re-querying the database even if the cache attr is not set.
3123 The contents of the cache can also be populated by using the
3124 L</prefetch> attribute to L</search>.
3129 my ( $self, $data ) = @_;
3130 $self->throw_exception("set_cache requires an arrayref")
3131 if defined($data) && (ref $data ne 'ARRAY');
3132 $self->{all_cache} = $data;
3139 =item Arguments: none
3141 =item Return Value: undef
3145 Clears the cache for the resultset.
3150 shift->set_cache(undef);
3157 =item Arguments: none
3159 =item Return Value: true, if the resultset has been paginated
3167 return !!$self->{attrs}{page};
3174 =item Arguments: none
3176 =item Return Value: true, if the resultset has been ordered with C<order_by>.
3184 return scalar $self->result_source->storage->_extract_order_criteria($self->{attrs}{order_by});
3187 =head2 related_resultset
3191 =item Arguments: $rel_name
3193 =item Return Value: L<$resultset|/search>
3197 Returns a related resultset for the supplied relationship name.
3199 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
3203 sub related_resultset {
3204 $_[0]->throw_exception(
3205 'Extra arguments to $rs->related_resultset() were always quietly '
3206 . 'discarded without consideration, you need to switch to '
3207 . '...->related_resultset( $relname )->search_rs( $search, $args ) instead.'
3210 return $_[0]->{related_resultsets}{$_[1]}
3211 if defined $_[0]->{related_resultsets}{$_[1]};
3213 my ($self, $rel) = @_;
3215 return $self->{related_resultsets}{$rel} = do {
3216 my $rsrc = $self->result_source;
3217 my $rel_info = $rsrc->relationship_info($rel);
3219 $self->throw_exception(
3220 "search_related: result source '" . $rsrc->source_name .
3221 "' has no such relationship $rel")
3224 my $attrs = $self->_chain_relationship($rel);
3226 my $storage = $rsrc->schema->storage;
3228 # Previously this atribute was deleted (instead of being set as it is now)
3229 # Doing so seems to be harmless in all available test permutations
3230 # See also 01d59a6a6 and mst's comment below
3232 $attrs->{alias} = $storage->relname_to_table_alias(
3234 $attrs->{seen_join}{$rel}
3237 # since this is search_related, and we already slid the select window inwards
3238 # (the select/as attrs were deleted in the beginning), we need to flip all
3239 # left joins to inner, so we get the expected results
3240 # read the comment on top of the actual function to see what this does
3241 $attrs->{from} = $storage->_inner_join_to_node( $attrs->{from}, $attrs->{alias} );
3243 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
3244 delete $attrs->{result_class};
3248 # The reason we do this now instead of passing the alias to the
3249 # search_rs below is that if you wrap/overload resultset on the
3250 # source you need to know what alias it's -going- to have for things
3251 # to work sanely (e.g. RestrictWithObject wants to be able to add
3252 # extra query restrictions, and these may need to be $alias.)
3253 # -- mst ~ 2007 (01d59a6a6)
3255 # FIXME - this seems to be no longer neccessary (perhaps due to the
3256 # advances in relcond resolution. Testing DBIC::S::RWO and its only
3257 # dependent (as of Jun 2015 ) does not yield any difference with or
3258 # without this line. Nevertheless keep it as is for now, to minimize
3259 # churn, there is enough potential for breakage in 0.0829xx as it is
3260 # -- ribasushi Jun 2015
3262 my $rel_source = $rsrc->related_source($rel);
3263 local $rel_source->resultset_attributes->{alias} = $attrs->{alias};
3265 $rel_source->resultset->search_rs( undef, $attrs );
3268 if (my $cache = $self->get_cache) {
3269 my @related_cache = map
3270 { $_->related_resultset($rel)->get_cache || () }
3274 $new->set_cache([ map @$_, @related_cache ]) if @related_cache == @$cache;
3281 =head2 current_source_alias
3285 =item Arguments: none
3287 =item Return Value: $source_alias
3291 Returns the current table alias for the result source this resultset is built
3292 on, that will be used in the SQL query. Usually it is C<me>.
3294 Currently the source alias that refers to the result set returned by a
3295 L</search>/L</find> family method depends on how you got to the resultset: it's
3296 C<me> by default, but eg. L</search_related> aliases it to the related result
3297 source name (and keeps C<me> referring to the original result set). The long
3298 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
3299 (and make this method unnecessary).
3301 Thus it's currently necessary to use this method in predefined queries (see
3302 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
3303 source alias of the current result set:
3305 # in a result set class
3307 my ($self, $user) = @_;
3309 my $me = $self->current_source_alias;
3311 return $self->search({
3312 "$me.modified" => $user->id,
3316 The alias of L<newly created resultsets|/search> can be altered by the
3317 L<alias attribute|/alias>.
3321 sub current_source_alias {
3322 return (shift->{attrs} || {})->{alias} || 'me';
3325 =head2 as_subselect_rs
3329 =item Arguments: none
3331 =item Return Value: L<$resultset|/search>
3335 Act as a barrier to SQL symbols. The resultset provided will be made into a
3336 "virtual view" by including it as a subquery within the from clause. From this
3337 point on, any joined tables are inaccessible to ->search on the resultset (as if
3338 it were simply where-filtered without joins). For example:
3340 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
3342 # 'x' now pollutes the query namespace
3344 # So the following works as expected
3345 my $ok_rs = $rs->search({'x.other' => 1});
3347 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
3348 # def) we look for one row with contradictory terms and join in another table
3349 # (aliased 'x_2') which we never use
3350 my $broken_rs = $rs->search({'x.name' => 'def'});
3352 my $rs2 = $rs->as_subselect_rs;
3354 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
3355 my $not_joined_rs = $rs2->search({'x.other' => 1});
3357 # works as expected: finds a 'table' row related to two x rows (abc and def)
3358 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
3360 Another example of when one might use this would be to select a subset of
3361 columns in a group by clause:
3363 my $rs = $schema->resultset('Bar')->search(undef, {
3364 group_by => [qw{ id foo_id baz_id }],
3365 })->as_subselect_rs->search(undef, {
3366 columns => [qw{ id foo_id }]
3369 In the above example normally columns would have to be equal to the group by,
3370 but because we isolated the group by into a subselect the above works.
