1 package DBIx::Class::ResultSet;
5 use base qw/DBIx::Class/;
7 use DBIx::Class::ResultSetColumn;
8 use Scalar::Util qw/blessed weaken reftype/;
9 use DBIx::Class::_Util 'fail_on_internal_wantarray';
11 use Data::Dumper::Concise ();
12 use Data::Query::Constants;
13 use Data::Query::ExprHelpers;
14 # not importing first() as it will clash with our own method
18 # De-duplication in _merge_attr() is disabled, but left in for reference
19 # (the merger is used for other things that ought not to be de-duped)
20 *__HM_DEDUP = sub () { 0 };
30 # this is real - CDBICompat overrides it with insanity
31 # yes, prototype won't matter, but that's for now ;)
34 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class result_source/);
38 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
42 my $users_rs = $schema->resultset('User');
43 while( $user = $users_rs->next) {
44 print $user->username;
47 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
48 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
52 A ResultSet is an object which stores a set of conditions representing
53 a query. It is the backbone of DBIx::Class (i.e. the really
54 important/useful bit).
56 No SQL is executed on the database when a ResultSet is created, it
57 just stores all the conditions needed to create the query.
59 A basic ResultSet representing the data of an entire table is returned
60 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
61 L<Source|DBIx::Class::Manual::Glossary/Source> name.
63 my $users_rs = $schema->resultset('User');
65 A new ResultSet is returned from calling L</search> on an existing
66 ResultSet. The new one will contain all the conditions of the
67 original, plus any new conditions added in the C<search> call.
69 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
70 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
73 The query that the ResultSet represents is B<only> executed against
74 the database when these methods are called:
75 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
77 If a resultset is used in a numeric context it returns the L</count>.
78 However, if it is used in a boolean context it is B<always> true. So if
79 you want to check if a resultset has any results, you must use C<if $rs
82 =head1 CUSTOM ResultSet CLASSES THAT USE Moose
84 If you want to make your custom ResultSet classes with L<Moose>, use a template
87 package MyApp::Schema::ResultSet::User;
90 use namespace::autoclean;
92 extends 'DBIx::Class::ResultSet';
94 sub BUILDARGS { $_[2] }
98 __PACKAGE__->meta->make_immutable;
102 The L<MooseX::NonMoose> is necessary so that the L<Moose> constructor does not
103 clash with the regular ResultSet constructor. Alternatively, you can use:
105 __PACKAGE__->meta->make_immutable(inline_constructor => 0);
107 The L<BUILDARGS|Moose::Manual::Construction/BUILDARGS> is necessary because the
108 signature of the ResultSet C<new> is C<< ->new($source, \%args) >>.
112 =head2 Chaining resultsets
114 Let's say you've got a query that needs to be run to return some data
115 to the user. But, you have an authorization system in place that
116 prevents certain users from seeing certain information. So, you want
117 to construct the basic query in one method, but add constraints to it in
122 my $request = $self->get_request; # Get a request object somehow.
123 my $schema = $self->result_source->schema;
125 my $cd_rs = $schema->resultset('CD')->search({
126 title => $request->param('title'),
127 year => $request->param('year'),
130 $cd_rs = $self->apply_security_policy( $cd_rs );
132 return $cd_rs->all();
135 sub apply_security_policy {
144 =head3 Resolving conditions and attributes
146 When a resultset is chained from another resultset (e.g.:
147 C<< my $new_rs = $old_rs->search(\%extra_cond, \%attrs) >>), conditions
148 and attributes with the same keys need resolving.
150 If any of L</columns>, L</select>, L</as> are present, they reset the
151 original selection, and start the selection "clean".
153 The L</join>, L</prefetch>, L</+columns>, L</+select>, L</+as> attributes
154 are merged into the existing ones from the original resultset.
156 The L</where> and L</having> attributes, and any search conditions, are
157 merged with an SQL C<AND> to the existing condition from the original
160 All other attributes are overridden by any new ones supplied in the
163 =head2 Multiple queries
165 Since a resultset just defines a query, you can do all sorts of
166 things with it with the same object.
168 # Don't hit the DB yet.
169 my $cd_rs = $schema->resultset('CD')->search({
170 title => 'something',
174 # Each of these hits the DB individually.
175 my $count = $cd_rs->count;
176 my $most_recent = $cd_rs->get_column('date_released')->max();
177 my @records = $cd_rs->all;
179 And it's not just limited to SELECT statements.
185 $cd_rs->create({ artist => 'Fred' });
187 Which is the same as:
189 $schema->resultset('CD')->create({
190 title => 'something',
195 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
203 =item Arguments: L<$source|DBIx::Class::ResultSource>, L<\%attrs?|/ATTRIBUTES>
205 =item Return Value: L<$resultset|/search>
209 The resultset constructor. Takes a source object (usually a
210 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
211 L</ATTRIBUTES> below). Does not perform any queries -- these are
212 executed as needed by the other methods.
214 Generally you never construct a resultset manually. Instead you get one
216 C<< $schema->L<resultset|DBIx::Class::Schema/resultset>('$source_name') >>
217 or C<< $another_resultset->L<search|/search>(...) >> (the later called in
220 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
226 If called on an object, proxies to L</new_result> instead, so
228 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
230 will return a CD object, not a ResultSet, and is equivalent to:
232 my $cd = $schema->resultset('CD')->new_result({ title => 'Spoon' });
234 Please also keep in mind that many internals call L</new_result> directly,
235 so overloading this method with the idea of intercepting new result object
236 creation B<will not work>. See also warning pertaining to L</create>.
244 return $class->new_result(@_) if ref $class;
246 my ($source, $attrs) = @_;
247 $source = $source->resolve
248 if $source->isa('DBIx::Class::ResultSourceHandle');
250 $attrs = { %{$attrs||{}} };
251 delete @{$attrs}{qw(_sqlmaker_select_args _related_results_construction)};
253 if ($attrs->{page}) {
254 $attrs->{rows} ||= 10;
257 $attrs->{alias} ||= 'me';
260 result_source => $source,
261 cond => $attrs->{where},
266 # if there is a dark selector, this means we are already in a
267 # chain and the cleanup/sanification was taken care of by
269 $self->_normalize_selection($attrs)
270 unless $attrs->{_dark_selector};
273 $attrs->{result_class} || $source->result_class
283 =item Arguments: L<$cond|DBIx::Class::SQLMaker> | undef, L<\%attrs?|/ATTRIBUTES>
285 =item Return Value: $resultset (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
289 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
290 my $new_rs = $cd_rs->search({ year => 2005 });
292 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
293 # year = 2005 OR year = 2004
295 In list context, C<< ->all() >> is called implicitly on the resultset, thus
296 returning a list of L<result|DBIx::Class::Manual::ResultClass> objects instead.
297 To avoid that, use L</search_rs>.
299 If you need to pass in additional attributes but no additional condition,
300 call it as C<search(undef, \%attrs)>.
302 # "SELECT name, artistid FROM $artist_table"
303 my @all_artists = $schema->resultset('Artist')->search(undef, {
304 columns => [qw/name artistid/],
307 For a list of attributes that can be passed to C<search>, see
308 L</ATTRIBUTES>. For more examples of using this function, see
309 L<Searching|DBIx::Class::Manual::Cookbook/SEARCHING>. For a complete
310 documentation for the first argument, see L<SQL::Abstract/"WHERE CLAUSES">
311 and its extension L<DBIx::Class::SQLMaker>.
313 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
317 Note that L</search> does not process/deflate any of the values passed in the
318 L<SQL::Abstract>-compatible search condition structure. This is unlike other
319 condition-bound methods L</new_result>, L</create> and L</find>. The user must ensure
320 manually that any value passed to this method will stringify to something the
321 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
322 objects, for more info see:
323 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
329 my $rs = $self->search_rs( @_ );
332 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_WANTARRAY and my $sog = fail_on_internal_wantarray($rs);
335 elsif (defined wantarray) {
339 # we can be called by a relationship helper, which in
340 # turn may be called in void context due to some braindead
341 # overload or whatever else the user decided to be clever
342 # at this particular day. Thus limit the exception to
343 # external code calls only
344 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense')
345 if (caller)[0] !~ /^\QDBIx::Class::/;
355 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
357 =item Return Value: L<$resultset|/search>
361 This method does the same exact thing as search() except it will
362 always return a resultset, even in list context.
369 my $rsrc = $self->result_source;
370 my ($call_cond, $call_attrs);
372 # Special-case handling for (undef, undef) or (undef)
373 # Note that (foo => undef) is valid deprecated syntax
374 @_ = () if not scalar grep { defined $_ } @_;
380 # fish out attrs in the ($condref, $attr) case
381 elsif (@_ == 2 and ( ! defined $_[0] or (ref $_[0]) ne '') ) {
382 ($call_cond, $call_attrs) = @_;
385 $self->throw_exception('Odd number of arguments to search')
389 carp_unique 'search( %condition ) is deprecated, use search( \%condition ) instead'
390 unless $rsrc->result_class->isa('DBIx::Class::CDBICompat');
392 for my $i (0 .. $#_) {
394 $self->throw_exception ('All keys in condition key/value pairs must be plain scalars')
395 if (! defined $_[$i] or ref $_[$i] ne '');
401 if (blessed($call_cond) and $call_cond->isa('Data::Query::ExprBuilder')) {
402 $call_cond = \$call_cond->{expr};
405 # see if we can keep the cache (no $rs changes)
407 my %safe = (alias => 1, cache => 1);
408 if ( ! List::Util::first { !$safe{$_} } keys %$call_attrs and (
411 ref $call_cond eq 'HASH' && ! keys %$call_cond
413 ref $call_cond eq 'ARRAY' && ! @$call_cond
415 $cache = $self->get_cache;
418 my $old_attrs = { %{$self->{attrs}} };
419 my ($old_having, $old_where) = delete @{$old_attrs}{qw(having where)};
421 my $new_attrs = { %$old_attrs };
423 # take care of call attrs (only if anything is changing)
424 if ($call_attrs and keys %$call_attrs) {
426 # copy for _normalize_selection
427 $call_attrs = { %$call_attrs };
429 my @selector_attrs = qw/select as columns cols +select +as +columns include_columns/;
431 # reset the current selector list if new selectors are supplied
432 if (List::Util::first { exists $call_attrs->{$_} } qw/columns cols select as/) {
433 delete @{$old_attrs}{(@selector_attrs, '_dark_selector')};
436 # Normalize the new selector list (operates on the passed-in attr structure)
437 # Need to do it on every chain instead of only once on _resolved_attrs, in
438 # order to allow detection of empty vs partial 'as'
439 $call_attrs->{_dark_selector} = $old_attrs->{_dark_selector}
440 if $old_attrs->{_dark_selector};
441 $self->_normalize_selection ($call_attrs);
443 # start with blind overwriting merge, exclude selector attrs
444 $new_attrs = { %{$old_attrs}, %{$call_attrs} };
445 delete @{$new_attrs}{@selector_attrs};
447 for (@selector_attrs) {
448 $new_attrs->{$_} = $self->_merge_attr($old_attrs->{$_}, $call_attrs->{$_})
449 if ( exists $old_attrs->{$_} or exists $call_attrs->{$_} );
452 # older deprecated name, use only if {columns} is not there
453 if (my $c = delete $new_attrs->{cols}) {
454 carp_unique( "Resultset attribute 'cols' is deprecated, use 'columns' instead" );
455 if ($new_attrs->{columns}) {
456 carp "Resultset specifies both the 'columns' and the legacy 'cols' attributes - ignoring 'cols'";
459 $new_attrs->{columns} = $c;
464 # join/prefetch use their own crazy merging heuristics
465 foreach my $key (qw/join prefetch/) {
466 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key})
467 if exists $call_attrs->{$key};
470 # stack binds together
471 $new_attrs->{bind} = [ @{ $old_attrs->{bind} || [] }, @{ $call_attrs->{bind} || [] } ];
475 for ($old_where, $call_cond) {
477 $new_attrs->{where} = $self->_stack_cond (
478 $_, $new_attrs->{where}
483 if (defined $old_having) {
484 $new_attrs->{having} = $self->_stack_cond (
485 $old_having, $new_attrs->{having}
489 my $rs = (ref $self)->new($rsrc, $new_attrs);
491 $rs->set_cache($cache) if ($cache);
