1 package DBIx::Class::ResultSet;
6 use base 'DBIx::Class';
9 use DBIx::Class::ResultSetColumn;
10 use DBIx::Class::ResultClass::HashRefInflator;
11 use Scalar::Util qw( blessed reftype );
12 use SQL::Abstract 'is_literal_value';
13 use DBIx::Class::_Util qw(
14 dbic_internal_try dbic_internal_catch dump_value emit_loud_diag
15 fail_on_internal_wantarray fail_on_internal_call UNRESOLVABLE_CONDITION
17 use DBIx::Class::SQLMaker::Util qw( normalize_sqla_condition extract_equality_conditions );
18 use DBIx::Class::ResultSource::FromSpec::Util 'find_join_path_to_alias';
21 # De-duplication in _merge_attr() is disabled, but left in for reference
22 # (the merger is used for other things that ought not to be de-duped)
23 *__HM_DEDUP = sub () { 0 };
33 # this is real - CDBICompat overrides it with insanity
34 # yes, prototype won't matter, but that's for now ;)
37 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class result_source/);
41 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
45 my $users_rs = $schema->resultset('User');
46 while( $user = $users_rs->next) {
47 print $user->username;
50 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
51 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
55 A ResultSet is an object which stores a set of conditions representing
56 a query. It is the backbone of DBIx::Class (i.e. the really
57 important/useful bit).
59 No SQL is executed on the database when a ResultSet is created, it
60 just stores all the conditions needed to create the query.
62 A basic ResultSet representing the data of an entire table is returned
63 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
64 L<Source|DBIx::Class::Manual::Glossary/ResultSource> name.
66 my $users_rs = $schema->resultset('User');
68 A new ResultSet is returned from calling L</search> on an existing
69 ResultSet. The new one will contain all the conditions of the
70 original, plus any new conditions added in the C<search> call.
72 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
73 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
76 The query that the ResultSet represents is B<only> executed against
77 the database when these methods are called:
78 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
80 If a resultset is used in a numeric context it returns the L</count>.
81 However, if it is used in a boolean context it is B<always> true. So if
82 you want to check if a resultset has any results, you must use C<if $rs
87 =head2 Chaining resultsets
89 Let's say you've got a query that needs to be run to return some data
90 to the user. But, you have an authorization system in place that
91 prevents certain users from seeing certain information. So, you want
92 to construct the basic query in one method, but add constraints to it in
97 my $request = $self->get_request; # Get a request object somehow.
98 my $schema = $self->result_source->schema;
100 my $cd_rs = $schema->resultset('CD')->search({
101 title => $request->param('title'),
102 year => $request->param('year'),
105 $cd_rs = $self->apply_security_policy( $cd_rs );
107 return $cd_rs->all();
110 sub apply_security_policy {
119 =head3 Resolving conditions and attributes
121 When a resultset is chained from another resultset (e.g.:
122 C<< my $new_rs = $old_rs->search(\%extra_cond, \%attrs) >>), conditions
123 and attributes with the same keys need resolving.
125 If any of L</columns>, L</select>, L</as> are present, they reset the
126 original selection, and start the selection "clean".
128 The L</join>, L</prefetch>, L</+columns>, L</+select>, L</+as> attributes
129 are merged into the existing ones from the original resultset.
131 The L</where> and L</having> attributes, and any search conditions, are
132 merged with an SQL C<AND> to the existing condition from the original
135 All other attributes are overridden by any new ones supplied in the
138 =head2 Multiple queries
140 Since a resultset just defines a query, you can do all sorts of
141 things with it with the same object.
143 # Don't hit the DB yet.
144 my $cd_rs = $schema->resultset('CD')->search({
145 title => 'something',
149 # Each of these hits the DB individually.
150 my $count = $cd_rs->count;
151 my $most_recent = $cd_rs->get_column('date_released')->max();
152 my @records = $cd_rs->all;
154 And it's not just limited to SELECT statements.
160 $cd_rs->create({ artist => 'Fred' });
162 Which is the same as:
164 $schema->resultset('CD')->create({
165 title => 'something',
170 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
172 =head2 Custom ResultSet classes
174 To add methods to your resultsets, you can subclass L<DBIx::Class::ResultSet>, similar to:
176 package MyApp::Schema::ResultSet::User;
181 use base 'DBIx::Class::ResultSet';
185 $self->search({ $self->current_source_alias . '.active' => 1 });
190 $self->search({ $self->current_source_alias . '.verified' => 0 });
193 sub created_n_days_ago {
194 my ($self, $days_ago) = @_;
196 $self->current_source_alias . '.create_date' => {
198 $self->result_source->schema->storage->datetime_parser->format_datetime(
199 DateTime->now( time_zone => 'UTC' )->subtract( days => $days_ago )
204 sub users_to_warn { shift->active->unverified->created_n_days_ago(7) }
208 See L<DBIx::Class::Schema/load_namespaces> on how DBIC can discover and
209 automatically attach L<Result|DBIx::Class::Manual::ResultClass>-specific
210 L<ResulSet|DBIx::Class::ResultSet> classes.
212 =head3 ResultSet subclassing with Moose and similar constructor-providers
214 Using L<Moose> or L<Moo> in your ResultSet classes is usually overkill, but
215 you may find it useful if your ResultSets contain a lot of business logic
216 (e.g. C<has xml_parser>, C<has json>, etc) or if you just prefer to organize
219 In order to write custom ResultSet classes with L<Moo> you need to use the
220 following template. The L<BUILDARGS|Moo/BUILDARGS> is necessary due to the
221 unusual signature of the L<constructor provided by DBIC
222 |DBIx::Class::ResultSet/new> C<< ->new($source, \%args) >>.
225 extends 'DBIx::Class::ResultSet';
226 sub BUILDARGS { $_[2] } # ::RS::new() expects my ($class, $rsrc, $args) = @_
232 If you want to build your custom ResultSet classes with L<Moose>, you need
233 a similar, though a little more elaborate template in order to interface the
234 inlining of the L<Moose>-provided
235 L<object constructor|Moose::Manual::Construction/WHERE'S THE CONSTRUCTOR?>,
238 package MyApp::Schema::ResultSet::User;
241 use MooseX::NonMoose;
242 extends 'DBIx::Class::ResultSet';
244 sub BUILDARGS { $_[2] } # ::RS::new() expects my ($class, $rsrc, $args) = @_
248 __PACKAGE__->meta->make_immutable;
252 The L<MooseX::NonMoose> is necessary so that the L<Moose> constructor does not
253 entirely overwrite the DBIC one (in contrast L<Moo> does this automatically).
254 Alternatively, you can skip L<MooseX::NonMoose> and get by with just L<Moose>
257 __PACKAGE__->meta->make_immutable(inline_constructor => 0);
265 =item Arguments: L<$source|DBIx::Class::ResultSource>, L<\%attrs?|/ATTRIBUTES>
267 =item Return Value: L<$resultset|/search>
271 The resultset constructor. Takes a source object (usually a
272 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
273 L</ATTRIBUTES> below). Does not perform any queries -- these are
274 executed as needed by the other methods.
276 Generally you never construct a resultset manually. Instead you get one
278 C<< $schema->L<resultset|DBIx::Class::Schema/resultset>('$source_name') >>
279 or C<< $another_resultset->L<search|/search>(...) >> (the later called in
282 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
288 If called on an object, proxies to L</new_result> instead, so
290 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
292 will return a CD object, not a ResultSet, and is equivalent to:
294 my $cd = $schema->resultset('CD')->new_result({ title => 'Spoon' });
296 Please also keep in mind that many internals call L</new_result> directly,
297 so overloading this method with the idea of intercepting new result object
298 creation B<will not work>. See also warning pertaining to L</create>.
308 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
309 return $class->new_result(@_);
312 my ($source, $attrs) = @_;
313 $source = $source->resolve
314 if $source->isa('DBIx::Class::ResultSourceHandle');
316 $attrs = { %{$attrs||{}} };
317 delete @{$attrs}{qw(_last_sqlmaker_alias_map _simple_passthrough_construction)};
319 if ($attrs->{page}) {
320 $attrs->{rows} ||= 10;
323 $attrs->{alias} ||= 'me';
326 result_source => $source,
327 cond => $attrs->{where},
332 # if there is a dark selector, this means we are already in a
333 # chain and the cleanup/sanification was taken care of by
335 $self->_normalize_selection($attrs)
336 unless $attrs->{_dark_selector};
339 $attrs->{result_class} || $source->result_class
349 =item Arguments: L<$cond|DBIx::Class::SQLMaker> | undef, L<\%attrs?|/ATTRIBUTES>
351 =item Return Value: $resultset (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
355 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
356 my $new_rs = $cd_rs->search({ year => 2005 });
358 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
359 # year = 2005 OR year = 2004
361 In list context, C<< ->all() >> is called implicitly on the resultset, thus
362 returning a list of L<result|DBIx::Class::Manual::ResultClass> objects instead.
363 To avoid that, use L</search_rs>.
365 If you need to pass in additional attributes but no additional condition,
366 call it as C<search(undef, \%attrs)>.
368 # "SELECT name, artistid FROM $artist_table"
369 my @all_artists = $schema->resultset('Artist')->search(undef, {
370 columns => [qw/name artistid/],
373 For a list of attributes that can be passed to C<search>, see
374 L</ATTRIBUTES>. For more examples of using this function, see
375 L<Searching|DBIx::Class::Manual::Cookbook/SEARCHING>. For a complete
376 documentation for the first argument, see L<SQL::Abstract/"WHERE CLAUSES">
377 and its extension L<DBIx::Class::SQLMaker>.
379 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
383 Note that L</search> does not process/deflate any of the values passed in the
384 L<SQL::Abstract>-compatible search condition structure. This is unlike other
385 condition-bound methods L</new_result>, L</create> and L</find>. The user must ensure
386 manually that any value passed to this method will stringify to something the
387 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
388 objects, for more info see:
389 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
395 my $rs = $self->search_rs( @_ );
398 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_WANTARRAY and my $sog = fail_on_internal_wantarray;
401 elsif (defined wantarray) {
405 # we can be called by a relationship helper, which in
406 # turn may be called in void context due to some braindead
407 # overload or whatever else the user decided to be clever
408 # at this particular day. Thus limit the exception to
409 # external code calls only
410 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense')
411 if (caller)[0] !~ /^\QDBIx::Class::/;
421 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
423 =item Return Value: L<$resultset|/search>
427 This method does the same exact thing as search() except it will
428 always return a resultset, even in list context.
435 my $rsrc = $self->result_source;
436 my ($call_cond, $call_attrs);
438 # Special-case handling for (undef, undef) or (undef)
439 # Note that (foo => undef) is valid deprecated syntax
440 @_ = () if not scalar grep { defined $_ } @_;
446 # fish out attrs in the ($condref, $attr) case
447 elsif (@_ == 2 and ( ! defined $_[0] or length ref $_[0] ) ) {
448 ($call_cond, $call_attrs) = @_;
451 $self->throw_exception('Odd number of arguments to search')
455 carp_unique 'search( %condition ) is deprecated, use search( \%condition ) instead'
456 unless $rsrc->result_class->isa('DBIx::Class::CDBICompat');
458 for my $i (0 .. $#_) {
460 $self->throw_exception ('All keys in condition key/value pairs must be plain scalars')
461 if (! defined $_[$i] or length ref $_[$i] );
467 # see if we can keep the cache (no $rs changes)
469 my %safe = (alias => 1, cache => 1);
470 if ( ! grep { !$safe{$_} } keys %$call_attrs and (
473 ref $call_cond eq 'HASH' && ! keys %$call_cond
475 ref $call_cond eq 'ARRAY' && ! @$call_cond
477 $cache = $self->get_cache;
480 my $old_attrs = { %{$self->{attrs}} };
481 my ($old_having, $old_where) = delete @{$old_attrs}{qw(having where)};
483 my $new_attrs = { %$old_attrs };
485 # take care of call attrs (only if anything is changing)
486 if ($call_attrs and keys %$call_attrs) {
488 # copy for _normalize_selection
489 $call_attrs = { %$call_attrs };
491 my @selector_attrs = qw/select as columns cols +select +as +columns include_columns/;
493 # reset the current selector list if new selectors are supplied
494 delete @{$old_attrs}{(@selector_attrs, '_dark_selector')}
495 if grep { exists $call_attrs->{$_} } qw(columns cols select as);
497 # Normalize the new selector list (operates on the passed-in attr structure)
498 # Need to do it on every chain instead of only once on _resolved_attrs, in
499 # order to allow detection of empty vs partial 'as'
500 $call_attrs->{_dark_selector} = $old_attrs->{_dark_selector}
501 if $old_attrs->{_dark_selector};
502 $self->_normalize_selection ($call_attrs);
504 # start with blind overwriting merge, exclude selector attrs
505 $new_attrs = { %{$old_attrs}, %{$call_attrs} };
506 delete @{$new_attrs}{@selector_attrs};
508 for (@selector_attrs) {
509 $new_attrs->{$_} = $self->_merge_attr($old_attrs->{$_}, $call_attrs->{$_})
510 if ( exists $old_attrs->{$_} or exists $call_attrs->{$_} );
513 # older deprecated name, use only if {columns} is not there
514 if (my $c = delete $new_attrs->{cols}) {
515 carp_unique( "Resultset attribute 'cols' is deprecated, use 'columns' instead" );
516 if ($new_attrs->{columns}) {
517 carp "Resultset specifies both the 'columns' and the legacy 'cols' attributes - ignoring 'cols'";
520 $new_attrs->{columns} = $c;
525 # join/prefetch use their own crazy merging heuristics
526 foreach my $key (qw/join prefetch/) {
527 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key})
528 if exists $call_attrs->{$key};
531 # stack binds together
532 $new_attrs->{bind} = [ @{ $old_attrs->{bind} || [] }, @{ $call_attrs->{bind} || [] } ];
536 for ($old_where, $call_cond) {
538 $new_attrs->{where} = $self->_stack_cond (
539 $_, $new_attrs->{where}
544 if (defined $old_having) {
545 $new_attrs->{having} = $self->_stack_cond (
546 $old_having, $new_attrs->{having}
550 my $rs = (ref $self)->new($rsrc, $new_attrs);
552 $rs->set_cache($cache) if ($cache);
557 sub _normalize_selection {
558 my ($self, $attrs) = @_;
561 if ( exists $attrs->{include_columns} ) {
562 carp_unique( "Resultset attribute 'include_columns' is deprecated, use '+columns' instead" );
563 $attrs->{'+columns'} = $self->_merge_attr(
564 $attrs->{'+columns'}, delete $attrs->{include_columns}
568 # columns are always placed first, however
570 # Keep the X vs +X separation until _resolved_attrs time - this allows to
571 # delay the decision on whether to use a default select list ($rsrc->columns)
572 # allowing stuff like the remove_columns helper to work
574 # select/as +select/+as pairs need special handling - the amount of select/as
575 # elements in each pair does *not* have to be equal (think multicolumn
576 # selectors like distinct(foo, bar) ). If the selector is bare (no 'as'
577 # supplied at all) - try to infer the alias, either from the -as parameter
578 # of the selector spec, or use the parameter whole if it looks like a column
579 # name (ugly legacy heuristic). If all fails - leave the selector bare (which
580 # is ok as well), but make sure no more additions to the 'as' chain take place
581 for my $pref ('', '+') {
583 my ($sel, $as) = map {
584 my $key = "${pref}${_}";
586 my $val = [ ref $attrs->{$key} eq 'ARRAY'
588 : $attrs->{$key} || ()
590 delete $attrs->{$key};
594 if (! @$as and ! @$sel ) {
597 elsif (@$as and ! @$sel) {
598 $self->throw_exception(
599 "Unable to handle ${pref}as specification (@$as) without a corresponding ${pref}select"
603 # no as part supplied at all - try to deduce (unless explicit end of named selection is declared)
604 # if any @$as has been supplied we assume the user knows what (s)he is doing
605 # and blindly keep stacking up pieces
606 unless ($attrs->{_dark_selector}) {
609 if ( ref $_ eq 'HASH' and exists $_->{-as} ) {
610 push @$as, $_->{-as};
612 # assume any plain no-space, no-parenthesis string to be a column spec
613 # FIXME - this is retarded but is necessary to support shit like 'count(foo)'
614 elsif ( ! ref $_ and $_ =~ /^ [^\s\(\)]+ $/x) {
617 # if all else fails - raise a flag that no more aliasing will be allowed
619 $attrs->{_dark_selector} = {
622 local $Data::Dumper::Indent = 0;
631 elsif (@$as < @$sel) {
632 $self->throw_exception(
633 "Unable to handle an ${pref}as specification (@$as) with less elements than the corresponding ${pref}select"
636 elsif ($pref and $attrs->{_dark_selector}) {
637 $self->throw_exception(
638 "Unable to process named '+select', resultset contains an unnamed selector $attrs->{_dark_selector}{string}"
644 $attrs->{"${pref}select"} = $self->_merge_attr($attrs->{"${pref}select"}, $sel);
645 $attrs->{"${pref}as"} = $self->_merge_attr($attrs->{"${pref}as"}, $as);
650 my ($self, $left, $right) = @_;
653 (ref $_ eq 'ARRAY' and !@$_)
655 (ref $_ eq 'HASH' and ! keys %$_)
656 ) and $_ = undef for ($left, $right);
659 # either one of the two undef
660 ( (defined $left) xor (defined $right) ) ? ( defined $left ? $left : $right )
663 : ( ! defined $left ) ? undef
665 : { -and => [$left, $right] }
669 =head2 search_literal
671 B<CAVEAT>: C<search_literal> is provided for Class::DBI compatibility and
672 should only be used in that context. C<search_literal> is a convenience
673 method. It is equivalent to calling C<< $schema->search(\[]) >>, but if you
674 want to ensure columns are bound correctly, use L</search>.