3374 sub as_subselect_rs {
3377 my $attrs = $self->_resolved_attrs;
3379 my $fresh_rs = (ref $self)->new (
3380 $self->result_source
3383 # these pieces will be locked in the subquery
3384 delete $fresh_rs->{cond};
3385 delete @{$fresh_rs->{attrs}}{qw/where bind/};
3387 return $fresh_rs->search( {}, {
3389 $attrs->{alias} => $self->as_query,
3390 -alias => $attrs->{alias},
3391 -rsrc => $self->result_source,
3393 alias => $attrs->{alias},
3397 # This code is called by search_related, and makes sure there
3398 # is clear separation between the joins before, during, and
3399 # after the relationship. This information is needed later
3400 # in order to properly resolve prefetch aliases (any alias
3401 # with a relation_chain_depth less than the depth of the
3402 # current prefetch is not considered)
3404 # The increments happen twice per join. An even number means a
3405 # relationship specified via a search_related, whereas an odd
3406 # number indicates a join/prefetch added via attributes
3408 # Also this code will wrap the current resultset (the one we
3409 # chain to) in a subselect IFF it contains limiting attributes
3410 sub _chain_relationship {
3411 my ($self, $rel) = @_;
3412 my $source = $self->result_source;
3413 my $attrs = { %{$self->{attrs}||{}} };
3415 # we need to take the prefetch the attrs into account before we
3416 # ->_resolve_join as otherwise they get lost - captainL
3417 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
3419 delete @{$attrs}{qw/join prefetch collapse group_by distinct _grouped_by_distinct select as columns +select +as +columns/};
3421 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
3424 my @force_subq_attrs = qw/offset rows group_by having/;
3427 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
3429 $self->_has_resolved_attr (@force_subq_attrs)
3431 # Nuke the prefetch (if any) before the new $rs attrs
3432 # are resolved (prefetch is useless - we are wrapping
3433 # a subquery anyway).
3434 my $rs_copy = $self->search;
3435 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
3436 $rs_copy->{attrs}{join},
3437 delete $rs_copy->{attrs}{prefetch},
3442 -alias => $attrs->{alias},
3443 $attrs->{alias} => $rs_copy->as_query,
3445 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
3446 $seen->{-relation_chain_depth} = 0;
3448 elsif ($attrs->{from}) { #shallow copy suffices
3449 $from = [ @{$attrs->{from}} ];
3454 -alias => $attrs->{alias},
3455 $attrs->{alias} => $source->from,
3459 my $jpath = ($seen->{-relation_chain_depth})
3460 ? $from->[-1][0]{-join_path}
3463 my @requested_joins = $source->_resolve_join(
3470 push @$from, @requested_joins;
3472 $seen->{-relation_chain_depth}++;
3474 # if $self already had a join/prefetch specified on it, the requested
3475 # $rel might very well be already included. What we do in this case
3476 # is effectively a no-op (except that we bump up the chain_depth on
3477 # the join in question so we could tell it *is* the search_related)
3480 # we consider the last one thus reverse
3481 for my $j (reverse @requested_joins) {
3482 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
3483 if ($rel eq $last_j) {
3484 $j->[0]{-relation_chain_depth}++;
3490 unless ($already_joined) {
3491 push @$from, $source->_resolve_join(
3499 $seen->{-relation_chain_depth}++;
3501 return {%$attrs, from => $from, seen_join => $seen};
3504 sub _resolved_attrs {
3506 return $self->{_attrs} if $self->{_attrs};
3508 my $attrs = { %{ $self->{attrs} || {} } };
3509 my $source = $attrs->{result_source} = $self->result_source;
3510 my $alias = $attrs->{alias};
3512 $self->throw_exception("Specifying distinct => 1 in conjunction with collapse => 1 is unsupported")
3513 if $attrs->{collapse} and $attrs->{distinct};
3516 # Sanity check the paging attributes
3517 # SQLMaker does it too, but in case of a software_limit we'll never get there
3518 if (defined $attrs->{offset}) {
3519 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
3520 if ( $attrs->{offset} =~ /[^0-9]/ or $attrs->{offset} < 0 );
3522 if (defined $attrs->{rows}) {
3523 $self->throw_exception("The rows attribute must be a positive integer if present")
3524 if ( $attrs->{rows} =~ /[^0-9]/ or $attrs->{rows} <= 0 );
3528 # default selection list
3529 $attrs->{columns} = [ $source->columns ]
3530 unless grep { exists $attrs->{$_} } qw/columns cols select as/;
3532 # merge selectors together
3533 for (qw/columns select as/) {
3534 $attrs->{$_} = $self->_merge_attr($attrs->{$_}, delete $attrs->{"+$_"})
3535 if $attrs->{$_} or $attrs->{"+$_"};
3538 # disassemble columns
3540 if (my $cols = delete $attrs->{columns}) {
3541 for my $c (ref $cols eq 'ARRAY' ? @$cols : $cols) {
3542 if (ref $c eq 'HASH') {
3543 for my $as (sort keys %$c) {
3544 push @sel, $c->{$as};
3555 # when trying to weed off duplicates later do not go past this point -
3556 # everything added from here on is unbalanced "anyone's guess" stuff
3557 my $dedup_stop_idx = $#as;
3559 push @as, @{ ref $attrs->{as} eq 'ARRAY' ? $attrs->{as} : [ $attrs->{as} ] }
3561 push @sel, @{ ref $attrs->{select} eq 'ARRAY' ? $attrs->{select} : [ $attrs->{select} ] }
3562 if $attrs->{select};
3564 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3565 $_ = (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" for @sel;