497 sub _normalize_selection {
498 my ($self, $attrs) = @_;
501 if ( exists $attrs->{include_columns} ) {
502 carp_unique( "Resultset attribute 'include_columns' is deprecated, use '+columns' instead" );
503 $attrs->{'+columns'} = $self->_merge_attr(
504 $attrs->{'+columns'}, delete $attrs->{include_columns}
508 # columns are always placed first, however
510 # Keep the X vs +X separation until _resolved_attrs time - this allows to
511 # delay the decision on whether to use a default select list ($rsrc->columns)
512 # allowing stuff like the remove_columns helper to work
514 # select/as +select/+as pairs need special handling - the amount of select/as
515 # elements in each pair does *not* have to be equal (think multicolumn
516 # selectors like distinct(foo, bar) ). If the selector is bare (no 'as'
517 # supplied at all) - try to infer the alias, either from the -as parameter
518 # of the selector spec, or use the parameter whole if it looks like a column
519 # name (ugly legacy heuristic). If all fails - leave the selector bare (which
520 # is ok as well), but make sure no more additions to the 'as' chain take place
521 for my $pref ('', '+') {
523 my ($sel, $as) = map {
524 my $key = "${pref}${_}";
526 my $val = [ ref $attrs->{$key} eq 'ARRAY'
528 : $attrs->{$key} || ()
530 delete $attrs->{$key};
534 if (! @$as and ! @$sel ) {
537 elsif (@$as and ! @$sel) {
538 $self->throw_exception(
539 "Unable to handle ${pref}as specification (@$as) without a corresponding ${pref}select"
543 # no as part supplied at all - try to deduce (unless explicit end of named selection is declared)
544 # if any @$as has been supplied we assume the user knows what (s)he is doing
545 # and blindly keep stacking up pieces
546 unless ($attrs->{_dark_selector}) {
549 if ( ref $_ eq 'HASH' and exists $_->{-as} ) {
550 push @$as, $_->{-as};
552 # assume any plain no-space, no-parenthesis string to be a column spec
553 # FIXME - this is retarded but is necessary to support shit like 'count(foo)'
554 elsif ( ! ref $_ and $_ =~ /^ [^\s\(\)]+ $/x) {
557 # if all else fails - raise a flag that no more aliasing will be allowed
559 $attrs->{_dark_selector} = {
561 string => ($dark_sel_dumper ||= do {
562 require Data::Dumper::Concise;
563 Data::Dumper::Concise::DumperObject()->Indent(0);
564 })->Values([$_])->Dump
572 elsif (@$as < @$sel) {
573 $self->throw_exception(
574 "Unable to handle an ${pref}as specification (@$as) with less elements than the corresponding ${pref}select"
577 elsif ($pref and $attrs->{_dark_selector}) {
578 $self->throw_exception(
579 "Unable to process named '+select', resultset contains an unnamed selector $attrs->{_dark_selector}{string}"
585 $attrs->{"${pref}select"} = $self->_merge_attr($attrs->{"${pref}select"}, $sel);
586 $attrs->{"${pref}as"} = $self->_merge_attr($attrs->{"${pref}as"}, $as);
591 my ($self, $left, $right) = @_;
593 my $source = $self->result_source;
595 my $converter = $source->schema->storage->sql_maker->converter;
597 my @top = map $source->_extract_top_level_conditions(
598 $converter->_expr_to_dq($_)
599 ), grep defined, $left, $right;
601 return undef unless @top;
605 my @uniq = grep { !$seen{Data::Dumper::Concise::Dumper($_)}++ } @top;
607 return \$uniq[0] if @uniq == 1;
609 return \Operator({ 'SQL.Naive' => 'AND' }, \@uniq);
612 =head2 search_literal
614 B<CAVEAT>: C<search_literal> is provided for Class::DBI compatibility and
615 should only be used in that context. C<search_literal> is a convenience
616 method. It is equivalent to calling C<< $schema->search(\[]) >>, but if you
617 want to ensure columns are bound correctly, use L</search>.
619 See L<DBIx::Class::Manual::Cookbook/SEARCHING> and
620 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
621 require C<search_literal>.
625 =item Arguments: $sql_fragment, @standalone_bind_values
627 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
631 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
632 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
634 Pass a literal chunk of SQL to be added to the conditional part of the
637 Example of how to use C<search> instead of C<search_literal>
639 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
640 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
645 my ($self, $sql, @bind) = @_;
647 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
650 return $self->search(\[ $sql, map [ {} => $_ ], @bind ], ($attr || () ));
657 =item Arguments: \%columns_values | @pk_values, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
659 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
663 Finds and returns a single row based on supplied criteria. Takes either a
664 hashref with the same format as L</create> (including inference of foreign
665 keys from related objects), or a list of primary key values in the same
666 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
667 declaration on the L</result_source>.
669 In either case an attempt is made to combine conditions already existing on
670 the resultset with the condition passed to this method.
672 To aid with preparing the correct query for the storage you may supply the
673 C<key> attribute, which is the name of a
674 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
675 unique constraint corresponding to the
676 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
677 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
678 to construct a query that satisfies the named unique constraint fully (
679 non-NULL values for each column member of the constraint) an exception is
682 If no C<key> is specified, the search is carried over all unique constraints
683 which are fully defined by the available condition.
685 If no such constraint is found, C<find> currently defaults to a simple
686 C<< search->(\%column_values) >> which may or may not do what you expect.
687 Note that this fallback behavior may be deprecated in further versions. If
688 you need to search with arbitrary conditions - use L</search>. If the query
689 resulting from this fallback produces more than one row, a warning to the
690 effect is issued, though only the first row is constructed and returned as
693 In addition to C<key>, L</find> recognizes and applies standard
694 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
696 Note that if you have extra concerns about the correctness of the resulting
697 query you need to specify the C<key> attribute and supply the entire condition
698 as an argument to find (since it is not always possible to perform the
699 combination of the resultset condition with the supplied one, especially if
700 the resultset condition contains literal sql).
702 For example, to find a row by its primary key:
704 my $cd = $schema->resultset('CD')->find(5);
706 You can also find a row by a specific unique constraint:
708 my $cd = $schema->resultset('CD')->find(
710 artist => 'Massive Attack',
711 title => 'Mezzanine',
713 { key => 'cd_artist_title' }
716 See also L</find_or_create> and L</update_or_create>.
722 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
724 my $rsrc = $self->result_source;
727 if (exists $attrs->{key}) {
728 $constraint_name = defined $attrs->{key}
730 : $self->throw_exception("An undefined 'key' resultset attribute makes no sense")
734 # Parse out the condition from input
737 if (ref $_[0] eq 'HASH') {
738 $call_cond = { %{$_[0]} };
741 # if only values are supplied we need to default to 'primary'
742 $constraint_name = 'primary' unless defined $constraint_name;
744 my @c_cols = $rsrc->unique_constraint_columns($constraint_name);
746 $self->throw_exception(
747 "No constraint columns, maybe a malformed '$constraint_name' constraint?"
750 $self->throw_exception (
751 'find() expects either a column/value hashref, or a list of values '
752 . "corresponding to the columns of the specified unique constraint '$constraint_name'"
753 ) unless @c_cols == @_;
756 @{$call_cond}{@c_cols} = @_;
760 for my $key (keys %$call_cond) {
762 my $keyref = ref($call_cond->{$key})
764 my $relinfo = $rsrc->relationship_info($key)
766 my $val = delete $call_cond->{$key};
768 next if $keyref eq 'ARRAY'; # has_many for multi_create
770 my $rel_q = $rsrc->_resolve_condition(
771 $relinfo->{cond}, $val, $key, $key
773 die "Can't handle complex relationship conditions in find" if ref($rel_q) ne 'HASH';
774 @related{keys %$rel_q} = values %$rel_q;
778 # relationship conditions take precedence (?)
779 @{$call_cond}{keys %related} = values %related;
781 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
783 if (defined $constraint_name) {
784 $final_cond = $self->_qualify_cond_columns (
786 $self->_build_unique_cond (
794 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
795 # This means that we got here after a merger of relationship conditions
796 # in ::Relationship::Base::search_related (the row method), and furthermore
797 # the relationship is of the 'single' type. This means that the condition
798 # provided by the relationship (already attached to $self) is sufficient,
799 # as there can be only one row in the database that would satisfy the
803 # no key was specified - fall down to heuristics mode:
804 # run through all unique queries registered on the resultset, and
805 # 'OR' all qualifying queries together
806 my (@unique_queries, %seen_column_combinations);
807 for my $c_name ($rsrc->unique_constraint_names) {
808 next if $seen_column_combinations{
809 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
812 push @unique_queries, try {
813 $self->_build_unique_cond ($c_name, $call_cond, 'croak_on_nulls')
817 $final_cond = @unique_queries
818 ? [ map { $self->_qualify_cond_columns($_, $alias) } @unique_queries ]
819 : $self->_non_unique_find_fallback ($call_cond, $attrs)
823 # Run the query, passing the result_class since it should propagate for find
824 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
825 if ($rs->_resolved_attrs->{collapse}) {
827 carp "Query returned more than one row" if $rs->next;
835 # This is a stop-gap method as agreed during the discussion on find() cleanup:
836 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
838 # It is invoked when find() is called in legacy-mode with insufficiently-unique
839 # condition. It is provided for overrides until a saner way forward is devised
841 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
842 # the road. Please adjust your tests accordingly to catch this situation early
843 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
845 # The method will not be removed without an adequately complete replacement
846 # for strict-mode enforcement
847 sub _non_unique_find_fallback {
848 my ($self, $cond, $attrs) = @_;
850 return $self->_qualify_cond_columns(
852 exists $attrs->{alias}
854 : $self->{attrs}{alias}
859 sub _qualify_cond_columns {
860 my ($self, $cond, $alias) = @_;
862 my %aliased = %$cond;
863 for (keys %aliased) {
864 $aliased{"$alias.$_"} = delete $aliased{$_}
871 sub _build_unique_cond {
872 my ($self, $constraint_name, $extra_cond, $croak_on_null) = @_;
874 my @c_cols = $self->result_source->unique_constraint_columns($constraint_name);
876 # combination may fail if $self->{cond} is non-trivial
877 my ($final_cond) = try {
878 $self->_merge_with_rscond ($extra_cond)
883 # trim out everything not in $columns
884 $final_cond = { map {
885 exists $final_cond->{$_}
886 ? ( $_ => $final_cond->{$_} )
890 if (my @missing = grep
891 { ! ($croak_on_null ? defined $final_cond->{$_} : exists $final_cond->{$_}) }
894 $self->throw_exception( sprintf ( "Unable to satisfy requested constraint '%s', no values for column(s): %s",
896 join (', ', map { "'$_'" } @missing),
903 !$ENV{DBIC_NULLABLE_KEY_NOWARN}
905 my @undefs = sort grep { ! defined $final_cond->{$_} } (keys %$final_cond)
907 carp_unique ( sprintf (
908 "NULL/undef values supplied for requested unique constraint '%s' (NULL "
909 . 'values in column(s): %s). This is almost certainly not what you wanted, '
910 . 'though you can set DBIC_NULLABLE_KEY_NOWARN to disable this warning.',
912 join (', ', map { "'$_'" } @undefs),
919 =head2 search_related
923 =item Arguments: $rel_name, $cond?, L<\%attrs?|/ATTRIBUTES>
925 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
929 $new_rs = $cd_rs->search_related('artist', {
933 Searches the specified relationship, optionally specifying a condition and
934 attributes for matching records. See L</ATTRIBUTES> for more information.
936 In list context, C<< ->all() >> is called implicitly on the resultset, thus
937 returning a list of result objects instead. To avoid that, use L</search_related_rs>.
939 See also L</search_related_rs>.
944 return shift->related_resultset(shift)->search(@_);
947 =head2 search_related_rs
949 This method works exactly the same as search_related, except that
950 it guarantees a resultset, even in list context.