676 See L<DBIx::Class::Manual::Cookbook/SEARCHING> and
677 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
678 require C<search_literal>.
682 =item Arguments: $sql_fragment, @standalone_bind_values
684 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
688 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
689 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
691 Pass a literal chunk of SQL to be added to the conditional part of the
694 Example of how to use C<search> instead of C<search_literal>
696 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
697 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
702 my ($self, $sql, @bind) = @_;
704 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
707 return $self->search(\[ $sql, map [ {} => $_ ], @bind ], ($attr || () ));
714 =item Arguments: \%columns_values | @pk_values, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
716 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
720 Finds and returns a single row based on supplied criteria. Takes either a
721 hashref with the same format as L</create> (including inference of foreign
722 keys from related objects), or a list of primary key values in the same
723 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
724 declaration on the L</result_source>.
726 In either case an attempt is made to combine conditions already existing on
727 the resultset with the condition passed to this method.
729 To aid with preparing the correct query for the storage you may supply the
730 C<key> attribute, which is the name of a
731 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
732 unique constraint corresponding to the
733 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
734 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
735 to construct a query that satisfies the named unique constraint fully (
736 non-NULL values for each column member of the constraint) an exception is
739 If no C<key> is specified, the search is carried over all unique constraints
740 which are fully defined by the available condition.
742 If no such constraint is found, C<find> currently defaults to a simple
743 C<< search->(\%column_values) >> which may or may not do what you expect.
744 Note that this fallback behavior may be deprecated in further versions. If
745 you need to search with arbitrary conditions - use L</search>. If the query
746 resulting from this fallback produces more than one row, a warning to the
747 effect is issued, though only the first row is constructed and returned as
750 In addition to C<key>, L</find> recognizes and applies standard
751 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
753 Note that if you have extra concerns about the correctness of the resulting
754 query you need to specify the C<key> attribute and supply the entire condition
755 as an argument to find (since it is not always possible to perform the
756 combination of the resultset condition with the supplied one, especially if
757 the resultset condition contains literal sql).
759 For example, to find a row by its primary key:
761 my $cd = $schema->resultset('CD')->find(5);
763 You can also find a row by a specific unique constraint:
765 my $cd = $schema->resultset('CD')->find(
767 artist => 'Massive Attack',
768 title => 'Mezzanine',
770 { key => 'cd_artist_title' }
773 See also L</find_or_create> and L</update_or_create>.
779 my $attrs = (@_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {});
783 if (exists $attrs->{key}) {
784 $constraint_name = defined $attrs->{key}
786 : $self->throw_exception("An undefined 'key' resultset attribute makes no sense")
790 # Parse out the condition from input
793 my $rsrc = $self->result_source;
795 if (ref $_[0] eq 'HASH') {
796 $call_cond = { %{$_[0]} };
799 # if only values are supplied we need to default to 'primary'
800 $constraint_name = 'primary' unless defined $constraint_name;
802 my @c_cols = $rsrc->unique_constraint_columns($constraint_name);
804 $self->throw_exception(
805 "No constraint columns, maybe a malformed '$constraint_name' constraint?"
808 $self->throw_exception (
809 'find() expects either a column/value hashref, or a list of values '
810 . "corresponding to the columns of the specified unique constraint '$constraint_name'"
811 ) unless @c_cols == @_;
813 @{$call_cond}{@c_cols} = @_;
816 # process relationship data if any
819 for my $key (keys %$call_cond) {
821 length ref($call_cond->{$key})
823 ( $rel_list ||= { map { $_ => 1 } $rsrc->relationships } )
826 ! is_literal_value( $call_cond->{$key} )
828 # implicitly skip has_many's (likely MC), via the delete()
829 ( ref( my $val = delete $call_cond->{$key} ) ne 'ARRAY' )
832 # FIXME: it seems wrong that relationship conditions take precedence...?
836 %{ $rsrc->_resolve_relationship_condition(
838 foreign_values => $val,
839 infer_values_based_on => {},
841 self_alias => "\xFE", # irrelevant
842 foreign_alias => "\xFF", # irrelevant
843 )->{inferred_values} },
848 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
850 if (defined $constraint_name) {
851 $final_cond = $self->_qualify_cond_columns (
853 $rsrc->_minimal_valueset_satisfying_constraint(
854 constraint_name => $constraint_name,
855 values => ($self->_merge_with_rscond($call_cond))[0],
862 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
863 # This means that we got here after a merger of relationship conditions
864 # in ::Relationship::Base::search_related (the row method), and furthermore
865 # the relationship is of the 'single' type. This means that the condition
866 # provided by the relationship (already attached to $self) is sufficient,
867 # as there can be only one row in the database that would satisfy the
871 my (@unique_queries, %seen_column_combinations, $ci, @fc_exceptions);
873 # no key was specified - fall down to heuristics mode:
874 # run through all unique queries registered on the resultset, and
875 # 'OR' all qualifying queries together
877 # always start from 'primary' if it exists at all
878 for my $c_name ( sort {
880 : $b eq 'primary' ? 1
882 } $rsrc->unique_constraint_names) {
884 next if $seen_column_combinations{
885 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
889 push @unique_queries, $self->_qualify_cond_columns(
890 $rsrc->_minimal_valueset_satisfying_constraint(
891 constraint_name => $c_name,
892 values => ($self->_merge_with_rscond($call_cond))[0],
893 columns_info => ($ci ||= $rsrc->columns_info),
898 dbic_internal_catch {
899 push @fc_exceptions, $_ if $_ =~ /\bFilterColumn\b/;
904 @unique_queries ? \@unique_queries
905 : @fc_exceptions ? $self->throw_exception(join "; ", map { $_ =~ /(.*) at .+ line \d+$/s } @fc_exceptions )
906 : $self->_non_unique_find_fallback ($call_cond, $attrs)
910 # Run the query, passing the result_class since it should propagate for find
911 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
912 if ($rs->_resolved_attrs->{collapse}) {
914 carp "Query returned more than one row" if $rs->next;
922 # This is a stop-gap method as agreed during the discussion on find() cleanup:
923 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
925 # It is invoked when find() is called in legacy-mode with insufficiently-unique
926 # condition. It is provided for overrides until a saner way forward is devised
928 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
929 # the road. Please adjust your tests accordingly to catch this situation early
930 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
932 # The method will not be removed without an adequately complete replacement
933 # for strict-mode enforcement
934 sub _non_unique_find_fallback {
935 my ($self, $cond, $attrs) = @_;
937 return $self->_qualify_cond_columns(
939 exists $attrs->{alias}
941 : $self->{attrs}{alias}
946 sub _qualify_cond_columns {
947 my ($self, $cond, $alias) = @_;
949 my %aliased = %$cond;
950 for (keys %aliased) {
951 $aliased{"$alias.$_"} = delete $aliased{$_}
958 sub _build_unique_cond {
960 '_build_unique_cond is a private method, and moreover is about to go '
961 . 'away. Please contact the development team at %s if you believe you '
962 . 'have a genuine use for this method, in order to discuss alternatives.',
963 DBIx::Class::_ENV_::HELP_URL,
966 my ($self, $constraint_name, $cond, $croak_on_null) = @_;
968 $self->result_source->_minimal_valueset_satisfying_constraint(
969 constraint_name => $constraint_name,
971 carp_on_nulls => !$croak_on_null
975 =head2 search_related
979 =item Arguments: $rel_name, $cond?, L<\%attrs?|/ATTRIBUTES>
981 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
985 $new_rs = $cd_rs->search_related('artist', {
989 Searches the specified relationship, optionally specifying a condition and
990 attributes for matching records. See L</ATTRIBUTES> for more information.
992 In list context, C<< ->all() >> is called implicitly on the resultset, thus
993 returning a list of result objects instead. To avoid that, use L</search_related_rs>.
995 See also L</search_related_rs>.
999 sub search_related :DBIC_method_is_indirect_sugar {
1000 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
1001 return shift->related_resultset(shift)->search(@_);
1004 =head2 search_related_rs
1006 This method works exactly the same as search_related, except that
1007 it guarantees a resultset, even in list context.
1011 sub search_related_rs :DBIC_method_is_indirect_sugar {
1012 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
1013 return shift->related_resultset(shift)->search_rs(@_);
1020 =item Arguments: none
1022 =item Return Value: L<$cursor|DBIx::Class::Cursor>
1026 Returns a storage-driven cursor to the given resultset. See
1027 L<DBIx::Class::Cursor> for more information.
1034 return $self->{cursor} ||= do {
1035 my $attrs = $self->_resolved_attrs;
1036 $self->result_source->schema->storage->select(
1037 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
1046 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1048 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1052 my $cd = $schema->resultset('CD')->single({ year => 2001 });
1054 Inflates the first result without creating a cursor if the resultset has
1055 any records in it; if not returns C<undef>. Used by L</find> as a lean version
1058 While this method can take an optional search condition (just like L</search>)
1059 being a fast-code-path it does not recognize search attributes. If you need to
1060 add extra joins or similar, call L</search> and then chain-call L</single> on the
1061 L<DBIx::Class::ResultSet> returned.
1067 As of 0.08100, this method enforces the assumption that the preceding
1068 query returns only one row. If more than one row is returned, you will receive
1071 Query returned more than one row
1073 In this case, you should be using L</next> or L</find> instead, or if you really
1074 know what you are doing, use the L</rows> attribute to explicitly limit the size
1077 This method will also throw an exception if it is called on a resultset prefetching
1078 has_many, as such a prefetch implies fetching multiple rows from the database in
1079 order to assemble the resulting object.
1086 my ($self, $where) = @_;
1088 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
1091 my $attrs = { %{$self->_resolved_attrs} };
1093 $self->throw_exception(
1094 'single() can not be used on resultsets collapsing a has_many. Use find( \%cond ) or next() instead'
1095 ) if $attrs->{collapse};
1098 if (defined $attrs->{where}) {
1101 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
1102 $where, delete $attrs->{where} ]
1105 $attrs->{where} = $where;
1109 my $data = [ $self->result_source->schema->storage->select_single(
1110 $attrs->{from}, $attrs->{select},
1111 $attrs->{where}, $attrs
1114 return undef unless @$data;
1115 $self->{_stashed_rows} = [ $data ];
1116 $self->_construct_results->[0];
1123 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1125 =item Return Value: L<$resultsetcolumn|DBIx::Class::ResultSetColumn>
1129 my $max_length = $rs->get_column('length')->max;
1131 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
1136 DBIx::Class::ResultSetColumn->new(@_);
1143 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1145 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1149 # WHERE title LIKE '%blue%'
1150 $cd_rs = $rs->search_like({ title => '%blue%'});
1152 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
1153 that this is simply a convenience method retained for ex Class::DBI users.
1154 You most likely want to use L</search> with specific operators.
1156 For more information, see L<DBIx::Class::Manual::Cookbook>.
1158 This method is deprecated and will be removed in 0.09. Use L<search()|/search>
1159 instead. An example conversion is:
1161 ->search_like({ foo => 'bar' });
1165 ->search({ foo => { like => 'bar' } });
1172 'search_like() is deprecated and will be removed in DBIC version 0.09.'
1173 .' Instead use ->search({ x => { -like => "y%" } })'
1174 .' (note the outer pair of {}s - they are important!)'
1176 my $attrs = (@_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {});
1177 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
1178 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
1179 return $class->search($query, { %$attrs });
1186 =item Arguments: $first, $last
1188 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1192 Returns a resultset or object list representing a subset of elements from the
1193 resultset slice is called on. Indexes are from 0, i.e., to get the first
1194 three records, call:
1196 my ($one, $two, $three) = $rs->slice(0, 2);
1201 my ($self, $min, $max) = @_;
1202 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
1203 $attrs->{offset} = $self->{attrs}{offset} || 0;
1204 $attrs->{offset} += $min;
1205 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
1206 return $self->search(undef, $attrs);
1213 =item Arguments: none
1215 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1219 Returns the next element in the resultset (C<undef> is there is none).
1221 Can be used to efficiently iterate over records in the resultset:
1223 my $rs = $schema->resultset('CD')->search;
1224 while (my $cd = $rs->next) {
1228 Note that you need to store the resultset object, and call C<next> on it.
1229 Calling C<< resultset('Table')->next >> repeatedly will always return the
1230 first record from the resultset.
1237 if (my $cache = $self->get_cache) {
1238 $self->{all_cache_position} ||= 0;
1239 return $cache->[$self->{all_cache_position}++];
1242 if ($self->{attrs}{cache}) {
1243 delete $self->{pager};
1244 $self->{all_cache_position} = 1;
1245 return ($self->all)[0];
1248 return shift(@{$self->{_stashed_results}}) if @{ $self->{_stashed_results}||[] };
1250 $self->{_stashed_results} = $self->_construct_results
1253 return shift @{$self->{_stashed_results}};
1256 # Constructs as many results as it can in one pass while respecting
1257 # cursor laziness. Several modes of operation:
1259 # * Always builds everything present in @{$self->{_stashed_rows}}
1260 # * If called with $fetch_all true - pulls everything off the cursor and
1261 # builds all result structures (or objects) in one pass
1262 # * If $self->_resolved_attrs->{collapse} is true, checks the order_by
1263 # and if the resultset is ordered properly by the left side:
1264 # * Fetches stuff off the cursor until the "master object" changes,
1265 # and saves the last extra row (if any) in @{$self->{_stashed_rows}}
1267 # * Just fetches, and collapses/constructs everything as if $fetch_all
1268 # was requested (there is no other way to collapse except for an
1270 # * If no collapse is requested - just get the next row, construct and
1272 sub _construct_results {
1273 my ($self, $fetch_all) = @_;
1275 my $rsrc = $self->result_source;
1276 my $attrs = $self->_resolved_attrs;
1281 ! $attrs->{order_by}
1285 my @pcols = $rsrc->primary_columns
1287 # default order for collapsing unless the user asked for something
1288 $attrs->{order_by} = [ map { join '.', $attrs->{alias}, $_} @pcols ];
1289 $attrs->{_ordered_for_collapse} = 1;
1290 $attrs->{_order_is_artificial} = 1;
1293 # this will be used as both initial raw-row collector AND as a RV of
1294 # _construct_results. Not regrowing the array twice matters a lot...