3567 # disqualify all $alias.col as-bits (inflate-map mandated)
3568 $_ = ($_ =~ /^\Q$alias.\E(.+)$/) ? $1 : $_ for @as;
3570 # de-duplicate the result (remove *identical* select/as pairs)
3571 # and also die on duplicate {as} pointing to different {select}s
3572 # not using a c-style for as the condition is prone to shrinkage
3575 while ($i <= $dedup_stop_idx) {
3576 if ($seen->{"$sel[$i] \x00\x00 $as[$i]"}++) {
3581 elsif ($seen->{$as[$i]}++) {
3582 $self->throw_exception(
3583 "inflate_result() alias '$as[$i]' specified twice with different SQL-side {select}-ors"
3591 $attrs->{select} = \@sel;
3592 $attrs->{as} = \@as;
3594 $attrs->{from} ||= [{
3596 -alias => $self->{attrs}{alias},
3597 $self->{attrs}{alias} => $source->from,
3600 if ( $attrs->{join} || $attrs->{prefetch} ) {
3602 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3603 if ref $attrs->{from} ne 'ARRAY';
3605 my $join = (delete $attrs->{join}) || {};
3607 if ( defined $attrs->{prefetch} ) {
3608 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3611 $attrs->{from} = # have to copy here to avoid corrupting the original
3613 @{ $attrs->{from} },
3614 $source->_resolve_join(
3617 { %{ $attrs->{seen_join} || {} } },
3618 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3619 ? $attrs->{from}[-1][0]{-join_path}
3627 for my $attr (qw(order_by group_by)) {
3629 if ( defined $attrs->{$attr} ) {
3631 ref( $attrs->{$attr} ) eq 'ARRAY'
3632 ? [ @{ $attrs->{$attr} } ]
3633 : [ $attrs->{$attr} || () ]
3636 delete $attrs->{$attr} unless @{$attrs->{$attr}};
3641 # set collapse default based on presence of prefetch
3644 defined $attrs->{prefetch}
3646 $prefetch = $self->_merge_joinpref_attr( {}, delete $attrs->{prefetch} )
3648 $self->throw_exception("Specifying prefetch in conjunction with an explicit collapse => 0 is unsupported")
3649 if defined $attrs->{collapse} and ! $attrs->{collapse};
3651 $attrs->{collapse} = 1;
3655 # run through the resulting joinstructure (starting from our current slot)
3656 # and unset collapse if proven unnecessary
3658 # also while we are at it find out if the current root source has
3659 # been premultiplied by previous related_source chaining
3661 # this allows to predict whether a root object with all other relation
3662 # data set to NULL is in fact unique
3663 if ($attrs->{collapse}) {
3665 if (ref $attrs->{from} eq 'ARRAY') {
3667 if (@{$attrs->{from}} == 1) {
3668 # no joins - no collapse
3669 $attrs->{collapse} = 0;
3672 # find where our table-spec starts
3673 my @fromlist = @{$attrs->{from}};
3675 my $t = shift @fromlist;
3678 # me vs join from-spec distinction - a ref means non-root
3679 if (ref $t eq 'ARRAY') {
3681 $is_multi ||= ! $t->{-is_single};
3683 last if ($t->{-alias} && $t->{-alias} eq $alias);
3684 $attrs->{_main_source_premultiplied} ||= $is_multi;
3687 # no non-singles remaining, nor any premultiplication - nothing to collapse
3689 ! $attrs->{_main_source_premultiplied}
3691 ! grep { ! $_->[0]{-is_single} } @fromlist
3693 $attrs->{collapse} = 0;
3699 # if we can not analyze the from - err on the side of safety
3700 $attrs->{_main_source_premultiplied} = 1;
3705 # generate the distinct induced group_by before injecting the prefetched select/as parts
3706 if (delete $attrs->{distinct}) {
3707 if ($attrs->{group_by}) {
3708 carp_unique ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3711 $attrs->{_grouped_by_distinct} = 1;
3712 # distinct affects only the main selection part, not what prefetch may add below
3713 ($attrs->{group_by}, my $new_order) = $source->storage->_group_over_selection($attrs);
3715 # FIXME possibly ignore a rewritten order_by (may turn out to be an issue)
3716 # The thinking is: if we are collapsing the subquerying prefetch engine will
3717 # rip stuff apart for us anyway, and we do not want to have a potentially
3718 # function-converted external order_by
3719 # ( there is an explicit if ( collapse && _grouped_by_distinct ) check in DBIHacks )
3720 $attrs->{order_by} = $new_order unless $attrs->{collapse};
3725 # generate selections based on the prefetch helper
3728 $self->throw_exception("Unable to prefetch, resultset contains an unnamed selector $attrs->{_dark_selector}{string}")
3729 if $attrs->{_dark_selector};
3731 # this is a separate structure (we don't look in {from} directly)
3732 # as the resolver needs to shift things off the lists to work
3733 # properly (identical-prefetches on different branches)
3734 my $joined_node_aliases_map = {};
3735 if (ref $attrs->{from} eq 'ARRAY') {
3737 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3739 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3740 next unless $j->[0]{-alias};
3741 next unless $j->[0]{-join_path};
3742 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3744 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3746 my $p = $joined_node_aliases_map;
3747 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3748 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3752 ( push @{$attrs->{select}}, $_->[0] ) and ( push @{$attrs->{as}}, $_->[1] )
3753 for $source->_resolve_selection_from_prefetch( $prefetch, $joined_node_aliases_map );
3757 $attrs->{_simple_passthrough_construction} = !(
3760 grep { $_ =~ /\./ } @{$attrs->{as}}
3764 # if both page and offset are specified, produce a combined offset