954 sub search_related_rs {
955 return shift->related_resultset(shift)->search_rs(@_);
962 =item Arguments: none
964 =item Return Value: L<$cursor|DBIx::Class::Cursor>
968 Returns a storage-driven cursor to the given resultset. See
969 L<DBIx::Class::Cursor> for more information.
976 return $self->{cursor} ||= do {
977 my $attrs = $self->_resolved_attrs;
978 $self->result_source->storage->select(
979 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
988 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
990 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
994 my $cd = $schema->resultset('CD')->single({ year => 2001 });
996 Inflates the first result without creating a cursor if the resultset has
997 any records in it; if not returns C<undef>. Used by L</find> as a lean version
1000 While this method can take an optional search condition (just like L</search>)
1001 being a fast-code-path it does not recognize search attributes. If you need to
1002 add extra joins or similar, call L</search> and then chain-call L</single> on the
1003 L<DBIx::Class::ResultSet> returned.
1009 As of 0.08100, this method enforces the assumption that the preceding
1010 query returns only one row. If more than one row is returned, you will receive
1013 Query returned more than one row
1015 In this case, you should be using L</next> or L</find> instead, or if you really
1016 know what you are doing, use the L</rows> attribute to explicitly limit the size
1019 This method will also throw an exception if it is called on a resultset prefetching
1020 has_many, as such a prefetch implies fetching multiple rows from the database in
1021 order to assemble the resulting object.
1028 my ($self, $where) = @_;
1030 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
1033 my $attrs = { %{$self->_resolved_attrs} };
1035 $self->throw_exception(
1036 'single() can not be used on resultsets collapsing a has_many. Use find( \%cond ) or next() instead'
1037 ) if $attrs->{collapse};
1040 if (defined $attrs->{where}) {
1043 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
1044 $where, delete $attrs->{where} ]
1047 $attrs->{where} = $where;
1051 my $data = [ $self->result_source->storage->select_single(
1052 $attrs->{from}, $attrs->{select},
1053 $attrs->{where}, $attrs
1056 return undef unless @$data;
1057 $self->{_stashed_rows} = [ $data ];
1058 $self->_construct_results->[0];
1064 # Recursively collapse the query, accumulating values for each column.
1066 sub _collapse_query {
1067 my ($self, $query, $collapsed) = @_;
1071 if (ref $query eq 'ARRAY') {
1072 foreach my $subquery (@$query) {
1073 next unless ref $subquery; # -or
1074 $collapsed = $self->_collapse_query($subquery, $collapsed);
1077 elsif (ref $query eq 'HASH') {
1078 if (keys %$query and (keys %$query)[0] eq '-and') {
1079 foreach my $subquery (@{$query->{-and}}) {
1080 $collapsed = $self->_collapse_query($subquery, $collapsed);
1084 foreach my $col (keys %$query) {
1085 my $value = $query->{$col};
1086 $collapsed->{$col}{$value}++;
1098 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1100 =item Return Value: L<$resultsetcolumn|DBIx::Class::ResultSetColumn>
1104 my $max_length = $rs->get_column('length')->max;
1106 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
1111 my ($self, $column) = @_;
1112 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
1120 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1122 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1126 # WHERE title LIKE '%blue%'
1127 $cd_rs = $rs->search_like({ title => '%blue%'});
1129 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
1130 that this is simply a convenience method retained for ex Class::DBI users.
1131 You most likely want to use L</search> with specific operators.
1133 For more information, see L<DBIx::Class::Manual::Cookbook>.
1135 This method is deprecated and will be removed in 0.09. Use L</search()>
1136 instead. An example conversion is:
1138 ->search_like({ foo => 'bar' });
1142 ->search({ foo => { like => 'bar' } });
1149 'search_like() is deprecated and will be removed in DBIC version 0.09.'
1150 .' Instead use ->search({ x => { -like => "y%" } })'
1151 .' (note the outer pair of {}s - they are important!)'
1153 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
1154 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
1155 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
1156 return $class->search($query, { %$attrs });
1163 =item Arguments: $first, $last
1165 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1169 Returns a resultset or object list representing a subset of elements from the
1170 resultset slice is called on. Indexes are from 0, i.e., to get the first
1171 three records, call:
1173 my ($one, $two, $three) = $rs->slice(0, 2);
1178 my ($self, $min, $max) = @_;
1179 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
1180 $attrs->{offset} = $self->{attrs}{offset} || 0;
1181 $attrs->{offset} += $min;
1182 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
1183 return $self->search(undef, $attrs);
1190 =item Arguments: none
1192 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1196 Returns the next element in the resultset (C<undef> is there is none).
1198 Can be used to efficiently iterate over records in the resultset:
1200 my $rs = $schema->resultset('CD')->search;
1201 while (my $cd = $rs->next) {
1205 Note that you need to store the resultset object, and call C<next> on it.
1206 Calling C<< resultset('Table')->next >> repeatedly will always return the
1207 first record from the resultset.
1214 if (my $cache = $self->get_cache) {
1215 $self->{all_cache_position} ||= 0;
1216 return $cache->[$self->{all_cache_position}++];
1219 if ($self->{attrs}{cache}) {
1220 delete $self->{pager};
1221 $self->{all_cache_position} = 1;
1222 return ($self->all)[0];
1225 return shift(@{$self->{_stashed_results}}) if @{ $self->{_stashed_results}||[] };
1227 $self->{_stashed_results} = $self->_construct_results
1230 return shift @{$self->{_stashed_results}};
1233 # Constructs as many results as it can in one pass while respecting
1234 # cursor laziness. Several modes of operation:
1236 # * Always builds everything present in @{$self->{_stashed_rows}}
1237 # * If called with $fetch_all true - pulls everything off the cursor and
1238 # builds all result structures (or objects) in one pass
1239 # * If $self->_resolved_attrs->{collapse} is true, checks the order_by
1240 # and if the resultset is ordered properly by the left side:
1241 # * Fetches stuff off the cursor until the "master object" changes,
1242 # and saves the last extra row (if any) in @{$self->{_stashed_rows}}
1244 # * Just fetches, and collapses/constructs everything as if $fetch_all
1245 # was requested (there is no other way to collapse except for an
1247 # * If no collapse is requested - just get the next row, construct and
1249 sub _construct_results {
1250 my ($self, $fetch_all) = @_;
1252 my $rsrc = $self->result_source;
1253 my $attrs = $self->_resolved_attrs;
1258 ! $attrs->{order_by}
1262 my @pcols = $rsrc->primary_columns
1264 # default order for collapsing unless the user asked for something
1265 $attrs->{order_by} = [ map { join '.', $attrs->{alias}, $_} @pcols ];
1266 $attrs->{_ordered_for_collapse} = 1;
1267 $attrs->{_order_is_artificial} = 1;
1270 # this will be used as both initial raw-row collector AND as a RV of
1271 # _construct_results. Not regrowing the array twice matters a lot...
1272 # a surprising amount actually
1273 my $rows = delete $self->{_stashed_rows};
1275 my $cursor; # we may not need one at all
1277 my $did_fetch_all = $fetch_all;
1280 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1281 $rows = [ ($rows ? @$rows : ()), $self->cursor->all ];
1283 elsif( $attrs->{collapse} ) {
1285 # a cursor will need to be closed over in case of collapse
1286 $cursor = $self->cursor;
1288 $attrs->{_ordered_for_collapse} = (
1294 ->_main_source_order_by_portion_is_stable($rsrc, $attrs->{order_by}, $attrs->{where})
1296 ) unless defined $attrs->{_ordered_for_collapse};
1298 if (! $attrs->{_ordered_for_collapse}) {
1301 # instead of looping over ->next, use ->all in stealth mode
1302 # *without* calling a ->reset afterwards
1303 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1304 if (! $cursor->{_done}) {
1305 $rows = [ ($rows ? @$rows : ()), $cursor->all ];
1306 $cursor->{_done} = 1;
1311 if (! $did_fetch_all and ! @{$rows||[]} ) {
1312 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1313 $cursor ||= $self->cursor;
1314 if (scalar (my @r = $cursor->next) ) {
1319 return undef unless @{$rows||[]};
1321 # sanity check - people are too clever for their own good
1322 if ($attrs->{collapse} and my $aliastypes = $attrs->{_sqlmaker_select_args}[3]{_aliastypes} ) {
1324 my $multiplied_selectors;
1325 for my $sel_alias ( grep { $_ ne $attrs->{alias} } keys %{ $aliastypes->{selecting} } ) {
1327 $aliastypes->{multiplying}{$sel_alias}
1329 $aliastypes->{premultiplied}{$sel_alias}
1331 $multiplied_selectors->{$_} = 1 for values %{$aliastypes->{selecting}{$sel_alias}{-seen_columns}}
1335 for my $i (0 .. $#{$attrs->{as}} ) {
1336 my $sel = $attrs->{select}[$i];
1338 if (ref $sel eq 'SCALAR') {
1341 elsif( ref $sel eq 'REF' and ref $$sel eq 'ARRAY' ) {
1345 $self->throw_exception(
1346 'Result collapse not possible - selection from a has_many source redirected to the main object'
1347 ) if ($multiplied_selectors->{$sel} and $attrs->{as}[$i] !~ /\./);
1351 # hotspot - skip the setter
1352 my $res_class = $self->_result_class;
1354 my $inflator_cref = $self->{_result_inflator}{cref} ||= do {
1355 $res_class->can ('inflate_result')
1356 or $self->throw_exception("Inflator $res_class does not provide an inflate_result() method");
1359 my $infmap = $attrs->{as};
1361 $self->{_result_inflator}{is_core_row} = ( (
1364 ( \&DBIx::Class::Row::inflate_result || die "No ::Row::inflate_result() - can't happen" )
1365 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_core_row};
1367 $self->{_result_inflator}{is_hri} = ( (
1368 ! $self->{_result_inflator}{is_core_row}
1371 require DBIx::Class::ResultClass::HashRefInflator
1373 DBIx::Class::ResultClass::HashRefInflator->can('inflate_result')
1375 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_hri};
1378 if (! $attrs->{_related_results_construction}) {
1379 # construct a much simpler array->hash folder for the one-table cases right here
1380 if ($self->{_result_inflator}{is_hri}) {
1381 for my $r (@$rows) {
1382 $r = { map { $infmap->[$_] => $r->[$_] } 0..$#$infmap };
1385 # FIXME SUBOPTIMAL this is a very very very hot spot
1386 # while rather optimal we can *still* do much better, by
1387 # building a smarter Row::inflate_result(), and
1388 # switch to feeding it data via a much leaner interface
1390 # crude unscientific benchmarking indicated the shortcut eval is not worth it for
1391 # this particular resultset size
1392 elsif (@$rows < 60) {
1393 for my $r (@$rows) {
1394 $r = $inflator_cref->($res_class, $rsrc, { map { $infmap->[$_] => $r->[$_] } (0..$#$infmap) } );
1399 '$_ = $inflator_cref->($res_class, $rsrc, { %s }) for @$rows',
1400 join (', ', map { "\$infmap->[$_] => \$_->[$_]" } 0..$#$infmap )
1406 $self->{_result_inflator}{is_hri} ? 'hri'
1407 : $self->{_result_inflator}{is_core_row} ? 'classic_pruning'
1408 : 'classic_nonpruning'
1411 # $args and $attrs to _mk_row_parser are separated to delineate what is
1412 # core collapser stuff and what is dbic $rs specific
1413 @{$self->{_row_parser}{$parser_type}}{qw(cref nullcheck)} = $rsrc->_mk_row_parser({
1415 inflate_map => $infmap,
1416 collapse => $attrs->{collapse},
1417 premultiplied => $attrs->{_main_source_premultiplied},
1418 hri_style => $self->{_result_inflator}{is_hri},
1419 prune_null_branches => $self->{_result_inflator}{is_hri} || $self->{_result_inflator}{is_core_row},
1420 }, $attrs) unless $self->{_row_parser}{$parser_type}{cref};
1422 # column_info metadata historically hasn't been too reliable.