1295 # a surprising amount actually
1296 my $rows = delete $self->{_stashed_rows};
1298 my $cursor; # we may not need one at all
1300 my $did_fetch_all = $fetch_all;
1303 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1304 $rows = [ ($rows ? @$rows : ()), $self->cursor->all ];
1306 elsif( $attrs->{collapse} ) {
1308 # a cursor will need to be closed over in case of collapse
1309 $cursor = $self->cursor;
1311 $attrs->{_ordered_for_collapse} = (
1317 ->_extract_colinfo_of_stable_main_source_order_by_portion($attrs)
1319 ) unless defined $attrs->{_ordered_for_collapse};
1321 if (! $attrs->{_ordered_for_collapse}) {
1324 # instead of looping over ->next, use ->all in stealth mode
1325 # *without* calling a ->reset afterwards
1326 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1327 if (! $cursor->{_done}) {
1328 $rows = [ ($rows ? @$rows : ()), $cursor->all ];
1329 $cursor->{_done} = 1;
1334 if (! $did_fetch_all and ! @{$rows||[]} ) {
1335 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1336 $cursor ||= $self->cursor;
1337 if (scalar (my @r = $cursor->next) ) {
1342 return undef unless @{$rows||[]};
1344 # sanity check - people are too clever for their own good
1345 if ($attrs->{collapse} and my $aliastypes = $attrs->{_last_sqlmaker_alias_map} ) {
1347 my $multiplied_selectors;
1348 for my $sel_alias ( grep { $_ ne $attrs->{alias} } keys %{ $aliastypes->{selecting} } ) {
1350 $aliastypes->{multiplying}{$sel_alias}
1352 $aliastypes->{premultiplied}{$sel_alias}
1354 $multiplied_selectors->{$_} = 1 for values %{$aliastypes->{selecting}{$sel_alias}{-seen_columns}}
1358 for my $i (0 .. $#{$attrs->{as}} ) {
1359 my $sel = $attrs->{select}[$i];
1361 if (ref $sel eq 'SCALAR') {
1364 elsif( ref $sel eq 'REF' and ref $$sel eq 'ARRAY' ) {
1368 $self->throw_exception(
1369 'Result collapse not possible - selection from a has_many source redirected to the main object'
1370 ) if ($multiplied_selectors->{$sel} and $attrs->{as}[$i] !~ /\./);
1374 # hotspot - skip the setter
1375 my $res_class = $self->_result_class;
1377 my $inflator_cref = $self->{_result_inflator}{cref} ||= do {
1378 $res_class->can ('inflate_result')
1379 or $self->throw_exception("Inflator $res_class does not provide an inflate_result() method");
1382 my $infmap = $attrs->{as};
1384 $self->{_result_inflator}{is_core_row} = ( (
1387 ( \&DBIx::Class::Row::inflate_result || die "No ::Row::inflate_result() - can't happen" )
1388 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_core_row};
1390 $self->{_result_inflator}{is_hri} = ( (
1391 ! $self->{_result_inflator}{is_core_row}
1393 $inflator_cref == \&DBIx::Class::ResultClass::HashRefInflator::inflate_result
1394 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_hri};
1397 if ($attrs->{_simple_passthrough_construction}) {
1398 # construct a much simpler array->hash folder for the one-table HRI cases right here
1399 if ($self->{_result_inflator}{is_hri}) {
1400 for my $r (@$rows) {
1401 $r = { map { $infmap->[$_] => $r->[$_] } 0..$#$infmap };
1404 # FIXME SUBOPTIMAL this is a very very very hot spot
1405 # while rather optimal we can *still* do much better, by
1406 # building a smarter Row::inflate_result(), and
1407 # switch to feeding it data via a much leaner interface
1409 # crude unscientific benchmarking indicated the shortcut eval is not worth it for
1410 # this particular resultset size
1411 elsif ( $self->{_result_inflator}{is_core_row} and @$rows < 60 ) {
1412 for my $r (@$rows) {
1413 $r = $inflator_cref->($res_class, $rsrc, { map { $infmap->[$_] => $r->[$_] } (0..$#$infmap) } );
1418 ( $self->{_result_inflator}{is_core_row}
1419 ? '$_ = $inflator_cref->($res_class, $rsrc, { %s }) for @$rows'
1420 # a custom inflator may be a multiplier/reductor - put it in direct list ctx
1421 : '@$rows = map { $inflator_cref->($res_class, $rsrc, { %s } ) } @$rows'
1423 ( join (', ', map { "\$infmap->[$_] => \$_->[$_]" } 0..$#$infmap ) )
1429 $self->{_result_inflator}{is_hri} ? 'hri'
1430 : $self->{_result_inflator}{is_core_row} ? 'classic_pruning'
1431 : 'classic_nonpruning'
1434 unless( $self->{_row_parser}{$parser_type}{cref} ) {
1436 # $args and $attrs to _mk_row_parser are separated to delineate what is
1437 # core collapser stuff and what is dbic $rs specific
1438 $self->{_row_parser}{$parser_type}{src} = $rsrc->_mk_row_parser({
1439 inflate_map => $infmap,
1440 collapse => $attrs->{collapse},
1441 premultiplied => $attrs->{_main_source_premultiplied},
1442 hri_style => $self->{_result_inflator}{is_hri},
1443 prune_null_branches => $self->{_result_inflator}{is_hri} || $self->{_result_inflator}{is_core_row},
1446 $self->{_row_parser}{$parser_type}{cref} = do {
1447 package # hide form PAUSE
1448 DBIx::Class::__GENERATED_ROW_PARSER__;
1450 eval $self->{_row_parser}{$parser_type}{src};
1454 # this needs to close over the *current* cursor, hence why it is not cached above
1455 my $next_cref = ($did_fetch_all or ! $attrs->{collapse})
1458 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1459 my @r = $cursor->next or return;
1464 $self->{_row_parser}{$parser_type}{cref}->(
1467 ( $self->{_stashed_rows} = [] ),
1468 ( my $null_violations = {} ),
1471 $self->throw_exception(
1472 'Collapse aborted - the following columns are declared (or defaulted to) '
1473 . 'non-nullable within DBIC but NULLs were retrieved from storage: '
1474 . join( ', ', map { "'$infmap->[$_]'" } sort { $a <=> $b } keys %$null_violations )
1475 . ' within data row ' . dump_value({
1478 ( ! defined $self->{_stashed_rows}[0][$_] or length $self->{_stashed_rows}[0][$_] < 50 )
1479 ? $self->{_stashed_rows}[0][$_]
1480 : substr( $self->{_stashed_rows}[0][$_], 0, 50 ) . '...'
1481 } 0 .. $#{$self->{_stashed_rows}[0]}
1483 ) if keys %$null_violations;
1485 # simple in-place substitution, does not regrow $rows
1486 if ($self->{_result_inflator}{is_core_row}) {
1487 $_ = $inflator_cref->($res_class, $rsrc, @$_) for @$rows
1489 # Special-case multi-object HRI - there is no $inflator_cref pass at all
1490 elsif ( ! $self->{_result_inflator}{is_hri} ) {
1491 # the inflator may be a multiplier/reductor - put it in list ctx
1492 @$rows = map { $inflator_cref->($res_class, $rsrc, @$_) } @$rows;
1496 # The @$rows check seems odd at first - why wouldn't we want to warn
1497 # regardless? The issue is things like find() etc, where the user
1498 # *knows* only one result will come back. In these cases the ->all
1499 # is not a pessimization, but rather something we actually want
1501 'Unable to properly collapse has_many results in iterator mode due '
1502 . 'to order criteria - performed an eager cursor slurp underneath. '
1503 . 'Consider using ->all() instead'
1504 ) if ( ! $fetch_all and @$rows > 1 );
1509 =head2 result_source
1513 =item Arguments: L<$result_source?|DBIx::Class::ResultSource>
1515 =item Return Value: L<$result_source|DBIx::Class::ResultSource>
1519 An accessor for the primary ResultSource object from which this ResultSet
1526 =item Arguments: $result_class?
1528 =item Return Value: $result_class
1532 An accessor for the class to use when creating result objects. Defaults to
1533 C<< result_source->result_class >> - which in most cases is the name of the
1534 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1536 Note that changing the result_class will also remove any components
1537 that were originally loaded in the source class via
1538 L<load_components|Class::C3::Componentised/load_components( @comps )>.
1539 Any overloaded methods in the original source class will not run.
1544 my ($self, $result_class) = @_;
1545 if ($result_class) {
1547 # don't fire this for an object
1548 $self->ensure_class_loaded($result_class)
1549 unless ref($result_class);
1551 if ($self->get_cache) {
1552 carp_unique('Changing the result_class of a ResultSet instance with cached results is a noop - the cache contents will not be altered');
1554 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1555 elsif ($self->{cursor} && $self->{cursor}{_pos}) {
1556 $self->throw_exception('Changing the result_class of a ResultSet instance with an active cursor is not supported');
1559 $self->_result_class($result_class);
1561 delete $self->{_result_inflator};
1563 $self->_result_class;
1570 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1572 =item Return Value: $count
1576 Performs an SQL C<COUNT> with the same query as the resultset was built
1577 with to find the number of elements. Passing arguments is equivalent to
1578 C<< $rs->search ($cond, \%attrs)->count >>
1584 return $self->search(@_)->count if @_ and defined $_[0];
1585 return scalar @{ $self->get_cache } if $self->get_cache;
1587 my $attrs = { %{ $self->_resolved_attrs } };
1589 # this is a little optimization - it is faster to do the limit
1590 # adjustments in software, instead of a subquery
1591 my ($rows, $offset) = delete @{$attrs}{qw/rows offset/};
1594 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1595 $crs = $self->_count_subq_rs ($attrs);
1598 $crs = $self->_count_rs ($attrs);
1600 my $count = $crs->next;
1602 $count -= $offset if $offset;
1603 $count = $rows if $rows and $rows < $count;
1604 $count = 0 if ($count < 0);
1613 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1615 =item Return Value: L<$count_rs|DBIx::Class::ResultSetColumn>
1619 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1620 This can be very handy for subqueries:
1622 ->search( { amount => $some_rs->count_rs->as_query } )
1624 As with regular resultsets the SQL query will be executed only after
1625 the resultset is accessed via L</next> or L</all>. That would return
1626 the same single value obtainable via L</count>.
1632 return $self->search(@_)->count_rs if @_;
1634 # this may look like a lack of abstraction (count() does about the same)
1635 # but in fact an _rs *must* use a subquery for the limits, as the
1636 # software based limiting can not be ported if this $rs is to be used
1637 # in a subquery itself (i.e. ->as_query)
1638 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1639 return $self->_count_subq_rs($self->{_attrs});
1642 return $self->_count_rs($self->{_attrs});
1647 # returns a ResultSetColumn object tied to the count query
1650 my ($self, $attrs) = @_;
1652 my $rsrc = $self->result_source;
1654 my $tmp_attrs = { %$attrs };
1655 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1656 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1658 # overwrite the selector (supplied by the storage)
1659 $rsrc->resultset_class->new($rsrc, {
1661 select => $rsrc->schema->storage->_count_select ($rsrc, $attrs),
1663 })->get_column ('count');
1667 # same as above but uses a subquery
1669 sub _count_subq_rs {
1670 my ($self, $attrs) = @_;
1672 my $rsrc = $self->result_source;
1674 my $sub_attrs = { %$attrs };
1675 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1676 delete @{$sub_attrs}{qw/collapse columns as select order_by for/};
1678 # if we multi-prefetch we group_by something unique, as this is what we would
1679 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1680 if ( $attrs->{collapse} ) {
1681 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } @{
1682 $rsrc->_identifying_column_set || $self->throw_exception(
1683 'Unable to construct a unique group_by criteria properly collapsing the '
1684 . 'has_many prefetch before count()'
1689 # Calculate subquery selector
1690 if (my $g = $sub_attrs->{group_by}) {
1692 my $sql_maker = $rsrc->schema->storage->sql_maker;
1694 # necessary as the group_by may refer to aliased functions
1696 for my $sel (@{$attrs->{select}}) {
1697 $sel_index->{$sel->{-as}} = $sel
1698 if (ref $sel eq 'HASH' and $sel->{-as});
1701 # anything from the original select mentioned on the group-by needs to make it to the inner selector
1702 # also look for named aggregates referred in the having clause
1703 # having often contains scalarrefs - thus parse it out entirely
1705 if ($attrs->{having}) {
1706 local $sql_maker->{having_bind};
1707 local $sql_maker->{quote_char} = $sql_maker->{quote_char};
1708 local $sql_maker->{name_sep} = $sql_maker->{name_sep};
1709 unless (defined $sql_maker->{quote_char} and length $sql_maker->{quote_char}) {
1710 $sql_maker->{quote_char} = [ "\x00", "\xFF" ];
1711 # if we don't unset it we screw up retarded but unfortunately working
1712 # 'MAX(foo.bar)' => { '>', 3 }
1713 $sql_maker->{name_sep} = '';
1716 my ($lquote, $rquote, $sep) = map { quotemeta $_ } ($sql_maker->_quote_chars, $sql_maker->name_sep);
1718 my $having_sql = $sql_maker->_parse_rs_attrs ({ having => $attrs->{having} });
1721 # search for both a proper quoted qualified string, for a naive unquoted scalarref
1722 # and if all fails for an utterly naive quoted scalar-with-function
1723 while ($having_sql =~ /
1724 $rquote $sep $lquote (.+?) $rquote
1726 [\s,] \w+ \. (\w+) [\s,]
1728 [\s,] $lquote (.+?) $rquote [\s,]
1730 my $part = $1 || $2 || $3; # one of them matched if we got here
1731 unless ($seen_having{$part}++) {
1738 my $colpiece = $sel_index->{$_} || $_;
1740 # unqualify join-based group_by's. Arcane but possible query
1741 # also horrible horrible hack to alias a column (not a func.)
1742 # (probably need to introduce SQLA syntax)
1743 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1746 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1748 push @{$sub_attrs->{select}}, $colpiece;
1752 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1753 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1756 return $rsrc->resultset_class
1757 ->new ($rsrc, $sub_attrs)
1759 ->search ({}, { columns => { count => $rsrc->schema->storage->_count_select ($rsrc, $attrs) } })
1760 ->get_column ('count');
1764 =head2 count_literal
1766 B<CAVEAT>: C<count_literal> is provided for Class::DBI compatibility and
1767 should only be used in that context. See L</search_literal> for further info.
1771 =item Arguments: $sql_fragment, @standalone_bind_values
1773 =item Return Value: $count
1777 Counts the results in a literal query. Equivalent to calling L</search_literal>
1778 with the passed arguments, then L</count>.
1782 sub count_literal :DBIC_method_is_indirect_sugar {
1783 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
1784 shift->search_literal(@_)->count
1791 =item Arguments: none
1793 =item Return Value: L<@result_objs|DBIx::Class::Manual::ResultClass>
1797 Returns all elements in the resultset.
1804 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1807 delete @{$self}{qw/_stashed_rows _stashed_results/};
1809 if (my $c = $self->get_cache) {
1813 $self->cursor->reset;
1815 my $objs = $self->_construct_results('fetch_all') || [];
1817 $self->set_cache($objs) if $self->{attrs}{cache};
1826 =item Arguments: none
1828 =item Return Value: $self
1832 Resets the resultset's cursor, so you can iterate through the elements again.
1833 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1841 delete @{$self}{qw/_stashed_rows _stashed_results/};
1842 $self->{all_cache_position} = 0;
1843 $self->cursor->reset;
1851 =item Arguments: none
1853 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1857 L<Resets|/reset> the resultset (causing a fresh query to storage) and returns
1858 an object for the first result (or C<undef> if the resultset is empty).
1862 sub first :DBIC_method_is_indirect_sugar {
1863 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
1864 return $_[0]->reset->next;
1870 # Determines whether and what type of subquery is required for the $rs operation.