3765 # even though it doesn't make much sense, this is what pre 081xx has
3767 if (my $page = delete $attrs->{page}) {
3769 ($attrs->{rows} * ($page - 1))
3771 ($attrs->{offset} || 0)
3775 return $self->{_attrs} = $attrs;
3779 my ($self, $attr) = @_;
3781 if (ref $attr eq 'HASH') {
3782 return $self->_rollout_hash($attr);
3783 } elsif (ref $attr eq 'ARRAY') {
3784 return $self->_rollout_array($attr);
3790 sub _rollout_array {
3791 my ($self, $attr) = @_;
3794 foreach my $element (@{$attr}) {
3795 if (ref $element eq 'HASH') {
3796 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3797 } elsif (ref $element eq 'ARRAY') {
3798 # XXX - should probably recurse here
3799 push( @rolled_array, @{$self->_rollout_array($element)} );
3801 push( @rolled_array, $element );
3804 return \@rolled_array;
3808 my ($self, $attr) = @_;
3811 foreach my $key (keys %{$attr}) {
3812 push( @rolled_array, { $key => $attr->{$key} } );
3814 return \@rolled_array;
3817 sub _calculate_score {
3818 my ($self, $a, $b) = @_;
3820 if (defined $a xor defined $b) {
3823 elsif (not defined $a) {
3827 if (ref $b eq 'HASH') {
3828 my ($b_key) = keys %{$b};
3829 $b_key = '' if ! defined $b_key;
3830 if (ref $a eq 'HASH') {
3831 my ($a_key) = keys %{$a};
3832 $a_key = '' if ! defined $a_key;
3833 if ($a_key eq $b_key) {
3834 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3839 return ($a eq $b_key) ? 1 : 0;
3842 if (ref $a eq 'HASH') {
3843 my ($a_key) = keys %{$a};
3844 return ($b eq $a_key) ? 1 : 0;
3846 return ($b eq $a) ? 1 : 0;
3851 sub _merge_joinpref_attr {
3852 my ($self, $orig, $import) = @_;
3854 return $import unless defined($orig);
3855 return $orig unless defined($import);
3857 $orig = $self->_rollout_attr($orig);
3858 $import = $self->_rollout_attr($import);
3861 foreach my $import_element ( @{$import} ) {
3862 # find best candidate from $orig to merge $b_element into
3863 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3864 foreach my $orig_element ( @{$orig} ) {
3865 my $score = $self->_calculate_score( $orig_element, $import_element );
3866 if ($score > $best_candidate->{score}) {
3867 $best_candidate->{position} = $position;
3868 $best_candidate->{score} = $score;
3872 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3873 $import_key = '' if not defined $import_key;
3875 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3876 push( @{$orig}, $import_element );
3878 my $orig_best = $orig->[$best_candidate->{position}];
3879 # merge orig_best and b_element together and replace original with merged
3880 if (ref $orig_best ne 'HASH') {
3881 $orig->[$best_candidate->{position}] = $import_element;
3882 } elsif (ref $import_element eq 'HASH') {
3883 my ($key) = keys %{$orig_best};
3884 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3887 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3890 return @$orig ? $orig : ();
3898 require Hash::Merge;
3899 my $hm = Hash::Merge->new;
3901 $hm->specify_behavior({
3904 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
3906 if ($defl xor $defr) {
3907 return [ $defl ? $_[0] : $_[1] ];
3912 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
3916 return [$_[0], $_[1]];
3920 return $_[1] if !defined $_[0];
3921 return $_[1] if __HM_DEDUP and grep { $_ eq $_[0] } @{$_[1]};
3922 return [$_[0], @{$_[1]}]
3925 return [] if !defined $_[0] and !keys %{$_[1]};
3926 return [ $_[1] ] if !defined $_[0];
3927 return [ $_[0] ] if !keys %{$_[1]};
3928 return [$_[0], $_[1]]
3933 return $_[0] if !defined $_[1];
3934 return $_[0] if __HM_DEDUP and grep { $_ eq $_[1] } @{$_[0]};
3935 return [@{$_[0]}, $_[1]]
3938 my @ret = @{$_[0]} or return $_[1];
3939 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
3940 my %idx = map { $_ => 1 } @ret;
3941 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
3945 return [ $_[1] ] if ! @{$_[0]};
3946 return $_[0] if !keys %{$_[1]};
3947 return $_[0] if __HM_DEDUP and grep { $_ eq $_[1] } @{$_[0]};
3948 return [ @{$_[0]}, $_[1] ];
3953 return [] if !keys %{$_[0]} and !defined $_[1];
3954 return [ $_[0] ] if !defined $_[1];
3955 return [ $_[1] ] if !keys %{$_[0]};
3956 return [$_[0], $_[1]]
3959 return [] if !keys %{$_[0]} and !@{$_[1]};
3960 return [ $_[0] ] if !@{$_[1]};
3961 return $_[1] if !keys %{$_[0]};
3962 return $_[1] if __HM_DEDUP and grep { $_ eq $_[0] } @{$_[1]};
3963 return [ $_[0], @{$_[1]} ];
3966 return [] if !keys %{$_[0]} and !keys %{$_[1]};
3967 return [ $_[0] ] if !keys %{$_[1]};
3968 return [ $_[1] ] if !keys %{$_[0]};
3969 return [ $_[0] ] if $_[0] eq $_[1];
3970 return [ $_[0], $_[1] ];
3973 } => 'DBIC_RS_ATTR_MERGER');
3977 return $hm->merge ($_[1], $_[2]);
3981 sub STORABLE_freeze {
3982 my ($self, $cloning) = @_;
3983 my $to_serialize = { %$self };
3985 # A cursor in progress can't be serialized (and would make little sense anyway)
3986 # the parser can be regenerated (and can't be serialized)
3987 delete @{$to_serialize}{qw/cursor _row_parser _result_inflator/};
3989 # nor is it sensical to store a not-yet-fired-count pager
3990 if ($to_serialize->{pager} and ref $to_serialize->{pager}{total_entries} eq 'CODE') {
3991 delete $to_serialize->{pager};
3994 Storable::nfreeze($to_serialize);
3997 # need this hook for symmetry
3999 my ($self, $cloning, $serialized) = @_;
4001 %$self = %{ Storable::thaw($serialized) };