1423 # We need to start fixing this somehow (the collapse resolver
1424 # can't work without it). Add an explicit check for the *main*
1425 # result, hopefully this will gradually weed out such errors
1427 # FIXME - this is a temporary kludge that reduces performance
1428 # It is however necessary for the time being
1429 my ($unrolled_non_null_cols_to_check, $err);
1431 if (my $check_non_null_cols = $self->{_row_parser}{$parser_type}{nullcheck} ) {
1434 'Collapse aborted due to invalid ResultSource metadata - the following '
1435 . 'selections are declared non-nullable but NULLs were retrieved: '
1439 COL: for my $i (@$check_non_null_cols) {
1440 ! defined $_->[$i] and push @violating_idx, $i and next COL for @$rows;
1443 $self->throw_exception( $err . join (', ', map { "'$infmap->[$_]'" } @violating_idx ) )
1446 $unrolled_non_null_cols_to_check = join (',', @$check_non_null_cols);
1450 ($did_fetch_all or ! $attrs->{collapse}) ? undef
1451 : defined $unrolled_non_null_cols_to_check ? eval sprintf <<'EOS', $unrolled_non_null_cols_to_check
1453 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1454 my @r = $cursor->next or return;
1455 if (my @violating_idx = grep { ! defined $r[$_] } (%s) ) {
1456 $self->throw_exception( $err . join (', ', map { "'$infmap->[$_]'" } @violating_idx ) )
1462 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1463 my @r = $cursor->next or return;
1468 $self->{_row_parser}{$parser_type}{cref}->(
1470 $next_cref ? ( $next_cref, $self->{_stashed_rows} = [] ) : (),
1473 # Special-case multi-object HRI - there is no $inflator_cref pass
1474 unless ($self->{_result_inflator}{is_hri}) {
1475 $_ = $inflator_cref->($res_class, $rsrc, @$_) for @$rows
1479 # The @$rows check seems odd at first - why wouldn't we want to warn
1480 # regardless? The issue is things like find() etc, where the user
1481 # *knows* only one result will come back. In these cases the ->all
1482 # is not a pessimization, but rather something we actually want
1484 'Unable to properly collapse has_many results in iterator mode due '
1485 . 'to order criteria - performed an eager cursor slurp underneath. '
1486 . 'Consider using ->all() instead'
1487 ) if ( ! $fetch_all and @$rows > 1 );
1492 =head2 result_source
1496 =item Arguments: L<$result_source?|DBIx::Class::ResultSource>
1498 =item Return Value: L<$result_source|DBIx::Class::ResultSource>
1502 An accessor for the primary ResultSource object from which this ResultSet
1509 =item Arguments: $result_class?
1511 =item Return Value: $result_class
1515 An accessor for the class to use when creating result objects. Defaults to
1516 C<< result_source->result_class >> - which in most cases is the name of the
1517 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1519 Note that changing the result_class will also remove any components
1520 that were originally loaded in the source class via
1521 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1522 in the original source class will not run.
1527 my ($self, $result_class) = @_;
1528 if ($result_class) {
1530 # don't fire this for an object
1531 $self->ensure_class_loaded($result_class)
1532 unless ref($result_class);
1534 if ($self->get_cache) {
1535 carp_unique('Changing the result_class of a ResultSet instance with cached results is a noop - the cache contents will not be altered');
1537 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1538 elsif ($self->{cursor} && $self->{cursor}{_pos}) {
1539 $self->throw_exception('Changing the result_class of a ResultSet instance with an active cursor is not supported');
1542 $self->_result_class($result_class);
1544 delete $self->{_result_inflator};
1546 $self->_result_class;
1553 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1555 =item Return Value: $count
1559 Performs an SQL C<COUNT> with the same query as the resultset was built
1560 with to find the number of elements. Passing arguments is equivalent to
1561 C<< $rs->search ($cond, \%attrs)->count >>
1567 return $self->search(@_)->count if @_ and defined $_[0];
1568 return scalar @{ $self->get_cache } if $self->get_cache;
1570 my $attrs = { %{ $self->_resolved_attrs } };
1572 # this is a little optimization - it is faster to do the limit
1573 # adjustments in software, instead of a subquery
1574 my ($rows, $offset) = delete @{$attrs}{qw/rows offset/};
1577 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1578 $crs = $self->_count_subq_rs ($attrs);
1581 $crs = $self->_count_rs ($attrs);
1583 my $count = $crs->next;
1585 $count -= $offset if $offset;
1586 $count = $rows if $rows and $rows < $count;
1587 $count = 0 if ($count < 0);
1596 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1598 =item Return Value: L<$count_rs|DBIx::Class::ResultSetColumn>
1602 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1603 This can be very handy for subqueries:
1605 ->search( { amount => $some_rs->count_rs->as_query } )
1607 As with regular resultsets the SQL query will be executed only after
1608 the resultset is accessed via L</next> or L</all>. That would return
1609 the same single value obtainable via L</count>.
1615 return $self->search(@_)->count_rs if @_;
1617 # this may look like a lack of abstraction (count() does about the same)
1618 # but in fact an _rs *must* use a subquery for the limits, as the
1619 # software based limiting can not be ported if this $rs is to be used
1620 # in a subquery itself (i.e. ->as_query)
1621 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1622 return $self->_count_subq_rs($self->{_attrs});
1625 return $self->_count_rs($self->{_attrs});
1630 # returns a ResultSetColumn object tied to the count query
1633 my ($self, $attrs) = @_;
1635 my $rsrc = $self->result_source;
1637 my $tmp_attrs = { %$attrs };
1638 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1639 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1641 # overwrite the selector (supplied by the storage)
1642 $rsrc->resultset_class->new($rsrc, {
1644 select => $rsrc->storage->_count_select ($rsrc, $attrs),
1646 })->get_column ('count');
1650 # same as above but uses a subquery
1652 sub _count_subq_rs {
1653 my ($self, $attrs) = @_;
1655 my $rsrc = $self->result_source;
1657 my $sub_attrs = { %$attrs };
1658 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1659 delete @{$sub_attrs}{qw/collapse columns as select order_by for/};
1661 # if we multi-prefetch we group_by something unique, as this is what we would
1662 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1663 if ( $attrs->{collapse} ) {
1664 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } @{
1665 $rsrc->_identifying_column_set || $self->throw_exception(
1666 'Unable to construct a unique group_by criteria properly collapsing the '
1667 . 'has_many prefetch before count()'
1672 # Calculate subquery selector
1673 if (my $g = $sub_attrs->{group_by}) {
1675 my $sql_maker = $rsrc->storage->sql_maker;
1677 # necessary as the group_by may refer to aliased functions
1679 for my $sel (@{$attrs->{select}}) {
1680 $sel_index->{$sel->{-as}} = $sel
1681 if (ref $sel eq 'HASH' and $sel->{-as});
1684 # anything from the original select mentioned on the group-by needs to make it to the inner selector
1685 # also look for named aggregates referred in the having clause
1686 # having often contains scalarrefs - thus parse it out entirely
1688 if ($attrs->{having}) {
1689 local $sql_maker->{having_bind};
1690 local $sql_maker->{quote_char} = $sql_maker->{quote_char};
1691 local $sql_maker->{name_sep} = $sql_maker->{name_sep};
1692 unless (defined $sql_maker->{quote_char} and length $sql_maker->{quote_char}) {
1693 $sql_maker->{quote_char} = [ "\x00", "\xFF" ];
1694 # if we don't unset it we screw up retarded but unfortunately working
1695 # 'MAX(foo.bar)' => { '>', 3 }
1696 $sql_maker->{name_sep} = '';
1699 $sql_maker->clear_renderer;
1700 $sql_maker->clear_converter;
1702 my ($lquote, $rquote, $sep) = map { quotemeta $_ } ($sql_maker->_quote_chars, $sql_maker->name_sep);
1704 my $having_sql = $sql_maker->_render_sqla(where => $attrs->{having});
1706 $sql_maker->clear_renderer;
1707 $sql_maker->clear_converter;
1711 # search for both a proper quoted qualified string, for a naive unquoted scalarref
1712 # and if all fails for an utterly naive quoted scalar-with-function
1713 while ($having_sql =~ /
1714 (?: $rquote $sep)? $lquote (.+?) $rquote
1716 [\s,] \w+ \. (\w+) [\s,]
1718 [\s,] $lquote (.+?) $rquote [\s,]
1720 my $part = $1 || $2 || $3; # one of them matched if we got here
1721 unless ($seen_having{$part}++) {
1728 my $colpiece = $sel_index->{$_} || $_;
1730 # unqualify join-based group_by's. Arcane but possible query
1731 # also horrible horrible hack to alias a column (not a func.)
1732 # (probably need to introduce SQLA syntax)
1733 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1736 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1738 push @{$sub_attrs->{select}}, $colpiece;
1742 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1743 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1746 return $rsrc->resultset_class
1747 ->new ($rsrc, $sub_attrs)
1749 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1750 ->get_column ('count');
1754 =head2 count_literal
1756 B<CAVEAT>: C<count_literal> is provided for Class::DBI compatibility and
1757 should only be used in that context. See L</search_literal> for further info.
1761 =item Arguments: $sql_fragment, @standalone_bind_values
1763 =item Return Value: $count
1767 Counts the results in a literal query. Equivalent to calling L</search_literal>
1768 with the passed arguments, then L</count>.
1772 sub count_literal { shift->search_literal(@_)->count; }
1778 =item Arguments: none
1780 =item Return Value: L<@result_objs|DBIx::Class::Manual::ResultClass>
1784 Returns all elements in the resultset.
1791 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1794 delete @{$self}{qw/_stashed_rows _stashed_results/};
1796 if (my $c = $self->get_cache) {
1800 $self->cursor->reset;
1802 my $objs = $self->_construct_results('fetch_all') || [];
1804 $self->set_cache($objs) if $self->{attrs}{cache};
1813 =item Arguments: none
1815 =item Return Value: $self
1819 Resets the resultset's cursor, so you can iterate through the elements again.
1820 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1828 delete @{$self}{qw/_stashed_rows _stashed_results/};
1829 $self->{all_cache_position} = 0;
1830 $self->cursor->reset;
1838 =item Arguments: none
1840 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1844 L<Resets|/reset> the resultset (causing a fresh query to storage) and returns
1845 an object for the first result (or C<undef> if the resultset is empty).
1850 return $_[0]->reset->next;
1856 # Determines whether and what type of subquery is required for the $rs operation.
1857 # If grouping is necessary either supplies its own, or verifies the current one
1858 # After all is done delegates to the proper storage method.
1860 sub _rs_update_delete {
1861 my ($self, $op, $values) = @_;
1863 my $rsrc = $self->result_source;
1864 my $storage = $rsrc->schema->storage;
1866 my $attrs = { %{$self->_resolved_attrs} };
1868 my $join_classifications;
1869 my ($existing_group_by) = delete @{$attrs}{qw(group_by _grouped_by_distinct)};
1871 # do we need a subquery for any reason?
1873 defined $existing_group_by
1875 # if {from} is unparseable wrap a subq
1876 ref($attrs->{from}) ne 'ARRAY'
1878 # limits call for a subq
1879 $self->_has_resolved_attr(qw/rows offset/)
1882 # simplify the joinmap, so we can further decide if a subq is necessary
1883 if (!$needs_subq and @{$attrs->{from}} > 1) {
1885 ($attrs->{from}, $join_classifications) =
1886 $storage->_prune_unused_joins ($attrs);
1888 # any non-pruneable non-local restricting joins imply subq
1889 $needs_subq = defined List::Util::first { $_ ne $attrs->{alias} } keys %{ $join_classifications->{restricting} || {} };
1892 # check if the head is composite (by now all joins are thrown out unless $needs_subq)
1894 (ref $attrs->{from}[0]) ne 'HASH'
1896 ref $attrs->{from}[0]{ $attrs->{from}[0]{-alias} }
1900 # do we need anything like a subquery?
1901 if (! $needs_subq) {
1902 # Most databases do not allow aliasing of tables in UPDATE/DELETE. Thus
1903 # a condition containing 'me' or other table prefixes will not work
1904 # at all - so we convert the WHERE to a dq tree now, dequalify all
1905 # identifiers found therein via a scan across the tree, and then use
1906 # \{} style to pass the result onwards for use in the final query
1907 if ($self->{cond}) {
1909 my $converter = $rsrc->storage->sql_maker->converter;
1911 DQ_IDENTIFIER ,=> sub { $_ = [ $_->[-1] ] for $_[0]->{elements} }
1912 }, my $where_dq = $converter->_where_to_dq($self->{cond}));
1918 # we got this far - means it is time to wrap a subquery
1919 my $idcols = $rsrc->_identifying_column_set || $self->throw_exception(
1921 "Unable to perform complex resultset %s() without an identifying set of columns on source '%s'",
1927 # make a new $rs selecting only the PKs (that's all we really need for the subq)
1928 delete $attrs->{$_} for qw/select as collapse/;
1929 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } @$idcols ];
1931 # this will be consumed by the pruner waaaaay down the stack
1932 $attrs->{_force_prune_multiplying_joins} = 1;
1934 my $subrs = (ref $self)->new($rsrc, $attrs);
1936 if (@$idcols == 1) {
1937 $cond = { $idcols->[0] => { -in => \$subrs->_as_select_dq } };
1939 elsif ($storage->_use_multicolumn_in) {
1940 # no syntax for calling this properly yet
1941 # !!! EXPERIMENTAL API !!! WILL CHANGE !!!