1871 # If grouping is necessary either supplies its own, or verifies the current one
1872 # After all is done delegates to the proper storage method.
1874 sub _rs_update_delete {
1875 my ($self, $op, $values) = @_;
1877 my $rsrc = $self->result_source;
1878 my $storage = $rsrc->schema->storage;
1880 my $attrs = { %{$self->_resolved_attrs} };
1882 my $join_classifications;
1883 my ($existing_group_by) = delete @{$attrs}{qw(group_by _grouped_by_distinct)};
1885 # do we need a subquery for any reason?
1887 defined $existing_group_by
1889 # if {from} is unparseable wrap a subq
1890 ref($attrs->{from}) ne 'ARRAY'
1892 # limits call for a subq
1893 $self->_has_resolved_attr(qw/rows offset/)
1896 # simplify the joinmap, so we can further decide if a subq is necessary
1897 if (!$needs_subq and @{$attrs->{from}} > 1) {
1899 ($attrs->{from}, $join_classifications) =
1900 $storage->_prune_unused_joins ($attrs);
1902 # any non-pruneable non-local restricting joins imply subq
1903 $needs_subq = grep { $_ ne $attrs->{alias} } keys %{ $join_classifications->{restricting} || {} };
1906 # check if the head is composite (by now all joins are thrown out unless $needs_subq)
1908 (ref $attrs->{from}[0]) ne 'HASH'
1910 ref $attrs->{from}[0]{ $attrs->{from}[0]{-alias} }
1914 # do we need anything like a subquery?
1915 if (! $needs_subq) {
1916 # Most databases do not allow aliasing of tables in UPDATE/DELETE. Thus
1917 # a condition containing 'me' or other table prefixes will not work
1918 # at all. Tell SQLMaker to dequalify idents via a gross hack.
1920 my $sqla = $rsrc->schema->storage->sql_maker;
1921 local $sqla->{_dequalify_idents} = 1;
1922 \[ $sqla->_recurse_where($self->{cond}) ];
1926 # we got this far - means it is time to wrap a subquery
1927 my $idcols = $rsrc->_identifying_column_set || $self->throw_exception(
1929 "Unable to perform complex resultset %s() without an identifying set of columns on source '%s'",
1935 # make a new $rs selecting only the PKs (that's all we really need for the subq)
1936 delete $attrs->{$_} for qw/select as collapse/;
1937 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } @$idcols ];
1939 # this will be consumed by the pruner waaaaay down the stack
1940 $attrs->{_force_prune_multiplying_joins} = 1;
1942 my $subrs = (ref $self)->new($rsrc, $attrs);
1944 if (@$idcols == 1) {
1945 $cond = { $idcols->[0] => { -in => $subrs->as_query } };
1947 elsif ($storage->_use_multicolumn_in) {
1948 # no syntax for calling this properly yet
1949 # !!! EXPERIMENTAL API !!! WILL CHANGE !!!
1950 $cond = $storage->sql_maker->_where_op_multicolumn_in (
1951 $idcols, # how do I convey a list of idents...? can binds reside on lhs?
1956 # if all else fails - get all primary keys and operate over a ORed set
1957 # wrap in a transaction for consistency
1958 # this is where the group_by/multiplication starts to matter
1962 # we do not need to check pre-multipliers, since if the premulti is there, its
1963 # parent (who is multi) will be there too
1964 keys %{ $join_classifications->{multiplying} || {} }
1966 # make sure if there is a supplied group_by it matches the columns compiled above
1967 # perfectly. Anything else can not be sanely executed on most databases so croak
1968 # right then and there
1969 if ($existing_group_by) {
1970 my @current_group_by = map
1971 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1976 join ("\x00", sort @current_group_by)
1978 join ("\x00", sort @{$attrs->{columns}} )
1980 $self->throw_exception (
1981 "You have just attempted a $op operation on a resultset which does group_by"
1982 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1983 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1984 . ' kind of queries. Please retry the operation with a modified group_by or'
1985 . ' without using one at all.'
1990 $subrs = $subrs->search({}, { group_by => $attrs->{columns} });
1993 $guard = $storage->txn_scope_guard;
1995 for my $row ($subrs->cursor->all) {
1997 { $idcols->[$_] => $row->[$_] }
2004 my $res = $cond ? $storage->$op (
2006 $op eq 'update' ? $values : (),
2010 $guard->commit if $guard;
2019 =item Arguments: \%values
2021 =item Return Value: $underlying_storage_rv
2025 Sets the specified columns in the resultset to the supplied values in a
2026 single query. Note that this will not run any accessor/set_column/update
2027 triggers, nor will it update any result object instances derived from this
2028 resultset (this includes the contents of the L<resultset cache|/set_cache>
2029 if any). See L</update_all> if you need to execute any on-update
2030 triggers or cascades defined either by you or a
2031 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2033 The return value is a pass through of what the underlying
2034 storage backend returned, and may vary. See L<DBI/execute> for the most
2039 Note that L</update> does not process/deflate any of the values passed in.
2040 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
2041 ensure manually that any value passed to this method will stringify to
2042 something the RDBMS knows how to deal with. A notable example is the
2043 handling of L<DateTime> objects, for more info see:
2044 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
2049 my ($self, $values) = @_;
2050 $self->throw_exception('Values for update must be a hash')
2051 unless ref $values eq 'HASH';
2053 return $self->_rs_update_delete ('update', $values);
2060 =item Arguments: \%values
2062 =item Return Value: 1
2066 Fetches all objects and updates them one at a time via
2067 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
2068 triggers, while L</update> will not.
2073 my ($self, $values) = @_;
2074 $self->throw_exception('Values for update_all must be a hash')
2075 unless ref $values eq 'HASH';
2077 my $guard = $self->result_source->schema->txn_scope_guard;
2078 $_->update({%$values}) for $self->all; # shallow copy - update will mangle it
2087 =item Arguments: none
2089 =item Return Value: $underlying_storage_rv
2093 Deletes the rows matching this resultset in a single query. Note that this
2094 will not run any delete triggers, nor will it alter the
2095 L<in_storage|DBIx::Class::Row/in_storage> status of any result object instances
2096 derived from this resultset (this includes the contents of the
2097 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
2098 execute any on-delete triggers or cascades defined either by you or a
2099 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2101 The return value is a pass through of what the underlying storage backend
2102 returned, and may vary. See L<DBI/execute> for the most common case.
2108 $self->throw_exception('delete does not accept any arguments')
2111 return $self->_rs_update_delete ('delete');
2118 =item Arguments: none
2120 =item Return Value: 1
2124 Fetches all objects and deletes them one at a time via
2125 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
2126 triggers, while L</delete> will not.
2132 $self->throw_exception('delete_all does not accept any arguments')
2135 my $guard = $self->result_source->schema->txn_scope_guard;
2136 $_->delete for $self->all;
2145 =item Arguments: [ \@column_list, \@row_values+ ] | [ \%col_data+ ]
2147 =item Return Value: L<\@result_objects|DBIx::Class::Manual::ResultClass> (scalar context) | L<@result_objects|DBIx::Class::Manual::ResultClass> (list context)
2151 Accepts either an arrayref of hashrefs or alternatively an arrayref of
2158 The context of this method call has an important effect on what is
2159 submitted to storage. In void context data is fed directly to fastpath
2160 insertion routines provided by the underlying storage (most often
2161 L<DBI/execute_for_fetch>), bypassing the L<new|DBIx::Class::Row/new> and
2162 L<insert|DBIx::Class::Row/insert> calls on the
2163 L<Result|DBIx::Class::Manual::ResultClass> class, including any
2164 augmentation of these methods provided by components. For example if you
2165 are using something like L<DBIx::Class::UUIDColumns> to create primary
2166 keys for you, you will find that your PKs are empty. In this case you
2167 will have to explicitly force scalar or list context in order to create
2172 In non-void (scalar or list) context, this method is simply a wrapper
2173 for L</create>. Depending on list or scalar context either a list of
2174 L<Result|DBIx::Class::Manual::ResultClass> objects or an arrayref
2175 containing these objects is returned.
2177 When supplying data in "arrayref of arrayrefs" invocation style, the
2178 first element should be a list of column names and each subsequent
2179 element should be a data value in the earlier specified column order.
2182 $schema->resultset("Artist")->populate([
2183 [ qw( artistid name ) ],
2184 [ 100, 'A Formally Unknown Singer' ],
2185 [ 101, 'A singer that jumped the shark two albums ago' ],
2186 [ 102, 'An actually cool singer' ],
2189 For the arrayref of hashrefs style each hashref should be a structure
2190 suitable for passing to L</create>. Multi-create is also permitted with
2193 $schema->resultset("Artist")->populate([
2194 { artistid => 4, name => 'Manufactured Crap', cds => [
2195 { title => 'My First CD', year => 2006 },
2196 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2199 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
2200 { title => 'My parents sold me to a record company', year => 2005 },
2201 { title => 'Why Am I So Ugly?', year => 2006 },
2202 { title => 'I Got Surgery and am now Popular', year => 2007 }
2207 If you attempt a void-context multi-create as in the example above (each
2208 Artist also has the related list of CDs), and B<do not> supply the
2209 necessary autoinc foreign key information, this method will proxy to the
2210 less efficient L</create>, and then throw the Result objects away. In this
2211 case there are obviously no benefits to using this method over L</create>.
2218 # this is naive and just a quick check
2219 # the types will need to be checked more thoroughly when the
2220 # multi-source populate gets added
2222 ref $_[0] eq 'ARRAY'
2224 ( @{$_[0]} or return )
2226 ( ref $_[0][0] eq 'HASH' or ref $_[0][0] eq 'ARRAY' )
2229 ) or $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
2231 # FIXME - no cref handling
2232 # At this point assume either hashes or arrays
2234 my $rsrc = $self->result_source;
2235 my $storage = $rsrc->schema->storage;
2237 if(defined wantarray) {
2238 my (@results, $guard);
2240 if (ref $data->[0] eq 'ARRAY') {
2241 # column names only, nothing to do
2242 return if @$data == 1;
2244 $guard = $storage->txn_scope_guard
2248 { my $vals = $_; $self->new_result({ map { $data->[0][$_] => $vals->[$_] } 0..$#{$data->[0]} })->insert }
2249 @{$data}[1 .. $#$data]
2254 $guard = $storage->txn_scope_guard
2257 @results = map { $self->new_result($_)->insert } @$data;
2260 $guard->commit if $guard;
2261 return wantarray ? @results : \@results;
2264 # we have to deal with *possibly incomplete* related data
2265 # this means we have to walk the data structure twice
2266 # whether we want this or not
2267 # jnap, I hate you ;)
2268 my $rel_info = { map { $_ => $rsrc->relationship_info($_) } $rsrc->relationships };
2270 my ($colinfo, $colnames, $slices_with_rels);
2274 for my $i (0 .. $#$data) {
2276 my $current_slice_seen_rel_infos;
2278 ### Determine/Supplement collists
2279 ### BEWARE - This is a hot piece of code, a lot of weird idioms were used
2280 if( ref $data->[$i] eq 'ARRAY' ) {
2282 # positional(!) explicit column list
2284 # column names only, nothing to do
2285 return if @$data == 1;
2287 $colinfo->{$data->[0][$_]} = { pos => $_, name => $data->[0][$_] } and push @$colnames, $data->[0][$_]
2288 for 0 .. $#{$data->[0]};
2295 for (values %$colinfo) {
2296 if ($_->{is_rel} ||= (
2297 $rel_info->{$_->{name}}
2300 ref $data->[$i][$_->{pos}] eq 'ARRAY'
2302 ref $data->[$i][$_->{pos}] eq 'HASH'
2305 defined blessed $data->[$i][$_->{pos}]
2307 $data->[$i][$_->{pos}]->isa(
2308 $DBIx::Class::ResultSource::__expected_result_class_isa
2312 msg => 'Global $DBIx::Class::ResultSource::__expected_result_class_isa unexpectedly unset...'
2321 # moar sanity check... sigh
2322 for ( ref $data->[$i][$_->{pos}] eq 'ARRAY' ? @{$data->[$i][$_->{pos}]} : $data->[$i][$_->{pos}] ) {
2327 $DBIx::Class::ResultSource::__expected_result_class_isa
2331 msg => 'Global $DBIx::Class::ResultSource::__expected_result_class_isa unexpectedly unset...'
2335 carp_unique("Fast-path populate() with supplied related objects is not possible - falling back to regular create()");
2336 return my $throwaway = $self->populate(@_);
2340 push @$current_slice_seen_rel_infos, $rel_info->{$_->{name}};
2345 if ($current_slice_seen_rel_infos) {
2346 push @$slices_with_rels, { map { $colnames->[$_] => $data->[$i][$_] } 0 .. $#$colnames };
2348 # this is needed further down to decide whether or not to fallback to create()
2349 $colinfo->{$colnames->[$_]}{seen_null} ||= ! defined $data->[$i][$_]
2350 for 0 .. $#$colnames;
2353 elsif( ref $data->[$i] eq 'HASH' ) {
2355 for ( sort keys %{$data->[$i]} ) {
2357 $colinfo->{$_} ||= do {
2359 $self->throw_exception("Column '$_' must be present in supplied explicit column list")
2360 if $data_start; # it will be 0 on AoH, 1 on AoA
2362 push @$colnames, $_;
2365 { pos => $#$colnames, name => $_ }
2368 if ($colinfo->{$_}{is_rel} ||= (
2372 ref $data->[$i]{$_} eq 'ARRAY'
2374 ref $data->[$i]{$_} eq 'HASH'
2377 defined blessed $data->[$i]{$_}
2379 $data->[$i]{$_}->isa(
2380 $DBIx::Class::ResultSource::__expected_result_class_isa
2384 msg => 'Global $DBIx::Class::ResultSource::__expected_result_class_isa unexpectedly unset...'
2393 # moar sanity check... sigh
2394 for ( ref $data->[$i]{$_} eq 'ARRAY' ? @{$data->[$i]{$_}} : $data->[$i]{$_} ) {
2399 $DBIx::Class::ResultSource::__expected_result_class_isa
2403 msg => 'Global $DBIx::Class::ResultSource::__expected_result_class_isa unexpectedly unset...'