4007 =head2 throw_exception
4009 See L<DBIx::Class::Schema/throw_exception> for details.
4013 sub throw_exception {
4016 if (ref $self and my $rsrc = $self->result_source) {
4017 $rsrc->throw_exception(@_)
4020 DBIx::Class::Exception->throw(@_);
4028 # XXX: FIXME: Attributes docs need clearing up
4032 Attributes are used to refine a ResultSet in various ways when
4033 searching for data. They can be passed to any method which takes an
4034 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
4037 Default attributes can be set on the result class using
4038 L<DBIx::Class::ResultSource/resultset_attributes>. (Please read
4039 the CAVEATS on that feature before using it!)
4041 These are in no particular order:
4047 =item Value: ( $order_by | \@order_by | \%order_by )
4051 Which column(s) to order the results by.
4053 [The full list of suitable values is documented in
4054 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
4057 If a single column name, or an arrayref of names is supplied, the
4058 argument is passed through directly to SQL. The hashref syntax allows
4059 for connection-agnostic specification of ordering direction:
4061 For descending order:
4063 order_by => { -desc => [qw/col1 col2 col3/] }
4065 For explicit ascending order:
4067 order_by => { -asc => 'col' }
4069 The old scalarref syntax (i.e. order_by => \'year DESC') is still
4070 supported, although you are strongly encouraged to use the hashref
4071 syntax as outlined above.
4077 =item Value: \@columns | \%columns | $column
4081 Shortcut to request a particular set of columns to be retrieved. Each
4082 column spec may be a string (a table column name), or a hash (in which
4083 case the key is the C<as> value, and the value is used as the C<select>
4084 expression). Adds the L</current_source_alias> onto the start of any column without a C<.> in
4085 it and sets C<select> from that, then auto-populates C<as> from
4086 C<select> as normal. (You may also use the C<cols> attribute, as in
4087 earlier versions of DBIC, but this is deprecated)
4089 Essentially C<columns> does the same as L</select> and L</as>.
4091 columns => [ 'some_column', { dbic_slot => 'another_column' } ]
4095 select => [qw(some_column another_column)],
4096 as => [qw(some_column dbic_slot)]
4098 If you want to individually retrieve related columns (in essence perform
4099 manual L</prefetch>) you have to make sure to specify the correct inflation slot
4100 chain such that it matches existing relationships:
4102 my $rs = $schema->resultset('Artist')->search({}, {
4103 # required to tell DBIC to collapse has_many relationships
4105 join => { cds => 'tracks' },
4107 'cds.cdid' => 'cds.cdid',
4108 'cds.tracks.title' => 'tracks.title',
4112 Like elsewhere, literal SQL or literal values can be included by using a
4113 scalar reference or a literal bind value, and these values will be available
4114 in the result with C<get_column> (see also
4115 L<SQL::Abstract/Literal SQL and value type operators>):
4117 # equivalent SQL: SELECT 1, 'a string', IF(my_column,?,?) ...
4118 # bind values: $true_value, $false_value
4122 bar => \q{'a string'},
4123 baz => \[ 'IF(my_column,?,?)', $true_value, $false_value ],
4129 B<NOTE:> You B<MUST> explicitly quote C<'+columns'> when using this attribute.
4130 Not doing so causes Perl to incorrectly interpret C<+columns> as a bareword
4131 with a unary plus operator before it, which is the same as simply C<columns>.
4135 =item Value: \@extra_columns
4139 Indicates additional columns to be selected from storage. Works the same as
4140 L</columns> but adds columns to the current selection. (You may also use the
4141 C<include_columns> attribute, as in earlier versions of DBIC, but this is
4144 $schema->resultset('CD')->search(undef, {
4145 '+columns' => ['artist.name'],
4149 would return all CDs and include a 'name' column to the information
4150 passed to object inflation. Note that the 'artist' is the name of the
4151 column (or relationship) accessor, and 'name' is the name of the column
4152 accessor in the related table.
4158 =item Value: \@select_columns
4162 Indicates which columns should be selected from the storage. You can use
4163 column names, or in the case of RDBMS back ends, function or stored procedure
4166 $rs = $schema->resultset('Employee')->search(undef, {
4169 { count => 'employeeid' },
4170 { max => { length => 'name' }, -as => 'longest_name' }
4175 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
4177 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
4178 use L</select>, to instruct DBIx::Class how to store the result of the column.
4180 Also note that the L</as> attribute has B<nothing to do> with the SQL-side
4181 C<AS> identifier aliasing. You B<can> alias a function (so you can use it e.g.
4182 in an C<ORDER BY> clause), however this is done via the C<-as> B<select
4183 function attribute> supplied as shown in the example above.
4187 B<NOTE:> You B<MUST> explicitly quote C<'+select'> when using this attribute.
4188 Not doing so causes Perl to incorrectly interpret C<+select> as a bareword
4189 with a unary plus operator before it, which is the same as simply C<select>.
4193 =item Value: \@extra_select_columns
4197 Indicates additional columns to be selected from storage. Works the same as
4198 L</select> but adds columns to the current selection, instead of specifying
4199 a new explicit list.
4205 =item Value: \@inflation_names
4209 Indicates DBIC-side names for object inflation. That is L</as> indicates the
4210 slot name in which the column value will be stored within the
4211 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
4212 identifier by the C<get_column> method (or via the object accessor B<if one
4213 with the same name already exists>) as shown below.
4215 The L</as> attribute has B<nothing to do> with the SQL-side identifier
4216 aliasing C<AS>. See L</select> for details.