1942 my $left = $storage->sql_maker->_render_sqla(select_select => $idcols);
1943 $left =~ s/^SELECT //i;
1944 my $right = $storage->sql_maker
1946 ->_literal_to_dq(${$subrs->as_query});
1948 { 'SQL.Naive' => 'in' },
1949 [ Literal(SQL => "( $left )"), $right ],
1953 # if all else fails - get all primary keys and operate over a ORed set
1954 # wrap in a transaction for consistency
1955 # this is where the group_by/multiplication starts to matter
1959 # we do not need to check pre-multipliers, since if the premulti is there, its
1960 # parent (who is multi) will be there too
1961 keys %{ $join_classifications->{multiplying} || {} }
1963 # make sure if there is a supplied group_by it matches the columns compiled above
1964 # perfectly. Anything else can not be sanely executed on most databases so croak
1965 # right then and there
1966 if ($existing_group_by) {
1967 my @current_group_by = map
1968 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1973 join ("\x00", sort @current_group_by)
1975 join ("\x00", sort @{$attrs->{columns}} )
1977 $self->throw_exception (
1978 "You have just attempted a $op operation on a resultset which does group_by"
1979 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1980 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1981 . ' kind of queries. Please retry the operation with a modified group_by or'
1982 . ' without using one at all.'
1987 $subrs = $subrs->search({}, { group_by => $attrs->{columns} });
1990 $guard = $storage->txn_scope_guard;
1993 for my $row ($subrs->cursor->all) {
1995 { $idcols->[$_] => $row->[$_] }
2002 my $res = $storage->$op (
2004 $op eq 'update' ? $values : (),
2008 $guard->commit if $guard;
2017 =item Arguments: \%values
2019 =item Return Value: $underlying_storage_rv
2023 Sets the specified columns in the resultset to the supplied values in a
2024 single query. Note that this will not run any accessor/set_column/update
2025 triggers, nor will it update any result object instances derived from this
2026 resultset (this includes the contents of the L<resultset cache|/set_cache>
2027 if any). See L</update_all> if you need to execute any on-update
2028 triggers or cascades defined either by you or a
2029 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2031 The return value is a pass through of what the underlying
2032 storage backend returned, and may vary. See L<DBI/execute> for the most
2037 Note that L</update> does not process/deflate any of the values passed in.
2038 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
2039 ensure manually that any value passed to this method will stringify to
2040 something the RDBMS knows how to deal with. A notable example is the
2041 handling of L<DateTime> objects, for more info see:
2042 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
2047 my ($self, $values) = @_;
2048 $self->throw_exception('Values for update must be a hash')
2049 unless ref $values eq 'HASH';
2051 return $self->_rs_update_delete ('update', $values);
2058 =item Arguments: \%values
2060 =item Return Value: 1
2064 Fetches all objects and updates them one at a time via
2065 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
2066 triggers, while L</update> will not.
2071 my ($self, $values) = @_;
2072 $self->throw_exception('Values for update_all must be a hash')
2073 unless ref $values eq 'HASH';
2075 my $guard = $self->result_source->schema->txn_scope_guard;
2076 $_->update({%$values}) for $self->all; # shallow copy - update will mangle it
2085 =item Arguments: none
2087 =item Return Value: $underlying_storage_rv
2091 Deletes the rows matching this resultset in a single query. Note that this
2092 will not run any delete triggers, nor will it alter the
2093 L<in_storage|DBIx::Class::Row/in_storage> status of any result object instances
2094 derived from this resultset (this includes the contents of the
2095 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
2096 execute any on-delete triggers or cascades defined either by you or a
2097 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2099 The return value is a pass through of what the underlying storage backend
2100 returned, and may vary. See L<DBI/execute> for the most common case.
2106 $self->throw_exception('delete does not accept any arguments')
2109 return $self->_rs_update_delete ('delete');
2116 =item Arguments: none
2118 =item Return Value: 1
2122 Fetches all objects and deletes them one at a time via
2123 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
2124 triggers, while L</delete> will not.
2130 $self->throw_exception('delete_all does not accept any arguments')
2133 my $guard = $self->result_source->schema->txn_scope_guard;
2134 $_->delete for $self->all;
2143 =item Arguments: [ \@column_list, \@row_values+ ] | [ \%col_data+ ]
2145 =item Return Value: L<\@result_objects|DBIx::Class::Manual::ResultClass> (scalar context) | L<@result_objects|DBIx::Class::Manual::ResultClass> (list context)
2149 Accepts either an arrayref of hashrefs or alternatively an arrayref of
2156 The context of this method call has an important effect on what is
2157 submitted to storage. In void context data is fed directly to fastpath
2158 insertion routines provided by the underlying storage (most often
2159 L<DBI/execute_for_fetch>), bypassing the L<new|DBIx::Class::Row/new> and
2160 L<insert|DBIx::Class::Row/insert> calls on the
2161 L<Result|DBIx::Class::Manual::ResultClass> class, including any
2162 augmentation of these methods provided by components. For example if you
2163 are using something like L<DBIx::Class::UUIDColumns> to create primary
2164 keys for you, you will find that your PKs are empty. In this case you
2165 will have to explicitly force scalar or list context in order to create
2170 In non-void (scalar or list) context, this method is simply a wrapper
2171 for L</create>. Depending on list or scalar context either a list of
2172 L<Result|DBIx::Class::Manual::ResultClass> objects or an arrayref
2173 containing these objects is returned.
2175 When supplying data in "arrayref of arrayrefs" invocation style, the
2176 first element should be a list of column names and each subsequent
2177 element should be a data value in the earlier specified column order.
2180 $schema->resultset("Artist")->populate([
2181 [ qw( artistid name ) ],
2182 [ 100, 'A Formally Unknown Singer' ],
2183 [ 101, 'A singer that jumped the shark two albums ago' ],
2184 [ 102, 'An actually cool singer' ],
2187 For the arrayref of hashrefs style each hashref should be a structure
2188 suitable for passing to L</create>. Multi-create is also permitted with
2191 $schema->resultset("Artist")->populate([
2192 { artistid => 4, name => 'Manufactured Crap', cds => [
2193 { title => 'My First CD', year => 2006 },
2194 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2197 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
2198 { title => 'My parents sold me to a record company', year => 2005 },
2199 { title => 'Why Am I So Ugly?', year => 2006 },
2200 { title => 'I Got Surgery and am now Popular', year => 2007 }
2205 If you attempt a void-context multi-create as in the example above (each
2206 Artist also has the related list of CDs), and B<do not> supply the
2207 necessary autoinc foreign key information, this method will proxy to the
2208 less efficient L</create>, and then throw the Result objects away. In this
2209 case there are obviously no benefits to using this method over L</create>.
2216 # cruft placed in standalone method
2217 my $data = $self->_normalize_populate_args(@_);
2219 return unless @$data;
2221 if(defined wantarray) {
2222 my @created = map { $self->create($_) } @$data;
2223 return wantarray ? @created : \@created;
2226 my $first = $data->[0];
2228 # if a column is a registered relationship, and is a non-blessed hash/array, consider
2229 # it relationship data
2230 my (@rels, @columns);
2231 my $rsrc = $self->result_source;
2232 my $rels = { map { $_ => $rsrc->relationship_info($_) } $rsrc->relationships };
2233 for (keys %$first) {
2234 my $ref = ref $first->{$_};
2235 $rels->{$_} && ($ref eq 'ARRAY' or $ref eq 'HASH')
2241 my @pks = $rsrc->primary_columns;
2243 ## do the belongs_to relationships
2244 foreach my $index (0..$#$data) {
2246 # delegate to create() for any dataset without primary keys with specified relationships
2247 if (grep { !defined $data->[$index]->{$_} } @pks ) {
2249 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
2250 my @ret = $self->populate($data);
2256 foreach my $rel (@rels) {
2257 next unless ref $data->[$index]->{$rel} eq "HASH";
2258 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
2259 my ($reverse_relname, $reverse_relinfo) = %{$rsrc->reverse_relationship_info($rel)};
2260 my $related = $result->result_source->_resolve_condition(
2261 $reverse_relinfo->{cond},
2267 delete $data->[$index]->{$rel};
2268 $data->[$index] = {%{$data->[$index]}, %$related};
2270 push @columns, keys %$related if $index == 0;
2274 ## inherit the data locked in the conditions of the resultset
2275 my ($rs_data) = $self->_merge_with_rscond({});
2276 delete @{$rs_data}{@columns};
2278 ## do bulk insert on current row
2279 $rsrc->storage->insert_bulk(
2281 [@columns, keys %$rs_data],
2282 [ map { [ @$_{@columns}, values %$rs_data ] } @$data ],
2285 ## do the has_many relationships
2286 foreach my $item (@$data) {
2290 foreach my $rel (@rels) {
2291 next unless ref $item->{$rel} eq "ARRAY" && @{ $item->{$rel} };
2293 $main_row ||= $self->new_result({map { $_ => $item->{$_} } @pks});
2295 my $child = $main_row->$rel;
2297 my $related = $child->result_source->_resolve_condition(
2298 $rels->{$rel}{cond},
2304 if (ref($related) eq 'REF' and ref($$related) eq 'HASH') {
2305 $related = $self->result_source
2306 ->_extract_fixed_values_for($$related, $rel);
2309 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
2310 my @populate = map { {%$_, %$related} } @rows_to_add;
2312 $child->populate( \@populate );
2318 # populate() arguments went over several incarnations
2319 # What we ultimately support is AoH
2320 sub _normalize_populate_args {
2321 my ($self, $arg) = @_;
2323 if (ref $arg eq 'ARRAY') {
2327 elsif (ref $arg->[0] eq 'HASH') {
2330 elsif (ref $arg->[0] eq 'ARRAY') {
2332 my @colnames = @{$arg->[0]};
2333 foreach my $values (@{$arg}[1 .. $#$arg]) {
2334 push @ret, { map { $colnames[$_] => $values->[$_] } (0 .. $#colnames) };
2340 $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
2347 =item Arguments: none
2349 =item Return Value: L<$pager|Data::Page>
2353 Returns a L<Data::Page> object for the current resultset. Only makes
2354 sense for queries with a C<page> attribute.
2356 To get the full count of entries for a paged resultset, call
2357 C<total_entries> on the L<Data::Page> object.
2364 return $self->{pager} if $self->{pager};
2366 my $attrs = $self->{attrs};
2367 if (!defined $attrs->{page}) {
2368 $self->throw_exception("Can't create pager for non-paged rs");
2370 elsif ($attrs->{page} <= 0) {
2371 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
2373 $attrs->{rows} ||= 10;
2375 # throw away the paging flags and re-run the count (possibly
2376 # with a subselect) to get the real total count
2377 my $count_attrs = { %$attrs };
2378 delete @{$count_attrs}{qw/rows offset page pager/};
2380 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
2382 require DBIx::Class::ResultSet::Pager;
2383 return $self->{pager} = DBIx::Class::ResultSet::Pager->new(
2384 sub { $total_rs->count }, #lazy-get the total
2386 $self->{attrs}{page},
2394 =item Arguments: $page_number
2396 =item Return Value: L<$resultset|/search>
2400 Returns a resultset for the $page_number page of the resultset on which page
2401 is called, where each page contains a number of rows equal to the 'rows'
2402 attribute set on the resultset (10 by default).
2407 my ($self, $page) = @_;
2408 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2415 =item Arguments: \%col_data
2417 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2421 Creates a new result object in the resultset's result class and returns
2422 it. The row is not inserted into the database at this point, call
2423 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2424 will tell you whether the result object has been inserted or not.
2426 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2431 my ($self, $values) = @_;
2433 $self->throw_exception( "new_result takes only one argument - a hashref of values" )
2436 $self->throw_exception( "new_result expects a hashref" )
2437 unless (ref $values eq 'HASH');
2439 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2441 my $new = $self->result_class->new({
2443 ( @$cols_from_relations
2444 ? (-cols_from_relations => $cols_from_relations)
2447 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2451 reftype($new) eq 'HASH'
2457 carp_unique (sprintf (
2458 "%s->new returned a blessed empty hashref - a strong indicator something is wrong with its inheritance chain",
2459 $self->result_class,
2466 # _merge_with_rscond
2468 # Takes a simple hash of K/V data and returns its copy merged with the
2469 # condition already present on the resultset. Additionally returns an
2470 # arrayref of value/condition names, which were inferred from related
2471 # objects (this is needed for in-memory related objects)
2472 sub _merge_with_rscond {
2473 my ($self, $data) = @_;
2475 my (%new_data, @cols_from_relations);
2477 my $alias = $self->{attrs}{alias};
2479 if (! defined $self->{cond}) {
2480 # just massage $data below
2482 elsif (ref $self->{cond} eq 'HASH') {
2483 # precedence must be given to passed values over values inherited from
2484 # the cond, so the order here is important.