2407 carp_unique("Fast-path populate() with supplied related objects is not possible - falling back to regular create()");
2408 return my $throwaway = $self->populate(@_);
2412 push @$current_slice_seen_rel_infos, $rel_info->{$_};
2416 if ($current_slice_seen_rel_infos) {
2417 push @$slices_with_rels, $data->[$i];
2419 # this is needed further down to decide whether or not to fallback to create()
2420 $colinfo->{$_}{seen_null} ||= ! defined $data->[$i]{$_}
2421 for keys %{$data->[$i]};
2425 $self->throw_exception('Unexpected populate() data structure member type: ' . ref $data->[$i] );
2429 { $_->{attrs}{is_depends_on} }
2430 @{ $current_slice_seen_rel_infos || [] }
2432 carp_unique("Fast-path populate() of belongs_to relationship data is not possible - falling back to regular create()");
2433 return my $throwaway = $self->populate(@_);
2437 if( $slices_with_rels ) {
2439 # need to exclude the rel "columns"
2440 $colnames = [ grep { ! $colinfo->{$_}{is_rel} } @$colnames ];
2442 # extra sanity check - ensure the main source is in fact identifiable
2443 # the localizing of nullability is insane, but oh well... the use-case is legit
2444 my $ci = $rsrc->columns_info($colnames);
2446 $ci->{$_} = { %{$ci->{$_}}, is_nullable => 0 }
2447 for grep { ! $colinfo->{$_}{seen_null} } keys %$ci;
2449 unless( $rsrc->_identifying_column_set($ci) ) {
2450 carp_unique("Fast-path populate() of non-uniquely identifiable rows with related data is not possible - falling back to regular create()");
2451 return my $throwaway = $self->populate(@_);
2455 ### inherit the data locked in the conditions of the resultset
2456 my ($rs_data) = $self->_merge_with_rscond({});
2457 delete @{$rs_data}{@$colnames}; # passed-in stuff takes precedence
2459 # if anything left - decompose rs_data
2461 if (keys %$rs_data) {
2462 push @$rs_data_vals, $rs_data->{$_}
2463 for sort keys %$rs_data;
2468 $guard = $storage->txn_scope_guard
2469 if $slices_with_rels;
2471 ### main source data
2472 # FIXME - need to switch entirely to a coderef-based thing,
2473 # so that large sets aren't copied several times... I think
2474 $storage->_insert_bulk(
2476 [ @$colnames, sort keys %$rs_data ],
2478 ref $data->[$_] eq 'ARRAY'
2480 $slices_with_rels ? [ @{$data->[$_]}[0..$#$colnames], @{$rs_data_vals||[]} ] # the collist changed
2481 : $rs_data_vals ? [ @{$data->[$_]}, @$rs_data_vals ]
2484 : [ @{$data->[$_]}{@$colnames}, @{$rs_data_vals||[]} ]
2485 } $data_start .. $#$data ],
2488 ### do the children relationships
2489 if ( $slices_with_rels ) {
2490 my @rels = grep { $colinfo->{$_}{is_rel} } keys %$colinfo
2491 or die 'wtf... please report a bug with DBIC_TRACE=1 output (stacktrace)';
2493 for my $sl (@$slices_with_rels) {
2495 my ($main_proto, $main_proto_rs);
2496 for my $rel (@rels) {
2497 next unless defined $sl->{$rel};
2501 (map { $_ => $sl->{$_} } @$colnames),
2504 unless (defined $colinfo->{$rel}{rs}) {
2506 $colinfo->{$rel}{rs} = $rsrc->related_source($rel)->resultset;
2508 $colinfo->{$rel}{fk_map} = { reverse %{ $rsrc->_resolve_relationship_condition(
2510 self_alias => "\xFE", # irrelevant
2511 foreign_alias => "\xFF", # irrelevant
2512 )->{identity_map} || {} } };
2516 $colinfo->{$rel}{rs}->search({ map # only so that we inherit them values properly, no actual search
2519 ( $main_proto_rs ||= $rsrc->resultset->search($main_proto) )
2520 ->get_column( $colinfo->{$rel}{fk_map}{$_} )
2524 keys %{$colinfo->{$rel}{fk_map}}
2525 })->populate( ref $sl->{$rel} eq 'ARRAY' ? $sl->{$rel} : [ $sl->{$rel} ] );
2532 $guard->commit if $guard;
2539 =item Arguments: none
2541 =item Return Value: L<$pager|Data::Page>
2545 Returns a L<Data::Page> object for the current resultset. Only makes
2546 sense for queries with a C<page> attribute.
2548 To get the full count of entries for a paged resultset, call
2549 C<total_entries> on the L<Data::Page> object.
2556 return $self->{pager} if $self->{pager};
2558 my $attrs = $self->{attrs};
2559 if (!defined $attrs->{page}) {
2560 $self->throw_exception("Can't create pager for non-paged rs");
2562 elsif ($attrs->{page} <= 0) {
2563 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
2565 $attrs->{rows} ||= 10;
2567 # throw away the paging flags and re-run the count (possibly
2568 # with a subselect) to get the real total count
2569 my $count_attrs = { %$attrs };
2570 delete @{$count_attrs}{qw/rows offset page pager/};
2572 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
2574 require DBIx::Class::ResultSet::Pager;
2575 return $self->{pager} = DBIx::Class::ResultSet::Pager->new(
2576 sub { $total_rs->count }, #lazy-get the total
2578 $self->{attrs}{page},
2586 =item Arguments: $page_number
2588 =item Return Value: L<$resultset|/search>
2592 Returns a resultset for the $page_number page of the resultset on which page
2593 is called, where each page contains a number of rows equal to the 'rows'
2594 attribute set on the resultset (10 by default).
2599 my ($self, $page) = @_;
2600 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2607 =item Arguments: \%col_data
2609 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2613 Creates a new result object in the resultset's result class and returns
2614 it. The row is not inserted into the database at this point, call
2615 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2616 will tell you whether the result object has been inserted or not.
2618 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2623 my ($self, $values) = @_;
2625 $self->throw_exception( "Result object instantiation requires a single hashref argument" )
2626 if @_ > 2 or ref $values ne 'HASH';
2628 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2630 my $new = $self->result_class->new({
2632 ( @$cols_from_relations
2633 ? (-cols_from_relations => $cols_from_relations)
2636 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2640 reftype($new) eq 'HASH'
2646 carp_unique (sprintf (
2647 "%s->new returned a blessed empty hashref - a strong indicator something is wrong with its inheritance chain",
2648 $self->result_class,
2655 # _merge_with_rscond
2657 # Takes a simple hash of K/V data and returns its copy merged with the
2658 # condition already present on the resultset. Additionally returns an
2659 # arrayref of value/condition names, which were inferred from related
2660 # objects (this is needed for in-memory related objects)
2661 sub _merge_with_rscond {
2662 my ($self, $data) = @_;
2664 my ($implied_data, @cols_from_relations);
2666 my $alias = $self->{attrs}{alias};
2668 if (! defined $self->{cond}) {
2669 # just massage $data below
2671 elsif ($self->{cond} eq UNRESOLVABLE_CONDITION) {
2672 $implied_data = $self->{attrs}{related_objects}; # nothing might have been inserted yet
2673 @cols_from_relations = keys %{ $implied_data || {} };
2676 my $eqs = extract_equality_conditions( $self->{cond}, 'consider_nulls' );
2677 $implied_data = { map {
2678 ( ($eqs->{$_}||'') eq UNRESOLVABLE_CONDITION ) ? () : ( $_ => $eqs->{$_} )
2684 { %{ $self->_remove_alias($_, $alias) } }
2685 # precedence must be given to passed values over values inherited from
2686 # the cond, so the order here is important.
2687 ( $implied_data||(), $data)
2689 \@cols_from_relations
2693 # _has_resolved_attr
2695 # determines if the resultset defines at least one
2696 # of the attributes supplied
2698 # used to determine if a subquery is necessary
2700 # supports some virtual attributes:
2702 # This will scan for any joins being present on the resultset.
2703 # It is not a mere key-search but a deep inspection of {from}
2706 sub _has_resolved_attr {
2707 my ($self, @attr_names) = @_;
2709 my $attrs = $self->_resolved_attrs;
2713 for my $n (@attr_names) {
2714 if (grep { $n eq $_ } (qw/-join/) ) {
2715 $extra_checks{$n}++;
2719 my $attr = $attrs->{$n};
2721 next if not defined $attr;
2723 if (ref $attr eq 'HASH') {
2724 return 1 if keys %$attr;
2726 elsif (ref $attr eq 'ARRAY') {
2734 # a resolved join is expressed as a multi-level from
2736 $extra_checks{-join}
2738 ref $attrs->{from} eq 'ARRAY'
2740 @{$attrs->{from}} > 1
2748 # Remove the specified alias from the specified query hash. A copy is made so
2749 # the original query is not modified.
2752 my ($self, $query, $alias) = @_;
2754 my %orig = %{ $query || {} };
2757 foreach my $key (keys %orig) {
2759 $unaliased{$key} = $orig{$key};
2762 $unaliased{$1} = $orig{$key}
2763 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2773 =item Arguments: none
2775 =item Return Value: \[ $sql, L<@bind_values|/DBIC BIND VALUES> ]
2779 Returns the SQL query and bind vars associated with the invocant.
2781 This is generally used as the RHS for a subquery.
2788 my $attrs = { %{ $self->_resolved_attrs } };
2790 my $aq = $self->result_source->schema->storage->_select_args_to_query (
2791 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
2801 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2803 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2807 my $artist = $schema->resultset('Artist')->find_or_new(
2808 { artist => 'fred' }, { key => 'artists' });
2810 $cd->cd_to_producer->find_or_new({ producer => $producer },
2811 { key => 'primary' });
2813 Find an existing record from this resultset using L</find>. if none exists,
2814 instantiate a new result object and return it. The object will not be saved
2815 into your storage until you call L<DBIx::Class::Row/insert> on it.
2817 You most likely want this method when looking for existing rows using a unique
2818 constraint that is not the primary key, or looking for related rows.
2820 If you want objects to be saved immediately, use L</find_or_create> instead.
2822 B<Note>: Make sure to read the documentation of L</find> and understand the
2823 significance of the C<key> attribute, as its lack may skew your search, and
2824 subsequently result in spurious new objects.
2826 B<Note>: Take care when using C<find_or_new> with a table having
2827 columns with default values that you intend to be automatically
2828 supplied by the database (e.g. an auto_increment primary key column).
2829 In normal usage, the value of such columns should NOT be included at
2830 all in the call to C<find_or_new>, even when set to C<undef>.
2836 my $attrs = (@_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {});
2837 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2838 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2841 return $self->new_result($hash);
2848 =item Arguments: \%col_data
2850 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2854 Attempt to create a single new row or a row with multiple related rows
2855 in the table represented by the resultset (and related tables). This
2856 will not check for duplicate rows before inserting, use
2857 L</find_or_create> to do that.
2859 To create one row for this resultset, pass a hashref of key/value
2860 pairs representing the columns of the table and the values you wish to
2861 store. If the appropriate relationships are set up, foreign key fields
2862 can also be passed an object representing the foreign row, and the
2863 value will be set to its primary key.
2865 To create related objects, pass a hashref of related-object column values
2866 B<keyed on the relationship name>. If the relationship is of type C<multi>
2867 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2868 The process will correctly identify columns holding foreign keys, and will
2869 transparently populate them from the keys of the corresponding relation.
2870 This can be applied recursively, and will work correctly for a structure
2871 with an arbitrary depth and width, as long as the relationships actually
2872 exists and the correct column data has been supplied.
2874 Instead of hashrefs of plain related data (key/value pairs), you may
2875 also pass new or inserted objects. New objects (not inserted yet, see
2876 L</new_result>), will be inserted into their appropriate tables.
2878 Effectively a shortcut for C<< ->new_result(\%col_data)->insert >>.
2880 Example of creating a new row.
2882 $person_rs->create({
2883 name=>"Some Person",
2884 email=>"somebody@someplace.com"
2887 Example of creating a new row and also creating rows in a related C<has_many>
2888 or C<has_one> resultset. Note Arrayref.
2891 { artistid => 4, name => 'Manufactured Crap', cds => [
2892 { title => 'My First CD', year => 2006 },
2893 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2898 Example of creating a new row and also creating a row in a related
2899 C<belongs_to> resultset. Note Hashref.
2902 title=>"Music for Silly Walks",
2905 name=>"Silly Musician",
2913 When subclassing ResultSet never attempt to override this method. Since
2914 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2915 lot of the internals simply never call it, so your override will be
2916 bypassed more often than not. Override either L<DBIx::Class::Row/new>
2917 or L<DBIx::Class::Row/insert> depending on how early in the
2918 L</create> process you need to intervene. See also warning pertaining to
2925 sub create :DBIC_method_is_indirect_sugar {
2926 #my ($self, $col_data) = @_;
2927 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_INDIRECT_CALLS and fail_on_internal_call;
2928 return shift->new_result(shift)->insert;
2931 =head2 find_or_create
2935 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2937 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2941 $cd->cd_to_producer->find_or_create({ producer => $producer },
2942 { key => 'primary' });
2944 Tries to find a record based on its primary key or unique constraints; if none
2945 is found, creates one and returns that instead.
2947 my $cd = $schema->resultset('CD')->find_or_create({
2949 artist => 'Massive Attack',
2950 title => 'Mezzanine',
2954 Also takes an optional C<key> attribute, to search by a specific key or unique
2955 constraint. For example:
2957 my $cd = $schema->resultset('CD')->find_or_create(
2959 artist => 'Massive Attack',
2960 title => 'Mezzanine',
2962 { key => 'cd_artist_title' }
2965 B<Note>: Make sure to read the documentation of L</find> and understand the
2966 significance of the C<key> attribute, as its lack may skew your search, and
2967 subsequently result in spurious row creation.
2969 B<Note>: Because find_or_create() reads from the database and then
2970 possibly inserts based on the result, this method is subject to a race
2971 condition. Another process could create a record in the table after
2972 the find has completed and before the create has started. To avoid
2973 this problem, use find_or_create() inside a transaction.
2975 B<Note>: Take care when using C<find_or_create> with a table having
2976 columns with default values that you intend to be automatically
2977 supplied by the database (e.g. an auto_increment primary key column).
2978 In normal usage, the value of such columns should NOT be included at
2979 all in the call to C<find_or_create>, even when set to C<undef>.
2981 See also L</find> and L</update_or_create>. For information on how to declare
2982 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2984 If you need to know if an existing row was found or a new one created use
2985 L</find_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2986 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2989 my $cd = $schema->resultset('CD')->find_or_new({
2991 artist => 'Massive Attack',
2992 title => 'Mezzanine',
2996 if( !$cd->in_storage ) {
3003 sub find_or_create {
3005 my $attrs = (@_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {});
3006 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
3007 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
3010 return $self->new_result($hash)->insert;
3013 =head2 update_or_create
3017 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
3019 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
3023 $resultset->update_or_create({ col => $val, ... });
3025 Like L</find_or_create>, but if a row is found it is immediately updated via
3026 C<< $found_row->update (\%col_data) >>.
3029 Takes an optional C<key> attribute to search on a specific unique constraint.
3032 # In your application
3033 my $cd = $schema->resultset('CD')->update_or_create(
3035 artist => 'Massive Attack',
3036 title => 'Mezzanine',
3039 { key => 'cd_artist_title' }
3042 $cd->cd_to_producer->update_or_create({
3043 producer => $producer,
3049 B<Note>: Make sure to read the documentation of L</find> and understand the
3050 significance of the C<key> attribute, as its lack may skew your search, and
3051 subsequently result in spurious row creation.
3053 B<Note>: Take care when using C<update_or_create> with a table having
3054 columns with default values that you intend to be automatically
3055 supplied by the database (e.g. an auto_increment primary key column).
3056 In normal usage, the value of such columns should NOT be included at
3057 all in the call to C<update_or_create>, even when set to C<undef>.
3059 See also L</find> and L</find_or_create>. For information on how to declare
3060 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
3062 If you need to know if an existing row was updated or a new one created use
3063 L</update_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
3064 to call L<DBIx::Class::Row/insert> to save the newly created row to the
3069 sub update_or_create {
3071 my $attrs = (@_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {});
3072 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
3074 my $row = $self->find($cond, $attrs);
3076 $row->update($cond);
3080 return $self->new_result($cond)->insert;
3083 =head2 update_or_new
3087 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
3089 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
3093 $resultset->update_or_new({ col => $val, ... });
3095 Like L</find_or_new> but if a row is found it is immediately updated via
3096 C<< $found_row->update (\%col_data) >>.
3100 # In your application
3101 my $cd = $schema->resultset('CD')->update_or_new(
3103 artist => 'Massive Attack',
3104 title => 'Mezzanine',
3107 { key => 'cd_artist_title' }
3110 if ($cd->in_storage) {
3111 # the cd was updated
3114 # the cd is not yet in the database, let's insert it
3118 B<Note>: Make sure to read the documentation of L</find> and understand the
3119 significance of the C<key> attribute, as its lack may skew your search, and
3120 subsequently result in spurious new objects.
3122 B<Note>: Take care when using C<update_or_new> with a table having
3123 columns with default values that you intend to be automatically
3124 supplied by the database (e.g. an auto_increment primary key column).
3125 In normal usage, the value of such columns should NOT be included at
3126 all in the call to C<update_or_new>, even when set to C<undef>.