4218 $rs = $schema->resultset('Employee')->search(undef, {
4221 { count => 'employeeid' },
4222 { max => { length => 'name' }, -as => 'longest_name' }
4231 If the object against which the search is performed already has an accessor
4232 matching a column name specified in C<as>, the value can be retrieved using
4233 the accessor as normal:
4235 my $name = $employee->name();
4237 If on the other hand an accessor does not exist in the object, you need to
4238 use C<get_column> instead:
4240 my $employee_count = $employee->get_column('employee_count');
4242 You can create your own accessors if required - see
4243 L<DBIx::Class::Manual::Cookbook> for details.
4247 B<NOTE:> You B<MUST> explicitly quote C<'+as'> when using this attribute.
4248 Not doing so causes Perl to incorrectly interpret C<+as> as a bareword
4249 with a unary plus operator before it, which is the same as simply C<as>.
4253 =item Value: \@extra_inflation_names
4257 Indicates additional inflation names for selectors added via L</+select>. See L</as>.
4263 =item Value: ($rel_name | \@rel_names | \%rel_names)
4267 Contains a list of relationships that should be joined for this query. For
4270 # Get CDs by Nine Inch Nails
4271 my $rs = $schema->resultset('CD')->search(
4272 { 'artist.name' => 'Nine Inch Nails' },
4273 { join => 'artist' }
4276 Can also contain a hash reference to refer to the other relation's relations.
4279 package MyApp::Schema::Track;
4280 use base qw/DBIx::Class/;
4281 __PACKAGE__->table('track');
4282 __PACKAGE__->add_columns(qw/trackid cd position title/);
4283 __PACKAGE__->set_primary_key('trackid');
4284 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
4287 # In your application
4288 my $rs = $schema->resultset('Artist')->search(
4289 { 'track.title' => 'Teardrop' },
4291 join => { cd => 'track' },
4292 order_by => 'artist.name',
4296 You need to use the relationship (not the table) name in conditions,
4297 because they are aliased as such. The current table is aliased as "me", so
4298 you need to use me.column_name in order to avoid ambiguity. For example:
4300 # Get CDs from 1984 with a 'Foo' track
4301 my $rs = $schema->resultset('CD')->search(
4304 'tracks.name' => 'Foo'
4306 { join => 'tracks' }
4309 If the same join is supplied twice, it will be aliased to <rel>_2 (and
4310 similarly for a third time). For e.g.
4312 my $rs = $schema->resultset('Artist')->search({
4313 'cds.title' => 'Down to Earth',
4314 'cds_2.title' => 'Popular',
4316 join => [ qw/cds cds/ ],
4319 will return a set of all artists that have both a cd with title 'Down
4320 to Earth' and a cd with title 'Popular'.
4322 If you want to fetch related objects from other tables as well, see L</prefetch>
4325 NOTE: An internal join-chain pruner will discard certain joins while
4326 constructing the actual SQL query, as long as the joins in question do not
4327 affect the retrieved result. This for example includes 1:1 left joins
4328 that are not part of the restriction specification (WHERE/HAVING) nor are
4329 a part of the query selection.
4331 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
4337 =item Value: (0 | 1)
4341 When set to a true value, indicates that any rows fetched from joined has_many
4342 relationships are to be aggregated into the corresponding "parent" object. For
4343 example, the resultset:
4345 my $rs = $schema->resultset('CD')->search({}, {
4346 '+columns' => [ qw/ tracks.title tracks.position / ],
4351 While executing the following query:
4353 SELECT me.*, tracks.title, tracks.position
4355 LEFT JOIN track tracks
4356 ON tracks.cdid = me.cdid
4358 Will return only as many objects as there are rows in the CD source, even
4359 though the result of the query may span many rows. Each of these CD objects
4360 will in turn have multiple "Track" objects hidden behind the has_many
4361 generated accessor C<tracks>. Without C<< collapse => 1 >>, the return values
4362 of this resultset would be as many CD objects as there are tracks (a "Cartesian
4363 product"), with each CD object containing exactly one of all fetched Track data.
4365 When a collapse is requested on a non-ordered resultset, an order by some
4366 unique part of the main source (the left-most table) is inserted automatically.
4367 This is done so that the resultset is allowed to be "lazy" - calling
4368 L<< $rs->next|/next >> will fetch only as many rows as it needs to build the next
4369 object with all of its related data.
4371 If an L</order_by> is already declared, and orders the resultset in a way that
4372 makes collapsing as described above impossible (e.g. C<< ORDER BY
4373 has_many_rel.column >> or C<ORDER BY RANDOM()>), DBIC will automatically
4374 switch to "eager" mode and slurp the entire resultset before constructing the
4375 first object returned by L</next>.
4377 Setting this attribute on a resultset that does not join any has_many
4378 relations is a no-op.
4380 For a more in-depth discussion, see L</PREFETCHING>.
4386 =item Value: ($rel_name | \@rel_names | \%rel_names)
4390 This attribute is a shorthand for specifying a L</join> spec, adding all
4391 columns from the joined related sources as L</+columns> and setting
4392 L</collapse> to a true value. It can be thought of as a rough B<superset>
4393 of the L</join> attribute.