2485 my $collapsed_cond = $self->_collapse_cond($self->{cond});
2486 my %implied = %{$self->_remove_alias($collapsed_cond, $alias)};
2488 while ( my($col, $value) = each %implied ) {
2489 my $vref = ref $value;
2495 (keys %$value)[0] eq '='
2497 $new_data{$col} = $value->{'='};
2499 elsif( !$vref or $vref eq 'SCALAR' or blessed($value) ) {
2500 $new_data{$col} = $value;
2504 elsif (ref $self->{cond} eq 'REF' and ref ${$self->{cond}} eq 'HASH') {
2505 if ((${$self->{cond}})->{'DBIx::Class::ResultSource.UNRESOLVABLE'}) {
2506 %new_data = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
2507 @cols_from_relations = keys %new_data;
2509 %new_data = %{$self->_remove_alias(
2510 $self->result_source
2511 ->_extract_fixed_values_for(${$self->{cond}}),
2517 $self->throw_exception(
2518 "Can't abstract implicit construct, resultset condition not a hash"
2524 %{ $self->_remove_alias($data, $alias) },
2527 return (\%new_data, \@cols_from_relations);
2530 # _has_resolved_attr
2532 # determines if the resultset defines at least one
2533 # of the attributes supplied
2535 # used to determine if a subquery is necessary
2537 # supports some virtual attributes:
2539 # This will scan for any joins being present on the resultset.
2540 # It is not a mere key-search but a deep inspection of {from}
2543 sub _has_resolved_attr {
2544 my ($self, @attr_names) = @_;
2546 my $attrs = $self->_resolved_attrs;
2550 for my $n (@attr_names) {
2551 if (grep { $n eq $_ } (qw/-join/) ) {
2552 $extra_checks{$n}++;
2556 my $attr = $attrs->{$n};
2558 next if not defined $attr;
2560 if (ref $attr eq 'HASH') {
2561 return 1 if keys %$attr;
2563 elsif (ref $attr eq 'ARRAY') {
2571 # a resolved join is expressed as a multi-level from
2573 $extra_checks{-join}
2575 ref $attrs->{from} eq 'ARRAY'
2577 @{$attrs->{from}} > 1
2585 # Recursively collapse the condition.
2587 sub _collapse_cond {
2588 my ($self, $cond, $collapsed) = @_;
2592 if (ref $cond eq 'ARRAY') {
2593 foreach my $subcond (@$cond) {
2594 next unless ref $subcond; # -or
2595 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2598 elsif (ref $cond eq 'HASH') {
2599 if (keys %$cond and (keys %$cond)[0] eq '-and') {
2600 foreach my $subcond (@{$cond->{-and}}) {
2601 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2605 foreach my $col (keys %$cond) {
2606 my $value = $cond->{$col};
2607 $collapsed->{$col} = $value;
2617 # Remove the specified alias from the specified query hash. A copy is made so
2618 # the original query is not modified.
2621 my ($self, $query, $alias) = @_;
2623 my %orig = %{ $query || {} };
2626 foreach my $key (keys %orig) {
2628 $unaliased{$key} = $orig{$key};
2631 $unaliased{$1} = $orig{$key}
2632 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2642 =item Arguments: none
2644 =item Return Value: \[ $sql, L<@bind_values|/DBIC BIND VALUES> ]
2648 Returns the SQL query and bind vars associated with the invocant.
2650 This is generally used as the RHS for a subquery.
2657 my $attrs = { %{ $self->_resolved_attrs } };
2659 my $aq = $self->result_source->storage->_select_args_to_query (
2660 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
2663 $self->{_attrs}{_sqlmaker_select_args} = $attrs->{_sqlmaker_select_args};
2670 my $attrs = { %{ $self->_resolved_attrs } };
2671 my $storage = $self->result_source->storage;
2672 my (undef, $ident, @args) = $storage->_select_args(
2673 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
2675 $ident = $ident->from if blessed($ident);
2676 $storage->sql_maker->converter->_select_to_dq(
2685 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2687 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2691 my $artist = $schema->resultset('Artist')->find_or_new(
2692 { artist => 'fred' }, { key => 'artists' });
2694 $cd->cd_to_producer->find_or_new({ producer => $producer },
2695 { key => 'primary' });
2697 Find an existing record from this resultset using L</find>. if none exists,
2698 instantiate a new result object and return it. The object will not be saved
2699 into your storage until you call L<DBIx::Class::Row/insert> on it.
2701 You most likely want this method when looking for existing rows using a unique
2702 constraint that is not the primary key, or looking for related rows.
2704 If you want objects to be saved immediately, use L</find_or_create> instead.
2706 B<Note>: Make sure to read the documentation of L</find> and understand the
2707 significance of the C<key> attribute, as its lack may skew your search, and
2708 subsequently result in spurious new objects.
2710 B<Note>: Take care when using C<find_or_new> with a table having
2711 columns with default values that you intend to be automatically
2712 supplied by the database (e.g. an auto_increment primary key column).
2713 In normal usage, the value of such columns should NOT be included at
2714 all in the call to C<find_or_new>, even when set to C<undef>.
2720 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2721 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2722 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2725 return $self->new_result($hash);
2732 =item Arguments: \%col_data
2734 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2738 Attempt to create a single new row or a row with multiple related rows
2739 in the table represented by the resultset (and related tables). This
2740 will not check for duplicate rows before inserting, use
2741 L</find_or_create> to do that.
2743 To create one row for this resultset, pass a hashref of key/value
2744 pairs representing the columns of the table and the values you wish to
2745 store. If the appropriate relationships are set up, foreign key fields
2746 can also be passed an object representing the foreign row, and the
2747 value will be set to its primary key.
2749 To create related objects, pass a hashref of related-object column values
2750 B<keyed on the relationship name>. If the relationship is of type C<multi>
2751 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2752 The process will correctly identify columns holding foreign keys, and will
2753 transparently populate them from the keys of the corresponding relation.
2754 This can be applied recursively, and will work correctly for a structure
2755 with an arbitrary depth and width, as long as the relationships actually
2756 exists and the correct column data has been supplied.
2758 Instead of hashrefs of plain related data (key/value pairs), you may
2759 also pass new or inserted objects. New objects (not inserted yet, see
2760 L</new_result>), will be inserted into their appropriate tables.
2762 Effectively a shortcut for C<< ->new_result(\%col_data)->insert >>.
2764 Example of creating a new row.
2766 $person_rs->create({
2767 name=>"Some Person",
2768 email=>"somebody@someplace.com"
2771 Example of creating a new row and also creating rows in a related C<has_many>
2772 or C<has_one> resultset. Note Arrayref.
2775 { artistid => 4, name => 'Manufactured Crap', cds => [
2776 { title => 'My First CD', year => 2006 },
2777 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2782 Example of creating a new row and also creating a row in a related
2783 C<belongs_to> resultset. Note Hashref.
2786 title=>"Music for Silly Walks",
2789 name=>"Silly Musician",
2797 When subclassing ResultSet never attempt to override this method. Since
2798 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2799 lot of the internals simply never call it, so your override will be
2800 bypassed more often than not. Override either L<DBIx::Class::Row/new>
2801 or L<DBIx::Class::Row/insert> depending on how early in the
2802 L</create> process you need to intervene. See also warning pertaining to
2810 my ($self, $col_data) = @_;
2811 $self->throw_exception( "create needs a hashref" )
2812 unless ref $col_data eq 'HASH';
2813 return $self->new_result($col_data)->insert;
2816 =head2 find_or_create
2820 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2822 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2826 $cd->cd_to_producer->find_or_create({ producer => $producer },
2827 { key => 'primary' });
2829 Tries to find a record based on its primary key or unique constraints; if none
2830 is found, creates one and returns that instead.
2832 my $cd = $schema->resultset('CD')->find_or_create({
2834 artist => 'Massive Attack',
2835 title => 'Mezzanine',
2839 Also takes an optional C<key> attribute, to search by a specific key or unique
2840 constraint. For example:
2842 my $cd = $schema->resultset('CD')->find_or_create(
2844 artist => 'Massive Attack',
2845 title => 'Mezzanine',
2847 { key => 'cd_artist_title' }
2850 B<Note>: Make sure to read the documentation of L</find> and understand the
2851 significance of the C<key> attribute, as its lack may skew your search, and
2852 subsequently result in spurious row creation.
2854 B<Note>: Because find_or_create() reads from the database and then
2855 possibly inserts based on the result, this method is subject to a race
2856 condition. Another process could create a record in the table after
2857 the find has completed and before the create has started. To avoid
2858 this problem, use find_or_create() inside a transaction.
2860 B<Note>: Take care when using C<find_or_create> with a table having
2861 columns with default values that you intend to be automatically
2862 supplied by the database (e.g. an auto_increment primary key column).
2863 In normal usage, the value of such columns should NOT be included at
2864 all in the call to C<find_or_create>, even when set to C<undef>.
2866 See also L</find> and L</update_or_create>. For information on how to declare
2867 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2869 If you need to know if an existing row was found or a new one created use
2870 L</find_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2871 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2874 my $cd = $schema->resultset('CD')->find_or_new({
2876 artist => 'Massive Attack',
2877 title => 'Mezzanine',
2881 if( !$cd->in_storage ) {
2888 sub find_or_create {
2890 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2891 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2892 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2895 return $self->create($hash);
2898 =head2 update_or_create
2902 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2904 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2908 $resultset->update_or_create({ col => $val, ... });
2910 Like L</find_or_create>, but if a row is found it is immediately updated via
2911 C<< $found_row->update (\%col_data) >>.
2914 Takes an optional C<key> attribute to search on a specific unique constraint.
2917 # In your application
2918 my $cd = $schema->resultset('CD')->update_or_create(
2920 artist => 'Massive Attack',
2921 title => 'Mezzanine',
2924 { key => 'cd_artist_title' }
2927 $cd->cd_to_producer->update_or_create({
2928 producer => $producer,
2934 B<Note>: Make sure to read the documentation of L</find> and understand the
2935 significance of the C<key> attribute, as its lack may skew your search, and
2936 subsequently result in spurious row creation.
2938 B<Note>: Take care when using C<update_or_create> with a table having
2939 columns with default values that you intend to be automatically
2940 supplied by the database (e.g. an auto_increment primary key column).
2941 In normal usage, the value of such columns should NOT be included at
2942 all in the call to C<update_or_create>, even when set to C<undef>.
2944 See also L</find> and L</find_or_create>. For information on how to declare
2945 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2947 If you need to know if an existing row was updated or a new one created use
2948 L</update_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2949 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2954 sub update_or_create {
2956 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2957 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2959 my $row = $self->find($cond, $attrs);
2961 $row->update($cond);
2965 return $self->create($cond);
2968 =head2 update_or_new
2972 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2974 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2978 $resultset->update_or_new({ col => $val, ... });
2980 Like L</find_or_new> but if a row is found it is immediately updated via
2981 C<< $found_row->update (\%col_data) >>.
2985 # In your application
2986 my $cd = $schema->resultset('CD')->update_or_new(
2988 artist => 'Massive Attack',
2989 title => 'Mezzanine',
2992 { key => 'cd_artist_title' }
2995 if ($cd->in_storage) {
2996 # the cd was updated
2999 # the cd is not yet in the database, let's insert it
3003 B<Note>: Make sure to read the documentation of L</find> and understand the
3004 significance of the C<key> attribute, as its lack may skew your search, and
3005 subsequently result in spurious new objects.
3007 B<Note>: Take care when using C<update_or_new> with a table having
3008 columns with default values that you intend to be automatically
3009 supplied by the database (e.g. an auto_increment primary key column).
3010 In normal usage, the value of such columns should NOT be included at
3011 all in the call to C<update_or_new>, even when set to C<undef>.