3128 See also L</find>, L</find_or_create> and L</find_or_new>.
3134 my $attrs = ( @_ > 1 && ref $_[-1] eq 'HASH' ? pop(@_) : {} );
3135 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
3137 my $row = $self->find( $cond, $attrs );
3138 if ( defined $row ) {
3139 $row->update($cond);
3143 return $self->new_result($cond);
3150 =item Arguments: none
3152 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass> | undef
3156 Gets the contents of the cache for the resultset, if the cache is set.
3158 The cache is populated either by using the L</prefetch> attribute to
3159 L</search> or by calling L</set_cache>.
3171 =item Arguments: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3173 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3177 Sets the contents of the cache for the resultset. Expects an arrayref
3178 of objects of the same class as those produced by the resultset. Note that
3179 if the cache is set, the resultset will return the cached objects rather
3180 than re-querying the database even if the cache attr is not set.
3182 The contents of the cache can also be populated by using the
3183 L</prefetch> attribute to L</search>.
3188 my ( $self, $data ) = @_;
3189 $self->throw_exception("set_cache requires an arrayref")
3190 if defined($data) && (ref $data ne 'ARRAY');
3191 $self->{all_cache} = $data;
3198 =item Arguments: none
3200 =item Return Value: undef
3204 Clears the cache for the resultset.
3209 shift->set_cache(undef);
3216 =item Arguments: none
3218 =item Return Value: true, if the resultset has been paginated
3226 return !!$self->{attrs}{page};
3233 =item Arguments: none
3235 =item Return Value: true, if the resultset has been ordered with C<order_by>.
3243 return scalar $self->result_source->schema->storage->_extract_order_criteria($self->{attrs}{order_by});
3246 =head2 related_resultset
3250 =item Arguments: $rel_name
3252 =item Return Value: L<$resultset|/search>
3256 Returns a related resultset for the supplied relationship name.
3258 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
3262 sub related_resultset {
3263 $_[0]->throw_exception(
3264 'Extra arguments to $rs->related_resultset() were always quietly '
3265 . 'discarded without consideration, you need to switch to '
3266 . '...->related_resultset( $relname )->search_rs( $search, $args ) instead.'
3269 return $_[0]->{related_resultsets}{$_[1]}
3270 if defined $_[0]->{related_resultsets}{$_[1]};
3272 my ($self, $rel) = @_;
3274 return $self->{related_resultsets}{$rel} = do {
3275 my $rsrc = $self->result_source;
3276 my $rel_info = $rsrc->relationship_info($rel);
3278 $self->throw_exception(
3279 "search_related: result source '" . $rsrc->source_name .
3280 "' has no such relationship $rel")
3283 my $attrs = $self->_chain_relationship($rel);
3285 # Previously this atribute was deleted (instead of being set as it is now)
3286 # Doing so seems to be harmless in all available test permutations
3287 # See also 01d59a6a6 and mst's comment below
3289 $attrs->{alias} = $rsrc->schema->storage->relname_to_table_alias(
3291 $attrs->{seen_join}{$rel}
3294 # since this is search_related, and we already slid the select window inwards
3295 # (the select/as attrs were deleted in the beginning), we need to flip all
3296 # left joins to inner, so we get the expected results
3298 # The DBIC relationship chaining implementation is pretty simple - every
3299 # new related_relationship is pushed onto the {from} stack, and the {select}
3300 # window simply slides further in. This means that when we count somewhere
3301 # in the middle, we got to make sure that everything in the join chain is an
3302 # actual inner join, otherwise the count will come back with unpredictable
3303 # results (a resultset may be generated with _some_ rows regardless of if
3304 # the relation which the $rs currently selects has rows or not). E.g.
3305 # $artist_rs->cds->count - normally generates:
3306 # SELECT COUNT( * ) FROM artist me LEFT JOIN cd cds ON cds.artist = me.artistid
3307 # which actually returns the number of artists * (number of cds || 1)
3309 # So what we do here is crawl {from}, determine if the current alias is at
3310 # the top of the stack, and if not - make sure the chain is inner-joined down
3313 my $switch_branch = find_join_path_to_alias(
3318 if ( @{ $switch_branch || [] } ) {
3320 # So it looks like we will have to switch some stuff around.
3321 # local() is useless here as we will be leaving the scope
3322 # anyway, and deep cloning is just too fucking expensive
3323 # So replace the first hashref in the node arrayref manually
3324 my @new_from = $attrs->{from}[0];
3325 my $sw_idx = { map { (values %$_), 1 } @$switch_branch }; #there's one k/v per join-path
3327 for my $j ( @{$attrs->{from}}[ 1 .. $#{$attrs->{from}} ] ) {
3328 my $jalias = $j->[0]{-alias};
3330 if ($sw_idx->{$jalias}) {
3331 my %attrs = %{$j->[0]};
3332 delete $attrs{-join_type};
3343 $attrs->{from} = \@new_from;
3347 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
3348 delete $attrs->{result_class};
3352 # The reason we do this now instead of passing the alias to the
3353 # search_rs below is that if you wrap/overload resultset on the
3354 # source you need to know what alias it's -going- to have for things
3355 # to work sanely (e.g. RestrictWithObject wants to be able to add
3356 # extra query restrictions, and these may need to be $alias.)
3357 # -- mst ~ 2007 (01d59a6a6)
3359 # FIXME - this seems to be no longer neccessary (perhaps due to the
3360 # advances in relcond resolution. Testing DBIC::S::RWO and its only
3361 # dependent (as of Jun 2015 ) does not yield any difference with or
3362 # without this line. Nevertheless keep it as is for now, to minimize
3363 # churn, there is enough potential for breakage in 0.0829xx as it is
3364 # -- ribasushi Jun 2015
3366 my $rel_source = $rsrc->related_source($rel);
3367 local $rel_source->resultset_attributes->{alias} = $attrs->{alias};
3369 $rel_source->resultset->search_rs( undef, $attrs );
3372 if (my $cache = $self->get_cache) {
3373 my @related_cache = map
3374 { $_->related_resultset($rel)->get_cache || () }
3378 $new->set_cache([ map @$_, @related_cache ]) if @related_cache == @$cache;
3385 =head2 current_source_alias
3389 =item Arguments: none
3391 =item Return Value: $source_alias
3395 Returns the current table alias for the result source this resultset is built
3396 on, that will be used in the SQL query. Usually it is C<me>.
3398 Currently the source alias that refers to the result set returned by a
3399 L</search>/L</find> family method depends on how you got to the resultset: it's
3400 C<me> by default, but eg. L</search_related> aliases it to the related result
3401 source name (and keeps C<me> referring to the original result set). The long
3402 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
3403 (and make this method unnecessary).
3405 Thus it's currently necessary to use this method in predefined queries (see
3406 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
3407 source alias of the current result set:
3409 # in a result set class
3411 my ($self, $user) = @_;
3413 my $me = $self->current_source_alias;
3415 return $self->search({
3416 "$me.modified" => $user->id,
3420 The alias of L<newly created resultsets|/search> can be altered by the
3421 L<alias attribute|/alias>.
3425 sub current_source_alias {
3426 return (shift->{attrs} || {})->{alias} || 'me';
3429 =head2 as_subselect_rs
3433 =item Arguments: none
3435 =item Return Value: L<$resultset|/search>
3439 Act as a barrier to SQL symbols. The resultset provided will be made into a
3440 "virtual view" by including it as a subquery within the from clause. From this
3441 point on, any joined tables are inaccessible to ->search on the resultset (as if
3442 it were simply where-filtered without joins). For example:
3444 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
3446 # 'x' now pollutes the query namespace
3448 # So the following works as expected
3449 my $ok_rs = $rs->search({'x.other' => 1});
3451 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
3452 # def) we look for one row with contradictory terms and join in another table
3453 # (aliased 'x_2') which we never use
3454 my $broken_rs = $rs->search({'x.name' => 'def'});
3456 my $rs2 = $rs->as_subselect_rs;
3458 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
3459 my $not_joined_rs = $rs2->search({'x.other' => 1});
3461 # works as expected: finds a 'table' row related to two x rows (abc and def)
3462 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
3464 Another example of when one might use this would be to select a subset of
3465 columns in a group by clause:
3467 my $rs = $schema->resultset('Bar')->search(undef, {
3468 group_by => [qw{ id foo_id baz_id }],
3469 })->as_subselect_rs->search(undef, {
3470 columns => [qw{ id foo_id }]
3473 In the above example normally columns would have to be equal to the group by,
3474 but because we isolated the group by into a subselect the above works.
3478 sub as_subselect_rs {
3481 my $attrs = $self->_resolved_attrs;
3483 my $fresh_rs = (ref $self)->new (
3484 $self->result_source
3487 # these pieces will be locked in the subquery
3488 delete $fresh_rs->{cond};
3489 delete @{$fresh_rs->{attrs}}{qw/where bind/};
3491 return $fresh_rs->search( {}, {
3493 $attrs->{alias} => $self->as_query,
3494 -alias => $attrs->{alias},
3495 -rsrc => $self->result_source,
3497 alias => $attrs->{alias},
3501 # This code is called by search_related, and makes sure there
3502 # is clear separation between the joins before, during, and
3503 # after the relationship. This information is needed later
3504 # in order to properly resolve prefetch aliases (any alias
3505 # with a relation_chain_depth less than the depth of the
3506 # current prefetch is not considered)
3508 # The increments happen twice per join. An even number means a
3509 # relationship specified via a search_related, whereas an odd
3510 # number indicates a join/prefetch added via attributes
3512 # Also this code will wrap the current resultset (the one we
3513 # chain to) in a subselect IFF it contains limiting attributes
3514 sub _chain_relationship {
3515 my ($self, $rel) = @_;
3516 my $source = $self->result_source;
3517 my $attrs = { %{$self->{attrs}||{}} };
3519 # we need to take the prefetch the attrs into account before we
3520 # ->_resolve_join as otherwise they get lost - captainL
3521 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
3523 delete @{$attrs}{qw/join prefetch collapse group_by distinct _grouped_by_distinct select as columns +select +as +columns/};
3525 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
3528 my @force_subq_attrs = qw/offset rows group_by having/;
3531 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
3533 $self->_has_resolved_attr (@force_subq_attrs)
3535 # Nuke the prefetch (if any) before the new $rs attrs
3536 # are resolved (prefetch is useless - we are wrapping
3537 # a subquery anyway).
3538 my $rs_copy = $self->search;
3539 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
3540 $rs_copy->{attrs}{join},
3541 delete $rs_copy->{attrs}{prefetch},
3546 -alias => $attrs->{alias},
3547 $attrs->{alias} => $rs_copy->as_query,
3549 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
3550 $seen->{-relation_chain_depth} = 0;
3552 elsif ($attrs->{from}) { #shallow copy suffices
3553 $from = [ @{$attrs->{from}} ];
3558 -alias => $attrs->{alias},
3559 $attrs->{alias} => $source->from,
3563 my $jpath = ($seen->{-relation_chain_depth})
3564 ? $from->[-1][0]{-join_path}
3567 my @requested_joins = $source->_resolve_join(
3574 push @$from, @requested_joins;
3576 $seen->{-relation_chain_depth}++;
3578 # if $self already had a join/prefetch specified on it, the requested
3579 # $rel might very well be already included. What we do in this case
3580 # is effectively a no-op (except that we bump up the chain_depth on
3581 # the join in question so we could tell it *is* the search_related)
3584 # we consider the last one thus reverse
3585 for my $j (reverse @requested_joins) {
3586 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
3587 if ($rel eq $last_j) {
3588 $j->[0]{-relation_chain_depth}++;
3594 unless ($already_joined) {
3595 push @$from, $source->_resolve_join(
3603 $seen->{-relation_chain_depth}++;
3605 return {%$attrs, from => $from, seen_join => $seen};
3608 sub _resolved_attrs {
3610 return $self->{_attrs} if $self->{_attrs};
3612 my $attrs = { %{ $self->{attrs} || {} } };
3613 my $source = $attrs->{result_source} = $self->result_source;
3614 my $alias = $attrs->{alias};
3616 $self->throw_exception("Specifying distinct => 1 in conjunction with collapse => 1 is unsupported")
3617 if $attrs->{collapse} and $attrs->{distinct};
3620 # Sanity check the paging attributes
3621 # SQLMaker does it too, but in case of a software_limit we'll never get there
3622 if (defined $attrs->{offset}) {
3623 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
3624 if ( $attrs->{offset} =~ /[^0-9]/ or $attrs->{offset} < 0 );
3626 if (defined $attrs->{rows}) {
3627 $self->throw_exception("The rows attribute must be a positive integer if present")
3628 if ( $attrs->{rows} =~ /[^0-9]/ or $attrs->{rows} <= 0 );
3631 # normalize where condition
3632 $attrs->{where} = normalize_sqla_condition( $attrs->{where} )
3635 # default selection list
3636 $attrs->{columns} = [ $source->columns ]
3637 unless grep { exists $attrs->{$_} } qw/columns cols select as/;
3639 # merge selectors together
3640 for (qw/columns select as/) {
3641 $attrs->{$_} = $self->_merge_attr($attrs->{$_}, delete $attrs->{"+$_"})
3642 if $attrs->{$_} or $attrs->{"+$_"};
3645 # disassemble columns
3647 if (my $cols = delete $attrs->{columns}) {
3648 for my $c (ref $cols eq 'ARRAY' ? @$cols : $cols) {
3649 if (ref $c eq 'HASH') {
3650 for my $as (sort keys %$c) {
3651 push @sel, $c->{$as};
3662 # when trying to weed off duplicates later do not go past this point -
3663 # everything added from here on is unbalanced "anyone's guess" stuff
3664 my $dedup_stop_idx = $#as;
3666 push @as, @{ ref $attrs->{as} eq 'ARRAY' ? $attrs->{as} : [ $attrs->{as} ] }
3668 push @sel, @{ ref $attrs->{select} eq 'ARRAY' ? $attrs->{select} : [ $attrs->{select} ] }
3669 if $attrs->{select};
3671 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3672 $_ = (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" for @sel;
3674 # disqualify all $alias.col as-bits (inflate-map mandated)
3675 $_ = ($_ =~ /^\Q$alias.\E(.+)$/) ? $1 : $_ for @as;
3677 # de-duplicate the result (remove *identical* select/as pairs)
3678 # and also die on duplicate {as} pointing to different {select}s
3679 # not using a c-style for as the condition is prone to shrinkage
3682 while ($i <= $dedup_stop_idx) {
3683 if ($seen->{"$sel[$i] \x00\x00 $as[$i]"}++) {
3688 elsif ($seen->{$as[$i]}++) {
3689 $self->throw_exception(
3690 "inflate_result() alias '$as[$i]' specified twice with different SQL-side {select}-ors"
3698 $attrs->{select} = \@sel;
3699 $attrs->{as} = \@as;
3701 $attrs->{from} ||= [{
3703 -alias => $self->{attrs}{alias},
3704 $self->{attrs}{alias} => $source->from,
3707 if ( $attrs->{join} || $attrs->{prefetch} ) {
3709 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3710 if ref $attrs->{from} ne 'ARRAY';
3712 my $join = (delete $attrs->{join}) || {};
3714 if ( defined $attrs->{prefetch} ) {
3715 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3718 $attrs->{from} = # have to copy here to avoid corrupting the original
3720 @{ $attrs->{from} },
3721 $source->_resolve_join(
3724 { %{ $attrs->{seen_join} || {} } },
3725 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3726 ? $attrs->{from}[-1][0]{-join_path}
3734 for my $attr (qw(order_by group_by)) {
3736 if ( defined $attrs->{$attr} ) {
3738 ref( $attrs->{$attr} ) eq 'ARRAY'
3739 ? [ @{ $attrs->{$attr} } ]
3740 : [ $attrs->{$attr} || () ]