4395 For example, the following two queries are equivalent:
4397 my $rs = $schema->resultset('Artist')->search({}, {
4398 prefetch => { cds => ['genre', 'tracks' ] },
4403 my $rs = $schema->resultset('Artist')->search({}, {
4404 join => { cds => ['genre', 'tracks' ] },
4408 { +{ "cds.$_" => "cds.$_" } }
4409 $schema->source('Artist')->related_source('cds')->columns
4412 { +{ "cds.genre.$_" => "genre.$_" } }
4413 $schema->source('Artist')->related_source('cds')->related_source('genre')->columns
4416 { +{ "cds.tracks.$_" => "tracks.$_" } }
4417 $schema->source('Artist')->related_source('cds')->related_source('tracks')->columns
4422 Both producing the following SQL:
4424 SELECT me.artistid, me.name, me.rank, me.charfield,
4425 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track,
4426 genre.genreid, genre.name,
4427 tracks.trackid, tracks.cd, tracks.position, tracks.title, tracks.last_updated_on, tracks.last_updated_at
4430 ON cds.artist = me.artistid
4431 LEFT JOIN genre genre
4432 ON genre.genreid = cds.genreid
4433 LEFT JOIN track tracks
4434 ON tracks.cd = cds.cdid
4435 ORDER BY me.artistid
4437 While L</prefetch> implies a L</join>, it is ok to mix the two together, as
4438 the arguments are properly merged and generally do the right thing. For
4439 example, you may want to do the following:
4441 my $artists_and_cds_without_genre = $schema->resultset('Artist')->search(
4442 { 'genre.genreid' => undef },
4444 join => { cds => 'genre' },
4449 Which generates the following SQL:
4451 SELECT me.artistid, me.name, me.rank, me.charfield,
4452 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track
4455 ON cds.artist = me.artistid
4456 LEFT JOIN genre genre
4457 ON genre.genreid = cds.genreid
4458 WHERE genre.genreid IS NULL
4459 ORDER BY me.artistid
4461 For a more in-depth discussion, see L</PREFETCHING>.
4467 =item Value: $source_alias
4471 Sets the source alias for the query. Normally, this defaults to C<me>, but
4472 nested search queries (sub-SELECTs) might need specific aliases set to
4473 reference inner queries. For example:
4476 ->related_resultset('CDs')
4477 ->related_resultset('Tracks')
4479 'track.id' => { -ident => 'none_search.id' },
4483 my $ids = $self->search({
4486 alias => 'none_search',
4487 group_by => 'none_search.id',
4488 })->get_column('id')->as_query;
4490 $self->search({ id => { -in => $ids } })
4492 This attribute is directly tied to L</current_source_alias>.
4502 Makes the resultset paged and specifies the page to retrieve. Effectively
4503 identical to creating a non-pages resultset and then calling ->page($page)
4506 If L</rows> attribute is not specified it defaults to 10 rows per page.
4508 When you have a paged resultset, L</count> will only return the number
4509 of rows in the page. To get the total, use the L</pager> and call
4510 C<total_entries> on it.
4520 Specifies the maximum number of rows for direct retrieval or the number of
4521 rows per page if the page attribute or method is used.
4527 =item Value: $offset
4531 Specifies the (zero-based) row number for the first row to be returned, or the
4532 of the first row of the first page if paging is used.
4534 =head2 software_limit
4538 =item Value: (0 | 1)
4542 When combined with L</rows> and/or L</offset> the generated SQL will not
4543 include any limit dialect stanzas. Instead the entire result will be selected
4544 as if no limits were specified, and DBIC will perform the limit locally, by
4545 artificially advancing and finishing the resulting L</cursor>.
4547 This is the recommended way of performing resultset limiting when no sane RDBMS
4548 implementation is available (e.g.
4549 L<Sybase ASE|DBIx::Class::Storage::DBI::Sybase::ASE> using the
4550 L<Generic Sub Query|DBIx::Class::SQLMaker::LimitDialects/GenericSubQ> hack)
4556 =item Value: \@columns
4560 A arrayref of columns to group by. Can include columns of joined tables.
4562 group_by => [qw/ column1 column2 ... /]
4568 =item Value: $condition
4572 The HAVING operator specifies a B<secondary> condition applied to the set
4573 after the grouping calculations have been done. In other words it is a
4574 constraint just like L</where> (and accepting the same
4575 L<SQL::Abstract syntax|SQL::Abstract/WHERE CLAUSES>) applied to the data
4576 as it exists after GROUP BY has taken place. Specifying L</having> without
4577 L</group_by> is a logical mistake, and a fatal error on most RDBMS engines.
4581 having => { 'count_employee' => { '>=', 100 } }
4583 or with an in-place function in which case literal SQL is required:
4585 having => \[ 'count(employee) >= ?', 100 ]
4591 =item Value: (0 | 1)
4595 Set to 1 to automatically generate a L</group_by> clause based on the selection
4596 (including intelligent handling of L</order_by> contents). Note that the group
4597 criteria calculation takes place over the B<final> selection. This includes
4598 any L</+columns>, L</+select> or L</order_by> additions in subsequent
4599 L</search> calls, and standalone columns selected via
4600 L<DBIx::Class::ResultSetColumn> (L</get_column>). A notable exception are the
4601 extra selections specified via L</prefetch> - such selections are explicitly
4602 excluded from group criteria calculations.
4604 If the final ResultSet also explicitly defines a L</group_by> attribute, this
4605 setting is ignored and an appropriate warning is issued.
4609 Adds extra conditions to the resultset, combined with the preexisting C<WHERE>
4610 conditions, same as the B<first> argument to the L<search operator|/search>
4612 # only return rows WHERE deleted IS NULL for all searches
4613 __PACKAGE__->resultset_attributes({ where => { deleted => undef } });
4615 Note that the above example is
4616 L<strongly discouraged|DBIx::Class::ResultSource/resultset_attributes>.
4620 Set to 1 to cache search results. This prevents extra SQL queries if you
4621 revisit rows in your ResultSet:
4623 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
4625 while( my $artist = $resultset->next ) {
4629 $resultset->first; # without cache, this would issue a query
4631 By default, searches are not cached.
4633 For more examples of using these attributes, see
4634 L<DBIx::Class::Manual::Cookbook>.
4640 =item Value: ( 'update' | 'shared' | \$scalar )
4644 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT
4645 ... FOR SHARED. If \$scalar is passed, this is taken directly and embedded in the
4650 DBIx::Class supports arbitrary related data prefetching from multiple related
4651 sources. Any combination of relationship types and column sets are supported.