3013 See also L</find>, L</find_or_create> and L</find_or_new>.
3019 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
3020 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
3022 my $row = $self->find( $cond, $attrs );
3023 if ( defined $row ) {
3024 $row->update($cond);
3028 return $self->new_result($cond);
3035 =item Arguments: none
3037 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass> | undef
3041 Gets the contents of the cache for the resultset, if the cache is set.
3043 The cache is populated either by using the L</prefetch> attribute to
3044 L</search> or by calling L</set_cache>.
3056 =item Arguments: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3058 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass>
3062 Sets the contents of the cache for the resultset. Expects an arrayref
3063 of objects of the same class as those produced by the resultset. Note that
3064 if the cache is set, the resultset will return the cached objects rather
3065 than re-querying the database even if the cache attr is not set.
3067 The contents of the cache can also be populated by using the
3068 L</prefetch> attribute to L</search>.
3073 my ( $self, $data ) = @_;
3074 $self->throw_exception("set_cache requires an arrayref")
3075 if defined($data) && (ref $data ne 'ARRAY');
3076 $self->{all_cache} = $data;
3083 =item Arguments: none
3085 =item Return Value: undef
3089 Clears the cache for the resultset.
3094 shift->set_cache(undef);
3101 =item Arguments: none
3103 =item Return Value: true, if the resultset has been paginated
3111 return !!$self->{attrs}{page};
3118 =item Arguments: none
3120 =item Return Value: true, if the resultset has been ordered with C<order_by>.
3128 return scalar $self->result_source->storage->_extract_order_criteria($self->{attrs}{order_by});
3131 =head2 related_resultset
3135 =item Arguments: $rel_name
3137 =item Return Value: L<$resultset|/search>
3141 Returns a related resultset for the supplied relationship name.
3143 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
3147 sub related_resultset {
3148 my ($self, $rel) = @_;
3150 return $self->{related_resultsets}{$rel}
3151 if defined $self->{related_resultsets}{$rel};
3153 return $self->{related_resultsets}{$rel} = do {
3154 my $rsrc = $self->result_source;
3155 my $rel_info = $rsrc->relationship_info($rel);
3157 $self->throw_exception(
3158 "search_related: result source '" . $rsrc->source_name .
3159 "' has no such relationship $rel")
3162 my $attrs = $self->_chain_relationship($rel);
3164 my $join_count = $attrs->{seen_join}{$rel};
3166 my $alias = $self->result_source->storage
3167 ->relname_to_table_alias($rel, $join_count);
3169 # since this is search_related, and we already slid the select window inwards
3170 # (the select/as attrs were deleted in the beginning), we need to flip all
3171 # left joins to inner, so we get the expected results
3172 # read the comment on top of the actual function to see what this does
3173 $attrs->{from} = $rsrc->schema->storage->_inner_join_to_node ($attrs->{from}, $alias);
3176 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
3177 delete @{$attrs}{qw(result_class alias)};
3179 my $rel_source = $rsrc->related_source($rel);
3183 # The reason we do this now instead of passing the alias to the
3184 # search_rs below is that if you wrap/overload resultset on the
3185 # source you need to know what alias it's -going- to have for things
3186 # to work sanely (e.g. RestrictWithObject wants to be able to add
3187 # extra query restrictions, and these may need to be $alias.)
3189 my $rel_attrs = $rel_source->resultset_attributes;
3190 local $rel_attrs->{alias} = $alias;
3192 $rel_source->resultset
3196 where => $attrs->{where},
3200 if (my $cache = $self->get_cache) {
3201 my @related_cache = map
3202 { @{$_->related_resultset($rel)->get_cache||[]} }
3206 $new->set_cache(\@related_cache) if @related_cache;
3213 =head2 current_source_alias
3217 =item Arguments: none
3219 =item Return Value: $source_alias
3223 Returns the current table alias for the result source this resultset is built
3224 on, that will be used in the SQL query. Usually it is C<me>.
3226 Currently the source alias that refers to the result set returned by a
3227 L</search>/L</find> family method depends on how you got to the resultset: it's
3228 C<me> by default, but eg. L</search_related> aliases it to the related result
3229 source name (and keeps C<me> referring to the original result set). The long
3230 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
3231 (and make this method unnecessary).
3233 Thus it's currently necessary to use this method in predefined queries (see
3234 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
3235 source alias of the current result set:
3237 # in a result set class
3239 my ($self, $user) = @_;
3241 my $me = $self->current_source_alias;
3243 return $self->search({
3244 "$me.modified" => $user->id,
3250 sub current_source_alias {
3251 return (shift->{attrs} || {})->{alias} || 'me';
3254 =head2 as_subselect_rs
3258 =item Arguments: none
3260 =item Return Value: L<$resultset|/search>
3264 Act as a barrier to SQL symbols. The resultset provided will be made into a
3265 "virtual view" by including it as a subquery within the from clause. From this
3266 point on, any joined tables are inaccessible to ->search on the resultset (as if
3267 it were simply where-filtered without joins). For example:
3269 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
3271 # 'x' now pollutes the query namespace
3273 # So the following works as expected
3274 my $ok_rs = $rs->search({'x.other' => 1});
3276 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
3277 # def) we look for one row with contradictory terms and join in another table
3278 # (aliased 'x_2') which we never use
3279 my $broken_rs = $rs->search({'x.name' => 'def'});
3281 my $rs2 = $rs->as_subselect_rs;
3283 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
3284 my $not_joined_rs = $rs2->search({'x.other' => 1});
3286 # works as expected: finds a 'table' row related to two x rows (abc and def)
3287 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
3289 Another example of when one might use this would be to select a subset of
3290 columns in a group by clause:
3292 my $rs = $schema->resultset('Bar')->search(undef, {
3293 group_by => [qw{ id foo_id baz_id }],
3294 })->as_subselect_rs->search(undef, {
3295 columns => [qw{ id foo_id }]
3298 In the above example normally columns would have to be equal to the group by,
3299 but because we isolated the group by into a subselect the above works.
3303 sub as_subselect_rs {
3306 my $attrs = $self->_resolved_attrs;
3308 my $fresh_rs = (ref $self)->new (
3309 $self->result_source
3312 # these pieces will be locked in the subquery
3313 delete $fresh_rs->{cond};
3314 delete @{$fresh_rs->{attrs}}{qw/where bind/};
3316 return $fresh_rs->search( {}, {
3318 $attrs->{alias} => $self->as_query,
3319 -alias => $attrs->{alias},
3320 -rsrc => $self->result_source,
3322 alias => $attrs->{alias},
3326 # This code is called by search_related, and makes sure there
3327 # is clear separation between the joins before, during, and
3328 # after the relationship. This information is needed later
3329 # in order to properly resolve prefetch aliases (any alias
3330 # with a relation_chain_depth less than the depth of the
3331 # current prefetch is not considered)
3333 # The increments happen twice per join. An even number means a
3334 # relationship specified via a search_related, whereas an odd
3335 # number indicates a join/prefetch added via attributes
3337 # Also this code will wrap the current resultset (the one we
3338 # chain to) in a subselect IFF it contains limiting attributes
3339 sub _chain_relationship {
3340 my ($self, $rel) = @_;
3341 my $source = $self->result_source;
3342 my $attrs = { %{$self->{attrs}||{}} };
3344 # we need to take the prefetch the attrs into account before we
3345 # ->_resolve_join as otherwise they get lost - captainL
3346 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
3348 delete @{$attrs}{qw/join prefetch collapse group_by distinct _grouped_by_distinct select as columns +select +as +columns/};
3350 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
3353 my @force_subq_attrs = qw/offset rows group_by having/;
3356 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
3358 $self->_has_resolved_attr (@force_subq_attrs)
3360 # Nuke the prefetch (if any) before the new $rs attrs
3361 # are resolved (prefetch is useless - we are wrapping
3362 # a subquery anyway).
3363 my $rs_copy = $self->search;
3364 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
3365 $rs_copy->{attrs}{join},
3366 delete $rs_copy->{attrs}{prefetch},
3371 -alias => $attrs->{alias},
3372 $attrs->{alias} => $rs_copy->as_query,
3374 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
3375 $seen->{-relation_chain_depth} = 0;
3377 elsif ($attrs->{from}) { #shallow copy suffices
3378 $from = [ @{$attrs->{from}} ];
3383 -alias => $attrs->{alias},
3384 $attrs->{alias} => $source->from,
3388 my $jpath = ($seen->{-relation_chain_depth})
3389 ? $from->[-1][0]{-join_path}
3392 my @requested_joins = $source->_resolve_join(
3399 push @$from, @requested_joins;
3401 $seen->{-relation_chain_depth}++;
3403 # if $self already had a join/prefetch specified on it, the requested
3404 # $rel might very well be already included. What we do in this case
3405 # is effectively a no-op (except that we bump up the chain_depth on
3406 # the join in question so we could tell it *is* the search_related)
3409 # we consider the last one thus reverse
3410 for my $j (reverse @requested_joins) {
3411 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
3412 if ($rel eq $last_j) {
3413 $j->[0]{-relation_chain_depth}++;
3419 unless ($already_joined) {
3420 push @$from, $source->_resolve_join(
3428 $seen->{-relation_chain_depth}++;
3430 return {%$attrs, from => $from, seen_join => $seen};
3433 sub _resolved_attrs {
3435 return $self->{_attrs} if $self->{_attrs};
3437 my $attrs = { %{ $self->{attrs} || {} } };
3438 my $source = $self->result_source;
3439 my $alias = $attrs->{alias};
3441 $self->throw_exception("Specifying distinct => 1 in conjunction with collapse => 1 is unsupported")
3442 if $attrs->{collapse} and $attrs->{distinct};
3444 # default selection list
3445 $attrs->{columns} = [ $source->columns ]
3446 unless List::Util::first { exists $attrs->{$_} } qw/columns cols select as/;
3448 # merge selectors together
3449 for (qw/columns select as/) {
3450 $attrs->{$_} = $self->_merge_attr($attrs->{$_}, delete $attrs->{"+$_"})
3451 if $attrs->{$_} or $attrs->{"+$_"};
3454 # disassemble columns
3456 if (my $cols = delete $attrs->{columns}) {
3457 for my $c (ref $cols eq 'ARRAY' ? @$cols : $cols) {
3458 if (ref $c eq 'HASH') {
3459 for my $as (sort keys %$c) {
3460 push @sel, $c->{$as};
3471 # when trying to weed off duplicates later do not go past this point -
3472 # everything added from here on is unbalanced "anyone's guess" stuff
3473 my $dedup_stop_idx = $#as;
3475 push @as, @{ ref $attrs->{as} eq 'ARRAY' ? $attrs->{as} : [ $attrs->{as} ] }
3477 push @sel, @{ ref $attrs->{select} eq 'ARRAY' ? $attrs->{select} : [ $attrs->{select} ] }
3478 if $attrs->{select};
3480 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3481 $_ = (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" for @sel;
3483 # disqualify all $alias.col as-bits (inflate-map mandated)
3484 $_ = ($_ =~ /^\Q$alias.\E(.+)$/) ? $1 : $_ for @as;
3486 # de-duplicate the result (remove *identical* select/as pairs)
3487 # and also die on duplicate {as} pointing to different {select}s
3488 # not using a c-style for as the condition is prone to shrinkage
3491 while ($i <= $dedup_stop_idx) {
3492 if ($seen->{"$sel[$i] \x00\x00 $as[$i]"}++) {
3497 elsif ($seen->{$as[$i]}++) {
3498 $self->throw_exception(
3499 "inflate_result() alias '$as[$i]' specified twice with different SQL-side {select}-ors"
3507 $attrs->{select} = \@sel;
3508 $attrs->{as} = \@as;
3510 $attrs->{from} ||= [{
3512 -alias => $self->{attrs}{alias},
3513 $self->{attrs}{alias} => $source->from,
3516 if ( $attrs->{join} || $attrs->{prefetch} ) {
3518 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3519 if ref $attrs->{from} ne 'ARRAY';
3521 my $join = (delete $attrs->{join}) || {};
3523 if ( defined $attrs->{prefetch} ) {
3524 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3527 $attrs->{from} = # have to copy here to avoid corrupting the original
3529 @{ $attrs->{from} },