3743 delete $attrs->{$attr} unless @{$attrs->{$attr}};
3748 # set collapse default based on presence of prefetch
3751 defined $attrs->{prefetch}
3753 $prefetch = $self->_merge_joinpref_attr( {}, delete $attrs->{prefetch} )
3755 $self->throw_exception("Specifying prefetch in conjunction with an explicit collapse => 0 is unsupported")
3756 if defined $attrs->{collapse} and ! $attrs->{collapse};
3758 $attrs->{collapse} = 1;
3762 # run through the resulting joinstructure (starting from our current slot)
3763 # and unset collapse if proven unnecessary
3765 # also while we are at it find out if the current root source has
3766 # been premultiplied by previous related_source chaining
3768 # this allows to predict whether a root object with all other relation
3769 # data set to NULL is in fact unique
3770 if ($attrs->{collapse}) {
3772 if (ref $attrs->{from} eq 'ARRAY') {
3774 if (@{$attrs->{from}} == 1) {
3775 # no joins - no collapse
3776 $attrs->{collapse} = 0;
3779 # find where our table-spec starts
3780 my @fromlist = @{$attrs->{from}};
3782 my $t = shift @fromlist;
3785 # me vs join from-spec distinction - a ref means non-root
3786 if (ref $t eq 'ARRAY') {
3788 $is_multi ||= ! $t->{-is_single};
3790 last if ($t->{-alias} && $t->{-alias} eq $alias);
3791 $attrs->{_main_source_premultiplied} ||= $is_multi;
3794 # no non-singles remaining, nor any premultiplication - nothing to collapse
3796 ! $attrs->{_main_source_premultiplied}
3798 ! grep { ! $_->[0]{-is_single} } @fromlist
3800 $attrs->{collapse} = 0;
3806 # if we can not analyze the from - err on the side of safety
3807 $attrs->{_main_source_premultiplied} = 1;
3812 # generate the distinct induced group_by before injecting the prefetched select/as parts
3813 if (delete $attrs->{distinct}) {
3814 if ($attrs->{group_by}) {
3815 carp_unique ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3818 $attrs->{_grouped_by_distinct} = 1;
3819 # distinct affects only the main selection part, not what prefetch may add below
3820 ($attrs->{group_by}, my $new_order) = $source->schema->storage->_group_over_selection($attrs);
3822 # FIXME possibly ignore a rewritten order_by (may turn out to be an issue)
3823 # The thinking is: if we are collapsing the subquerying prefetch engine will
3824 # rip stuff apart for us anyway, and we do not want to have a potentially
3825 # function-converted external order_by
3826 # ( there is an explicit if ( collapse && _grouped_by_distinct ) check in DBIHacks )
3827 $attrs->{order_by} = $new_order unless $attrs->{collapse};
3832 # generate selections based on the prefetch helper
3835 $self->throw_exception("Unable to prefetch, resultset contains an unnamed selector $attrs->{_dark_selector}{string}")
3836 if $attrs->{_dark_selector};
3838 # this is a separate structure (we don't look in {from} directly)
3839 # as the resolver needs to shift things off the lists to work
3840 # properly (identical-prefetches on different branches)
3841 my $joined_node_aliases_map = {};
3842 if (ref $attrs->{from} eq 'ARRAY') {
3844 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3846 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3847 next unless $j->[0]{-alias};
3848 next unless $j->[0]{-join_path};
3849 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3851 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3853 my $p = $joined_node_aliases_map;
3854 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3855 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3859 ( push @{$attrs->{select}}, $_->[0] ) and ( push @{$attrs->{as}}, $_->[1] )
3860 for $source->_resolve_selection_from_prefetch( $prefetch, $joined_node_aliases_map );
3864 $attrs->{_simple_passthrough_construction} = !(
3867 grep { $_ =~ /\./ } @{$attrs->{as}}
3871 # if both page and offset are specified, produce a combined offset
3872 # even though it doesn't make much sense, this is what pre 081xx has
3874 if (my $page = delete $attrs->{page}) {
3876 ($attrs->{rows} * ($page - 1))
3878 ($attrs->{offset} || 0)
3882 return $self->{_attrs} = $attrs;
3886 my ($self, $attr) = @_;
3888 if (ref $attr eq 'HASH') {
3889 return $self->_rollout_hash($attr);
3890 } elsif (ref $attr eq 'ARRAY') {
3891 return $self->_rollout_array($attr);
3897 sub _rollout_array {
3898 my ($self, $attr) = @_;
3901 foreach my $element (@{$attr}) {
3902 if (ref $element eq 'HASH') {
3903 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3904 } elsif (ref $element eq 'ARRAY') {
3905 # XXX - should probably recurse here
3906 push( @rolled_array, @{$self->_rollout_array($element)} );
3908 push( @rolled_array, $element );
3911 return \@rolled_array;
3915 my ($self, $attr) = @_;
3918 foreach my $key (keys %{$attr}) {
3919 push( @rolled_array, { $key => $attr->{$key} } );
3921 return \@rolled_array;
3924 sub _calculate_score {
3925 my ($self, $a, $b) = @_;
3927 if (defined $a xor defined $b) {
3930 elsif (not defined $a) {
3934 if (ref $b eq 'HASH') {
3935 my ($b_key) = keys %{$b};
3936 $b_key = '' if ! defined $b_key;
3937 if (ref $a eq 'HASH') {
3938 my ($a_key) = keys %{$a};
3939 $a_key = '' if ! defined $a_key;
3940 if ($a_key eq $b_key) {
3941 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3946 return ($a eq $b_key) ? 1 : 0;
3949 if (ref $a eq 'HASH') {
3950 my ($a_key) = keys %{$a};
3951 return ($b eq $a_key) ? 1 : 0;
3953 return ($b eq $a) ? 1 : 0;
3958 sub _merge_joinpref_attr {
3959 my ($self, $orig, $import) = @_;
3961 return $import unless defined($orig);
3962 return $orig unless defined($import);
3964 $orig = $self->_rollout_attr($orig);
3965 $import = $self->_rollout_attr($import);
3968 foreach my $import_element ( @{$import} ) {
3969 # find best candidate from $orig to merge $b_element into
3970 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3971 foreach my $orig_element ( @{$orig} ) {
3972 my $score = $self->_calculate_score( $orig_element, $import_element );
3973 if ($score > $best_candidate->{score}) {
3974 $best_candidate->{position} = $position;
3975 $best_candidate->{score} = $score;
3979 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3980 $import_key = '' if not defined $import_key;
3982 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3983 push( @{$orig}, $import_element );
3985 my $orig_best = $orig->[$best_candidate->{position}];
3986 # merge orig_best and b_element together and replace original with merged
3987 if (ref $orig_best ne 'HASH') {
3988 $orig->[$best_candidate->{position}] = $import_element;
3989 } elsif (ref $import_element eq 'HASH') {
3990 my ($key) = keys %{$orig_best};
3991 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3994 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3997 return @$orig ? $orig : ();
4005 require Hash::Merge;
4006 my $hm = Hash::Merge->new;
4008 $hm->specify_behavior({
4011 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
4013 if ($defl xor $defr) {
4014 return [ $defl ? $_[0] : $_[1] ];
4019 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
4023 return [$_[0], $_[1]];
4027 return $_[1] if !defined $_[0];
4028 return $_[1] if __HM_DEDUP and grep { $_ eq $_[0] } @{$_[1]};
4029 return [$_[0], @{$_[1]}]
4032 return [] if !defined $_[0] and !keys %{$_[1]};
4033 return [ $_[1] ] if !defined $_[0];
4034 return [ $_[0] ] if !keys %{$_[1]};
4035 return [$_[0], $_[1]]
4040 return $_[0] if !defined $_[1];
4041 return $_[0] if __HM_DEDUP and grep { $_ eq $_[1] } @{$_[0]};
4042 return [@{$_[0]}, $_[1]]
4045 my @ret = @{$_[0]} or return $_[1];
4046 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
4047 my %idx = map { $_ => 1 } @ret;
4048 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
4052 return [ $_[1] ] if ! @{$_[0]};
4053 return $_[0] if !keys %{$_[1]};
4054 return $_[0] if __HM_DEDUP and grep { $_ eq $_[1] } @{$_[0]};
4055 return [ @{$_[0]}, $_[1] ];
4060 return [] if !keys %{$_[0]} and !defined $_[1];
4061 return [ $_[0] ] if !defined $_[1];
4062 return [ $_[1] ] if !keys %{$_[0]};
4063 return [$_[0], $_[1]]
4066 return [] if !keys %{$_[0]} and !@{$_[1]};
4067 return [ $_[0] ] if !@{$_[1]};
4068 return $_[1] if !keys %{$_[0]};
4069 return $_[1] if __HM_DEDUP and grep { $_ eq $_[0] } @{$_[1]};
4070 return [ $_[0], @{$_[1]} ];
4073 return [] if !keys %{$_[0]} and !keys %{$_[1]};
4074 return [ $_[0] ] if !keys %{$_[1]};
4075 return [ $_[1] ] if !keys %{$_[0]};
4076 return [ $_[0] ] if $_[0] eq $_[1];
4077 return [ $_[0], $_[1] ];
4080 } => 'DBIC_RS_ATTR_MERGER');
4084 return $hm->merge ($_[1], $_[2]);
4088 sub STORABLE_freeze {
4089 my ($self, $cloning) = @_;
4090 my $to_serialize = { %$self };
4092 # A cursor in progress can't be serialized (and would make little sense anyway)
4093 # the parser can be regenerated (and can't be serialized)
4094 delete @{$to_serialize}{qw/cursor _row_parser _result_inflator/};
4096 # nor is it sensical to store a not-yet-fired-count pager
4097 if ($to_serialize->{pager} and ref $to_serialize->{pager}{total_entries} eq 'CODE') {
4098 delete $to_serialize->{pager};
4101 Storable::nfreeze($to_serialize);
4104 # need this hook for symmetry
4106 my ($self, $cloning, $serialized) = @_;
4108 %$self = %{ Storable::thaw($serialized) };
4114 =head2 throw_exception
4116 See L<DBIx::Class::Schema/throw_exception> for details.
4120 sub throw_exception {
4123 if (ref $self and my $rsrc = $self->result_source) {
4124 $rsrc->throw_exception(@_)
4127 DBIx::Class::Exception->throw(@_);
4135 # XXX: FIXME: Attributes docs need clearing up
4139 Attributes are used to refine a ResultSet in various ways when
4140 searching for data. They can be passed to any method which takes an
4141 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
4144 Default attributes can be set on the result class using
4145 L<DBIx::Class::ResultSource/resultset_attributes>. (Please read
4146 the CAVEATS on that feature before using it!)
4148 These are in no particular order:
4154 =item Value: ( $order_by | \@order_by | \%order_by )
4158 Which column(s) to order the results by.
4160 [The full list of suitable values is documented in
4161 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
4164 If a single column name, or an arrayref of names is supplied, the
4165 argument is passed through directly to SQL. The hashref syntax allows
4166 for connection-agnostic specification of ordering direction:
4168 For descending order:
4170 order_by => { -desc => [qw/col1 col2 col3/] }
4172 For explicit ascending order:
4174 order_by => { -asc => 'col' }
4176 The old scalarref syntax (i.e. order_by => \'year DESC') is still
4177 supported, although you are strongly encouraged to use the hashref
4178 syntax as outlined above.
4184 =item Value: \@columns | \%columns | $column
4188 Shortcut to request a particular set of columns to be retrieved. Each
4189 column spec may be a string (a table column name), or a hash (in which
4190 case the key is the C<as> value, and the value is used as the C<select>
4191 expression). Adds the L</current_source_alias> onto the start of any column without a C<.> in
4192 it and sets C<select> from that, then auto-populates C<as> from
4193 C<select> as normal. (You may also use the C<cols> attribute, as in
4194 earlier versions of DBIC, but this is deprecated)
4196 Essentially C<columns> does the same as L</select> and L</as>.
4198 columns => [ 'some_column', { dbic_slot => 'another_column' } ]
4202 select => [qw(some_column another_column)],
4203 as => [qw(some_column dbic_slot)]
4205 If you want to individually retrieve related columns (in essence perform
4206 manual L</prefetch>) you have to make sure to specify the correct inflation slot
4207 chain such that it matches existing relationships:
4209 my $rs = $schema->resultset('Artist')->search({}, {
4210 # required to tell DBIC to collapse has_many relationships
4212 join => { cds => 'tracks' },
4214 'cds.cdid' => 'cds.cdid',
4215 'cds.tracks.title' => 'tracks.title',
4219 Like elsewhere, literal SQL or literal values can be included by using a
4220 scalar reference or a literal bind value, and these values will be available
4221 in the result with C<get_column> (see also
4222 L<SQL::Abstract/Literal SQL and value type operators>):
4224 # equivalent SQL: SELECT 1, 'a string', IF(my_column,?,?) ...
4225 # bind values: $true_value, $false_value
4229 bar => \q{'a string'},
4230 baz => \[ 'IF(my_column,?,?)', $true_value, $false_value ],
4236 B<NOTE:> You B<MUST> explicitly quote C<'+columns'> when using this attribute.
4237 Not doing so causes Perl to incorrectly interpret C<+columns> as a bareword
4238 with a unary plus operator before it, which is the same as simply C<columns>.
4242 =item Value: \@extra_columns
4246 Indicates additional columns to be selected from storage. Works the same as
4247 L</columns> but adds columns to the current selection. (You may also use the
4248 C<include_columns> attribute, as in earlier versions of DBIC, but this is
4251 $schema->resultset('CD')->search(undef, {
4252 '+columns' => ['artist.name'],
4256 would return all CDs and include a 'name' column to the information
4257 passed to object inflation. Note that the 'artist' is the name of the
4258 column (or relationship) accessor, and 'name' is the name of the column
4259 accessor in the related table.
4265 =item Value: \@select_columns
4269 Indicates which columns should be selected from the storage. You can use
4270 column names, or in the case of RDBMS back ends, function or stored procedure
4273 $rs = $schema->resultset('Employee')->search(undef, {
4276 { count => 'employeeid' },
4277 { max => { length => 'name' }, -as => 'longest_name' }
4282 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
4284 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
4285 use L</select>, to instruct DBIx::Class how to store the result of the column.
4287 Also note that the L</as> attribute has B<nothing to do> with the SQL-side
4288 C<AS> identifier aliasing. You B<can> alias a function (so you can use it e.g.
4289 in an C<ORDER BY> clause), however this is done via the C<-as> B<select
4290 function attribute> supplied as shown in the example above.
4294 B<NOTE:> You B<MUST> explicitly quote C<'+select'> when using this attribute.
4295 Not doing so causes Perl to incorrectly interpret C<+select> as a bareword
4296 with a unary plus operator before it, which is the same as simply C<select>.
4300 =item Value: \@extra_select_columns
4304 Indicates additional columns to be selected from storage. Works the same as
4305 L</select> but adds columns to the current selection, instead of specifying
4306 a new explicit list.
4312 =item Value: \@inflation_names
4316 Indicates DBIC-side names for object inflation. That is L</as> indicates the
4317 slot name in which the column value will be stored within the
4318 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
4319 identifier by the C<get_column> method (or via the object accessor B<if one
4320 with the same name already exists>) as shown below.
4322 The L</as> attribute has B<nothing to do> with the SQL-side identifier
4323 aliasing C<AS>. See L</select> for details.
4325 $rs = $schema->resultset('Employee')->search(undef, {
4328 { count => 'employeeid' },
4329 { max => { length => 'name' }, -as => 'longest_name' }
4338 If the object against which the search is performed already has an accessor
4339 matching a column name specified in C<as>, the value can be retrieved using
4340 the accessor as normal:
4342 my $name = $employee->name();
4344 If on the other hand an accessor does not exist in the object, you need to
4345 use C<get_column> instead:
4347 my $employee_count = $employee->get_column('employee_count');
4349 You can create your own accessors if required - see
4350 L<DBIx::Class::Manual::Cookbook> for details.
4354 B<NOTE:> You B<MUST> explicitly quote C<'+as'> when using this attribute.
4355 Not doing so causes Perl to incorrectly interpret C<+as> as a bareword
4356 with a unary plus operator before it, which is the same as simply C<as>.