4652 If L<collapsing|/collapse> is requested, there is an additional requirement of
4653 selecting enough data to make every individual object uniquely identifiable.
4655 Here are some more involved examples, based on the following relationship map:
4658 My::Schema::CD->belongs_to( artist => 'My::Schema::Artist' );
4659 My::Schema::CD->might_have( liner_note => 'My::Schema::LinerNotes' );
4660 My::Schema::CD->has_many( tracks => 'My::Schema::Track' );
4662 My::Schema::Artist->belongs_to( record_label => 'My::Schema::RecordLabel' );
4664 My::Schema::Track->has_many( guests => 'My::Schema::Guest' );
4668 my $rs = $schema->resultset('Tag')->search(
4677 The initial search results in SQL like the following:
4679 SELECT tag.*, cd.*, artist.* FROM tag
4680 JOIN cd ON tag.cd = cd.cdid
4681 JOIN artist ON cd.artist = artist.artistid
4683 L<DBIx::Class> has no need to go back to the database when we access the
4684 C<cd> or C<artist> relationships, which saves us two SQL statements in this
4687 Simple prefetches will be joined automatically, so there is no need
4688 for a C<join> attribute in the above search.
4690 The L</prefetch> attribute can be used with any of the relationship types
4691 and multiple prefetches can be specified together. Below is a more complex
4692 example that prefetches a CD's artist, its liner notes (if present),
4693 the cover image, the tracks on that CD, and the guests on those
4696 my $rs = $schema->resultset('CD')->search(
4700 { artist => 'record_label'}, # belongs_to => belongs_to
4701 'liner_note', # might_have
4702 'cover_image', # has_one
4703 { tracks => 'guests' }, # has_many => has_many
4708 This will produce SQL like the following:
4710 SELECT cd.*, artist.*, record_label.*, liner_note.*, cover_image.*,
4714 ON artist.artistid = me.artistid
4715 JOIN record_label record_label
4716 ON record_label.labelid = artist.labelid
4717 LEFT JOIN track tracks
4718 ON tracks.cdid = me.cdid
4719 LEFT JOIN guest guests
4720 ON guests.trackid = track.trackid
4721 LEFT JOIN liner_notes liner_note
4722 ON liner_note.cdid = me.cdid
4723 JOIN cd_artwork cover_image
4724 ON cover_image.cdid = me.cdid
4727 Now the C<artist>, C<record_label>, C<liner_note>, C<cover_image>,
4728 C<tracks>, and C<guests> of the CD will all be available through the
4729 relationship accessors without the need for additional queries to the
4734 Prefetch does a lot of deep magic. As such, it may not behave exactly
4735 as you might expect.
4741 Prefetch uses the L</cache> to populate the prefetched relationships. This
4742 may or may not be what you want.
4746 If you specify a condition on a prefetched relationship, ONLY those
4747 rows that match the prefetched condition will be fetched into that relationship.
4748 This means that adding prefetch to a search() B<may alter> what is returned by
4749 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
4751 my $artist_rs = $schema->resultset('Artist')->search({
4757 my $count = $artist_rs->first->cds->count;
4759 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
4761 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
4763 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
4765 That cmp_ok() may or may not pass depending on the datasets involved. In other
4766 words the C<WHERE> condition would apply to the entire dataset, just like
4767 it would in regular SQL. If you want to add a condition only to the "right side"
4768 of a C<LEFT JOIN> - consider declaring and using a L<relationship with a custom
4769 condition|DBIx::Class::Relationship::Base/condition>
4773 =head1 DBIC BIND VALUES
4775 Because DBIC may need more information to bind values than just the column name
4776 and value itself, it uses a special format for both passing and receiving bind
4777 values. Each bind value should be composed of an arrayref of
4778 C<< [ \%args => $val ] >>. The format of C<< \%args >> is currently:
4784 If present (in any form), this is what is being passed directly to bind_param.
4785 Note that different DBD's expect different bind args. (e.g. DBD::SQLite takes
4786 a single numerical type, while DBD::Pg takes a hashref if bind options.)
4788 If this is specified, all other bind options described below are ignored.
4792 If present, this is used to infer the actual bind attribute by passing to
4793 C<< $resolved_storage->bind_attribute_by_data_type() >>. Defaults to the
4794 "data_type" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>.
4796 Note that the data type is somewhat freeform (hence the sqlt_ prefix);
4797 currently drivers are expected to "Do the Right Thing" when given a common
4798 datatype name. (Not ideal, but that's what we got at this point.)
4802 Currently used to correctly allocate buffers for bind_param_inout().
4803 Defaults to "size" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>,
4804 or to a sensible value based on the "data_type".
4808 Used to fill in missing sqlt_datatype and sqlt_size attributes (if they are
4809 explicitly specified they are never overridden). Also used by some weird DBDs,
4810 where the column name should be available at bind_param time (e.g. Oracle).
4814 For backwards compatibility and convenience, the following shortcuts are
4817 [ $name => $val ] === [ { dbic_colname => $name }, $val ]
4818 [ \$dt => $val ] === [ { sqlt_datatype => $dt }, $val ]
4819 [ undef, $val ] === [ {}, $val ]
4820 $val === [ {}, $val ]
4822 =head1 FURTHER QUESTIONS?
4824 Check the list of L<additional DBIC resources|DBIx::Class/GETTING HELP/SUPPORT>.
4826 =head1 COPYRIGHT AND LICENSE
4828 This module is free software L<copyright|DBIx::Class/COPYRIGHT AND LICENSE>
4829 by the L<DBIx::Class (DBIC) authors|DBIx::Class/AUTHORS>. You can
4830 redistribute it and/or modify it under the same terms as the
4831 L<DBIx::Class library|DBIx::Class/COPYRIGHT AND LICENSE>.