4360 =item Value: \@extra_inflation_names
4364 Indicates additional inflation names for selectors added via L</+select>. See L</as>.
4370 =item Value: ($rel_name | \@rel_names | \%rel_names)
4374 Contains a list of relationships that should be joined for this query. For
4377 # Get CDs by Nine Inch Nails
4378 my $rs = $schema->resultset('CD')->search(
4379 { 'artist.name' => 'Nine Inch Nails' },
4380 { join => 'artist' }
4383 Can also contain a hash reference to refer to the other relation's relations.
4386 package MyApp::Schema::Track;
4387 use base qw/DBIx::Class/;
4388 __PACKAGE__->table('track');
4389 __PACKAGE__->add_columns(qw/trackid cd position title/);
4390 __PACKAGE__->set_primary_key('trackid');
4391 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
4394 # In your application
4395 my $rs = $schema->resultset('Artist')->search(
4396 { 'track.title' => 'Teardrop' },
4398 join => { cd => 'track' },
4399 order_by => 'artist.name',
4403 You need to use the relationship (not the table) name in conditions,
4404 because they are aliased as such. The current table is aliased as "me", so
4405 you need to use me.column_name in order to avoid ambiguity. For example:
4407 # Get CDs from 1984 with a 'Foo' track
4408 my $rs = $schema->resultset('CD')->search(
4411 'tracks.name' => 'Foo'
4413 { join => 'tracks' }
4416 If the same join is supplied twice, it will be aliased to <rel>_2 (and
4417 similarly for a third time). For e.g.
4419 my $rs = $schema->resultset('Artist')->search({
4420 'cds.title' => 'Down to Earth',
4421 'cds_2.title' => 'Popular',
4423 join => [ qw/cds cds/ ],
4426 will return a set of all artists that have both a cd with title 'Down
4427 to Earth' and a cd with title 'Popular'.
4429 If you want to fetch related objects from other tables as well, see L</prefetch>
4432 NOTE: An internal join-chain pruner will discard certain joins while
4433 constructing the actual SQL query, as long as the joins in question do not
4434 affect the retrieved result. This for example includes 1:1 left joins
4435 that are not part of the restriction specification (WHERE/HAVING) nor are
4436 a part of the query selection.
4438 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
4444 =item Value: (0 | 1)
4448 When set to a true value, indicates that any rows fetched from joined has_many
4449 relationships are to be aggregated into the corresponding "parent" object. For
4450 example, the resultset:
4452 my $rs = $schema->resultset('CD')->search({}, {
4453 '+columns' => [ qw/ tracks.title tracks.position / ],
4458 While executing the following query:
4460 SELECT me.*, tracks.title, tracks.position
4462 LEFT JOIN track tracks
4463 ON tracks.cdid = me.cdid
4465 Will return only as many objects as there are rows in the CD source, even
4466 though the result of the query may span many rows. Each of these CD objects
4467 will in turn have multiple "Track" objects hidden behind the has_many
4468 generated accessor C<tracks>. Without C<< collapse => 1 >>, the return values
4469 of this resultset would be as many CD objects as there are tracks (a "Cartesian
4470 product"), with each CD object containing exactly one of all fetched Track data.
4472 When a collapse is requested on a non-ordered resultset, an order by some
4473 unique part of the main source (the left-most table) is inserted automatically.
4474 This is done so that the resultset is allowed to be "lazy" - calling
4475 L<< $rs->next|/next >> will fetch only as many rows as it needs to build the next
4476 object with all of its related data.
4478 If an L</order_by> is already declared, and orders the resultset in a way that
4479 makes collapsing as described above impossible (e.g. C<< ORDER BY
4480 has_many_rel.column >> or C<ORDER BY RANDOM()>), DBIC will automatically
4481 switch to "eager" mode and slurp the entire resultset before constructing the
4482 first object returned by L</next>.
4484 Setting this attribute on a resultset that does not join any has_many
4485 relations is a no-op.
4487 For a more in-depth discussion, see L</PREFETCHING>.
4493 =item Value: ($rel_name | \@rel_names | \%rel_names)
4497 This attribute is a shorthand for specifying a L</join> spec, adding all
4498 columns from the joined related sources as L</+columns> and setting
4499 L</collapse> to a true value. It can be thought of as a rough B<superset>
4500 of the L</join> attribute.
4502 For example, the following two queries are equivalent:
4504 my $rs = $schema->resultset('Artist')->search({}, {
4505 prefetch => { cds => ['genre', 'tracks' ] },
4510 my $rs = $schema->resultset('Artist')->search({}, {
4511 join => { cds => ['genre', 'tracks' ] },
4515 { +{ "cds.$_" => "cds.$_" } }
4516 $schema->source('Artist')->related_source('cds')->columns
4519 { +{ "cds.genre.$_" => "genre.$_" } }
4520 $schema->source('Artist')->related_source('cds')->related_source('genre')->columns
4523 { +{ "cds.tracks.$_" => "tracks.$_" } }
4524 $schema->source('Artist')->related_source('cds')->related_source('tracks')->columns
4529 Both producing the following SQL:
4531 SELECT me.artistid, me.name, me.rank, me.charfield,
4532 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track,
4533 genre.genreid, genre.name,
4534 tracks.trackid, tracks.cd, tracks.position, tracks.title, tracks.last_updated_on, tracks.last_updated_at
4537 ON cds.artist = me.artistid
4538 LEFT JOIN genre genre
4539 ON genre.genreid = cds.genreid
4540 LEFT JOIN track tracks
4541 ON tracks.cd = cds.cdid
4542 ORDER BY me.artistid
4544 While L</prefetch> implies a L</join>, it is ok to mix the two together, as
4545 the arguments are properly merged and generally do the right thing. For
4546 example, you may want to do the following:
4548 my $artists_and_cds_without_genre = $schema->resultset('Artist')->search(
4549 { 'genre.genreid' => undef },
4551 join => { cds => 'genre' },
4556 Which generates the following SQL:
4558 SELECT me.artistid, me.name, me.rank, me.charfield,
4559 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track
4562 ON cds.artist = me.artistid
4563 LEFT JOIN genre genre
4564 ON genre.genreid = cds.genreid
4565 WHERE genre.genreid IS NULL
4566 ORDER BY me.artistid
4568 For a more in-depth discussion, see L</PREFETCHING>.
4574 =item Value: $source_alias
4578 Sets the source alias for the query. Normally, this defaults to C<me>, but
4579 nested search queries (sub-SELECTs) might need specific aliases set to
4580 reference inner queries. For example:
4583 ->related_resultset('CDs')
4584 ->related_resultset('Tracks')
4586 'track.id' => { -ident => 'none_search.id' },
4590 my $ids = $self->search({
4593 alias => 'none_search',
4594 group_by => 'none_search.id',
4595 })->get_column('id')->as_query;
4597 $self->search({ id => { -in => $ids } })
4599 This attribute is directly tied to L</current_source_alias>.
4609 Makes the resultset paged and specifies the page to retrieve. Effectively
4610 identical to creating a non-pages resultset and then calling ->page($page)
4613 If L</rows> attribute is not specified it defaults to 10 rows per page.
4615 When you have a paged resultset, L</count> will only return the number
4616 of rows in the page. To get the total, use the L</pager> and call
4617 C<total_entries> on it.
4627 Specifies the maximum number of rows for direct retrieval or the number of
4628 rows per page if the page attribute or method is used.
4634 =item Value: $offset
4638 Specifies the (zero-based) row number for the first row to be returned, or the
4639 of the first row of the first page if paging is used.
4641 =head2 software_limit
4645 =item Value: (0 | 1)
4649 When combined with L</rows> and/or L</offset> the generated SQL will not
4650 include any limit dialect stanzas. Instead the entire result will be selected
4651 as if no limits were specified, and DBIC will perform the limit locally, by
4652 artificially advancing and finishing the resulting L</cursor>.
4654 This is the recommended way of performing resultset limiting when no sane RDBMS
4655 implementation is available (e.g.
4656 L<Sybase ASE|DBIx::Class::Storage::DBI::Sybase::ASE> using the
4657 L<Generic Sub Query|DBIx::Class::SQLMaker::LimitDialects/GenericSubQ> hack)
4663 =item Value: \@columns
4667 A arrayref of columns to group by. Can include columns of joined tables.
4669 group_by => [qw/ column1 column2 ... /]
4675 =item Value: $condition
4679 The HAVING operator specifies a B<secondary> condition applied to the set
4680 after the grouping calculations have been done. In other words it is a
4681 constraint just like L</where> (and accepting the same
4682 L<SQL::Abstract syntax|SQL::Abstract/WHERE CLAUSES>) applied to the data
4683 as it exists after GROUP BY has taken place. Specifying L</having> without
4684 L</group_by> is a logical mistake, and a fatal error on most RDBMS engines.
4688 having => { 'count_employee' => { '>=', 100 } }
4690 or with an in-place function in which case literal SQL is required:
4692 having => \[ 'count(employee) >= ?', 100 ]
4698 =item Value: (0 | 1)
4702 Set to 1 to automatically generate a L</group_by> clause based on the selection
4703 (including intelligent handling of L</order_by> contents). Note that the group
4704 criteria calculation takes place over the B<final> selection. This includes
4705 any L</+columns>, L</+select> or L</order_by> additions in subsequent
4706 L</search> calls, and standalone columns selected via
4707 L<DBIx::Class::ResultSetColumn> (L</get_column>). A notable exception are the
4708 extra selections specified via L</prefetch> - such selections are explicitly
4709 excluded from group criteria calculations.
4711 If the final ResultSet also explicitly defines a L</group_by> attribute, this
4712 setting is ignored and an appropriate warning is issued.
4716 Adds extra conditions to the resultset, combined with the preexisting C<WHERE>
4717 conditions, same as the B<first> argument to the L<search operator|/search>
4719 # only return rows WHERE deleted IS NULL for all searches
4720 __PACKAGE__->resultset_attributes({ where => { deleted => undef } });
4722 Note that the above example is
4723 L<strongly discouraged|DBIx::Class::ResultSource/resultset_attributes>.
4727 Set to 1 to cache search results. This prevents extra SQL queries if you
4728 revisit rows in your ResultSet:
4730 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
4732 while( my $artist = $resultset->next ) {
4736 $resultset->first; # without cache, this would issue a query
4738 By default, searches are not cached.
4740 For more examples of using these attributes, see
4741 L<DBIx::Class::Manual::Cookbook>.
4747 =item Value: ( 'update' | 'shared' | \$scalar )
4751 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT
4752 ... FOR SHARED. If \$scalar is passed, this is taken directly and embedded in the
4757 DBIx::Class supports arbitrary related data prefetching from multiple related
4758 sources. Any combination of relationship types and column sets are supported.
4759 If L<collapsing|/collapse> is requested, there is an additional requirement of
4760 selecting enough data to make every individual object uniquely identifiable.
4762 Here are some more involved examples, based on the following relationship map:
4765 My::Schema::CD->belongs_to( artist => 'My::Schema::Artist' );
4766 My::Schema::CD->might_have( liner_note => 'My::Schema::LinerNotes' );
4767 My::Schema::CD->has_many( tracks => 'My::Schema::Track' );
4769 My::Schema::Artist->belongs_to( record_label => 'My::Schema::RecordLabel' );
4771 My::Schema::Track->has_many( guests => 'My::Schema::Guest' );
4775 my $rs = $schema->resultset('Tag')->search(
4784 The initial search results in SQL like the following:
4786 SELECT tag.*, cd.*, artist.* FROM tag
4787 JOIN cd ON tag.cd = cd.cdid
4788 JOIN artist ON cd.artist = artist.artistid
4790 L<DBIx::Class> has no need to go back to the database when we access the
4791 C<cd> or C<artist> relationships, which saves us two SQL statements in this
4794 Simple prefetches will be joined automatically, so there is no need
4795 for a C<join> attribute in the above search.
4797 The L</prefetch> attribute can be used with any of the relationship types
4798 and multiple prefetches can be specified together. Below is a more complex
4799 example that prefetches a CD's artist, its liner notes (if present),
4800 the cover image, the tracks on that CD, and the guests on those
4803 my $rs = $schema->resultset('CD')->search(
4807 { artist => 'record_label'}, # belongs_to => belongs_to
4808 'liner_note', # might_have
4809 'cover_image', # has_one
4810 { tracks => 'guests' }, # has_many => has_many
4815 This will produce SQL like the following:
4817 SELECT cd.*, artist.*, record_label.*, liner_note.*, cover_image.*,
4821 ON artist.artistid = me.artistid
4822 JOIN record_label record_label
4823 ON record_label.labelid = artist.labelid
4824 LEFT JOIN track tracks
4825 ON tracks.cdid = me.cdid
4826 LEFT JOIN guest guests
4827 ON guests.trackid = track.trackid
4828 LEFT JOIN liner_notes liner_note
4829 ON liner_note.cdid = me.cdid
4830 JOIN cd_artwork cover_image
4831 ON cover_image.cdid = me.cdid
4834 Now the C<artist>, C<record_label>, C<liner_note>, C<cover_image>,
4835 C<tracks>, and C<guests> of the CD will all be available through the
4836 relationship accessors without the need for additional queries to the
4841 Prefetch does a lot of deep magic. As such, it may not behave exactly
4842 as you might expect.
4848 Prefetch uses the L</cache> to populate the prefetched relationships. This
4849 may or may not be what you want.
4853 If you specify a condition on a prefetched relationship, ONLY those
4854 rows that match the prefetched condition will be fetched into that relationship.
4855 This means that adding prefetch to a search() B<may alter> what is returned by
4856 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
4858 my $artist_rs = $schema->resultset('Artist')->search({
4864 my $count = $artist_rs->first->cds->count;
4866 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
4868 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
4870 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
4872 That cmp_ok() may or may not pass depending on the datasets involved. In other
4873 words the C<WHERE> condition would apply to the entire dataset, just like
4874 it would in regular SQL. If you want to add a condition only to the "right side"
4875 of a C<LEFT JOIN> - consider declaring and using a L<relationship with a custom
4876 condition|DBIx::Class::Relationship::Base/condition>
4880 =head1 DBIC BIND VALUES
4882 Because DBIC may need more information to bind values than just the column name
4883 and value itself, it uses a special format for both passing and receiving bind
4884 values. Each bind value should be composed of an arrayref of
4885 C<< [ \%args => $val ] >>. The format of C<< \%args >> is currently:
4891 If present (in any form), this is what is being passed directly to bind_param.
4892 Note that different DBD's expect different bind args. (e.g. DBD::SQLite takes
4893 a single numerical type, while DBD::Pg takes a hashref if bind options.)
4895 If this is specified, all other bind options described below are ignored.
4899 If present, this is used to infer the actual bind attribute by passing to
4900 C<< $resolved_storage->bind_attribute_by_data_type() >>. Defaults to the
4901 "data_type" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>.
4903 Note that the data type is somewhat freeform (hence the sqlt_ prefix);
4904 currently drivers are expected to "Do the Right Thing" when given a common
4905 datatype name. (Not ideal, but that's what we got at this point.)
4909 Currently used to correctly allocate buffers for bind_param_inout().
4910 Defaults to "size" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>,
4911 or to a sensible value based on the "data_type".
4915 Used to fill in missing sqlt_datatype and sqlt_size attributes (if they are
4916 explicitly specified they are never overridden). Also used by some weird DBDs,
4917 where the column name should be available at bind_param time (e.g. Oracle).
4921 For backwards compatibility and convenience, the following shortcuts are
4924 [ $name => $val ] === [ { dbic_colname => $name }, $val ]
4925 [ \$dt => $val ] === [ { sqlt_datatype => $dt }, $val ]
4926 [ undef, $val ] === [ {}, $val ]
4927 $val === [ {}, $val ]
4929 =head1 FURTHER QUESTIONS?
4931 Check the list of L<additional DBIC resources|DBIx::Class/GETTING HELP/SUPPORT>.
4933 =head1 COPYRIGHT AND LICENSE
4935 This module is free software L<copyright|DBIx::Class/COPYRIGHT AND LICENSE>
4936 by the L<DBIx::Class (DBIC) authors|DBIx::Class/AUTHORS>. You can
4937 redistribute it and/or modify it under the same terms as the
4938 L<DBIx::Class library|DBIx::Class/COPYRIGHT AND LICENSE>.