3530 $source->_resolve_join(
3533 { %{ $attrs->{seen_join} || {} } },
3534 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3535 ? $attrs->{from}[-1][0]{-join_path}
3542 if ( defined $attrs->{order_by} ) {
3543 $attrs->{order_by} = (
3544 ref( $attrs->{order_by} ) eq 'ARRAY'
3545 ? [ @{ $attrs->{order_by} } ]
3546 : [ $attrs->{order_by} || () ]
3550 if ($attrs->{group_by} and ref $attrs->{group_by} ne 'ARRAY') {
3551 $attrs->{group_by} = [ $attrs->{group_by} ];
3555 # generate selections based on the prefetch helper
3556 my ($prefetch, @prefetch_select, @prefetch_as);
3557 $prefetch = $self->_merge_joinpref_attr( {}, delete $attrs->{prefetch} )
3558 if defined $attrs->{prefetch};
3562 $self->throw_exception("Unable to prefetch, resultset contains an unnamed selector $attrs->{_dark_selector}{string}")
3563 if $attrs->{_dark_selector};
3565 $self->throw_exception("Specifying prefetch in conjunction with an explicit collapse => 0 is unsupported")
3566 if defined $attrs->{collapse} and ! $attrs->{collapse};
3568 $attrs->{collapse} = 1;
3570 # this is a separate structure (we don't look in {from} directly)
3571 # as the resolver needs to shift things off the lists to work
3572 # properly (identical-prefetches on different branches)
3574 if (ref $attrs->{from} eq 'ARRAY') {
3576 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3578 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3579 next unless $j->[0]{-alias};
3580 next unless $j->[0]{-join_path};
3581 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3583 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3586 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3587 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3591 my @prefetch = $source->_resolve_prefetch( $prefetch, $alias, $join_map );
3593 # save these for after distinct resolution
3594 @prefetch_select = map { $_->[0] } @prefetch;
3595 @prefetch_as = map { $_->[1] } @prefetch;
3598 # run through the resulting joinstructure (starting from our current slot)
3599 # and unset collapse if proven unnecessary
3601 # also while we are at it find out if the current root source has
3602 # been premultiplied by previous related_source chaining
3604 # this allows to predict whether a root object with all other relation
3605 # data set to NULL is in fact unique
3606 if ($attrs->{collapse}) {
3608 if (ref $attrs->{from} eq 'ARRAY') {
3610 if (@{$attrs->{from}} == 1) {
3611 # no joins - no collapse
3612 $attrs->{collapse} = 0;
3615 # find where our table-spec starts
3616 my @fromlist = @{$attrs->{from}};
3618 my $t = shift @fromlist;
3621 # me vs join from-spec distinction - a ref means non-root
3622 if (ref $t eq 'ARRAY') {
3624 $is_multi ||= ! $t->{-is_single};
3626 last if ($t->{-alias} && $t->{-alias} eq $alias);
3627 $attrs->{_main_source_premultiplied} ||= $is_multi;
3630 # no non-singles remaining, nor any premultiplication - nothing to collapse
3632 ! $attrs->{_main_source_premultiplied}
3634 ! List::Util::first { ! $_->[0]{-is_single} } @fromlist
3636 $attrs->{collapse} = 0;
3642 # if we can not analyze the from - err on the side of safety
3643 $attrs->{_main_source_premultiplied} = 1;
3647 # generate the distinct induced group_by before injecting the prefetched select/as parts
3648 if (delete $attrs->{distinct}) {
3649 if ($attrs->{group_by}) {
3650 carp_unique ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3653 $attrs->{_grouped_by_distinct} = 1;
3654 # distinct affects only the main selection part, not what prefetch may add below
3655 ($attrs->{group_by}, my $new_order) = $source->storage->_group_over_selection($attrs);
3657 # FIXME possibly ignore a rewritten order_by (may turn out to be an issue)
3658 # The thinking is: if we are collapsing the subquerying prefetch engine will
3659 # rip stuff apart for us anyway, and we do not want to have a potentially
3660 # function-converted external order_by
3661 # ( there is an explicit if ( collapse && _grouped_by_distinct ) check in DBIHacks )
3662 $attrs->{order_by} = $new_order unless $attrs->{collapse};
3666 # inject prefetch-bound selection (if any)
3667 push @{$attrs->{select}}, @prefetch_select;
3668 push @{$attrs->{as}}, @prefetch_as;
3670 # whether we can get away with the dumbest (possibly DBI-internal) collapser
3671 if ( List::Util::first { $_ =~ /\./ } @{$attrs->{as}} ) {
3672 $attrs->{_related_results_construction} = 1;
3675 # if both page and offset are specified, produce a combined offset
3676 # even though it doesn't make much sense, this is what pre 081xx has
3678 if (my $page = delete $attrs->{page}) {
3680 ($attrs->{rows} * ($page - 1))
3682 ($attrs->{offset} || 0)
3686 return $self->{_attrs} = $attrs;
3690 my ($self, $attr) = @_;
3692 if (ref $attr eq 'HASH') {
3693 return $self->_rollout_hash($attr);
3694 } elsif (ref $attr eq 'ARRAY') {
3695 return $self->_rollout_array($attr);
3701 sub _rollout_array {
3702 my ($self, $attr) = @_;
3705 foreach my $element (@{$attr}) {
3706 if (ref $element eq 'HASH') {
3707 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3708 } elsif (ref $element eq 'ARRAY') {
3709 # XXX - should probably recurse here
3710 push( @rolled_array, @{$self->_rollout_array($element)} );
3712 push( @rolled_array, $element );
3715 return \@rolled_array;
3719 my ($self, $attr) = @_;
3722 foreach my $key (keys %{$attr}) {
3723 push( @rolled_array, { $key => $attr->{$key} } );
3725 return \@rolled_array;
3728 sub _calculate_score {
3729 my ($self, $a, $b) = @_;
3731 if (defined $a xor defined $b) {
3734 elsif (not defined $a) {
3738 if (ref $b eq 'HASH') {
3739 my ($b_key) = keys %{$b};
3740 if (ref $a eq 'HASH') {
3741 my ($a_key) = keys %{$a};
3742 if ($a_key eq $b_key) {
3743 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3748 return ($a eq $b_key) ? 1 : 0;
3751 if (ref $a eq 'HASH') {
3752 my ($a_key) = keys %{$a};
3753 return ($b eq $a_key) ? 1 : 0;
3755 return ($b eq $a) ? 1 : 0;
3760 sub _merge_joinpref_attr {
3761 my ($self, $orig, $import) = @_;
3763 return $import unless defined($orig);
3764 return $orig unless defined($import);
3766 $orig = $self->_rollout_attr($orig);
3767 $import = $self->_rollout_attr($import);
3770 foreach my $import_element ( @{$import} ) {
3771 # find best candidate from $orig to merge $b_element into
3772 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3773 foreach my $orig_element ( @{$orig} ) {
3774 my $score = $self->_calculate_score( $orig_element, $import_element );
3775 if ($score > $best_candidate->{score}) {
3776 $best_candidate->{position} = $position;
3777 $best_candidate->{score} = $score;
3781 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3782 $import_key = '' if not defined $import_key;
3784 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3785 push( @{$orig}, $import_element );
3787 my $orig_best = $orig->[$best_candidate->{position}];
3788 # merge orig_best and b_element together and replace original with merged
3789 if (ref $orig_best ne 'HASH') {
3790 $orig->[$best_candidate->{position}] = $import_element;
3791 } elsif (ref $import_element eq 'HASH') {
3792 my ($key) = keys %{$orig_best};
3793 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3796 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3799 return @$orig ? $orig : ();
3807 require Hash::Merge;
3808 my $hm = Hash::Merge->new;
3810 $hm->specify_behavior({
3813 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
3815 if ($defl xor $defr) {
3816 return [ $defl ? $_[0] : $_[1] ];
3821 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
3825 return [$_[0], $_[1]];
3829 return $_[1] if !defined $_[0];
3830 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3831 return [$_[0], @{$_[1]}]
3834 return [] if !defined $_[0] and !keys %{$_[1]};
3835 return [ $_[1] ] if !defined $_[0];
3836 return [ $_[0] ] if !keys %{$_[1]};
3837 return [$_[0], $_[1]]
3842 return $_[0] if !defined $_[1];
3843 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3844 return [@{$_[0]}, $_[1]]
3847 my @ret = @{$_[0]} or return $_[1];
3848 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
3849 my %idx = map { $_ => 1 } @ret;
3850 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
3854 return [ $_[1] ] if ! @{$_[0]};
3855 return $_[0] if !keys %{$_[1]};
3856 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3857 return [ @{$_[0]}, $_[1] ];
3862 return [] if !keys %{$_[0]} and !defined $_[1];
3863 return [ $_[0] ] if !defined $_[1];
3864 return [ $_[1] ] if !keys %{$_[0]};
3865 return [$_[0], $_[1]]
3868 return [] if !keys %{$_[0]} and !@{$_[1]};
3869 return [ $_[0] ] if !@{$_[1]};
3870 return $_[1] if !keys %{$_[0]};
3871 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3872 return [ $_[0], @{$_[1]} ];
3875 return [] if !keys %{$_[0]} and !keys %{$_[1]};
3876 return [ $_[0] ] if !keys %{$_[1]};
3877 return [ $_[1] ] if !keys %{$_[0]};
3878 return [ $_[0] ] if $_[0] eq $_[1];
3879 return [ $_[0], $_[1] ];
3882 } => 'DBIC_RS_ATTR_MERGER');
3886 return $hm->merge ($_[1], $_[2]);
3890 sub STORABLE_freeze {
3891 my ($self, $cloning) = @_;
3892 my $to_serialize = { %$self };
3894 # A cursor in progress can't be serialized (and would make little sense anyway)
3895 # the parser can be regenerated (and can't be serialized)
3896 delete @{$to_serialize}{qw/cursor _row_parser _result_inflator/};
3898 # nor is it sensical to store a not-yet-fired-count pager
3899 if ($to_serialize->{pager} and ref $to_serialize->{pager}{total_entries} eq 'CODE') {
3900 delete $to_serialize->{pager};
3903 Storable::nfreeze($to_serialize);
3906 # need this hook for symmetry
3908 my ($self, $cloning, $serialized) = @_;
3910 %$self = %{ Storable::thaw($serialized) };
3916 =head2 throw_exception
3918 See L<DBIx::Class::Schema/throw_exception> for details.
3922 sub throw_exception {
3925 if (ref $self and my $rsrc = $self->result_source) {
3926 $rsrc->throw_exception(@_)
3929 DBIx::Class::Exception->throw(@_);
3937 # XXX: FIXME: Attributes docs need clearing up
3941 Attributes are used to refine a ResultSet in various ways when
3942 searching for data. They can be passed to any method which takes an
3943 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3946 Default attributes can be set on the result class using
3947 L<DBIx::Class::ResultSource/resultset_attributes>. (Please read
3948 the CAVEATS on that feature before using it!)
3950 These are in no particular order:
3956 =item Value: ( $order_by | \@order_by | \%order_by )
3960 Which column(s) to order the results by.
3962 [The full list of suitable values is documented in
3963 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3966 If a single column name, or an arrayref of names is supplied, the
3967 argument is passed through directly to SQL. The hashref syntax allows
3968 for connection-agnostic specification of ordering direction:
3970 For descending order:
3972 order_by => { -desc => [qw/col1 col2 col3/] }
3974 For explicit ascending order:
3976 order_by => { -asc => 'col' }
3978 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3979 supported, although you are strongly encouraged to use the hashref
3980 syntax as outlined above.
3986 =item Value: \@columns | \%columns | $column
3990 Shortcut to request a particular set of columns to be retrieved. Each
3991 column spec may be a string (a table column name), or a hash (in which
3992 case the key is the C<as> value, and the value is used as the C<select>
3993 expression). Adds C<me.> onto the start of any column without a C<.> in
3994 it and sets C<select> from that, then auto-populates C<as> from
3995 C<select> as normal. (You may also use the C<cols> attribute, as in
3996 earlier versions of DBIC, but this is deprecated.)
3998 Essentially C<columns> does the same as L</select> and L</as>.
4000 columns => [ 'foo', { bar => 'baz' } ]
4004 select => [qw/foo baz/],
4011 =item Value: \@columns
4015 Indicates additional columns to be selected from storage. Works the same as
4016 L</columns> but adds columns to the selection. (You may also use the
4017 C<include_columns> attribute, as in earlier versions of DBIC, but this is
4018 deprecated). For example:-
4020 $schema->resultset('CD')->search(undef, {
4021 '+columns' => ['artist.name'],
4025 would return all CDs and include a 'name' column to the information
4026 passed to object inflation. Note that the 'artist' is the name of the
4027 column (or relationship) accessor, and 'name' is the name of the column
4028 accessor in the related table.
4030 B<NOTE:> You need to explicitly quote '+columns' when defining the attribute.
4031 Not doing so causes Perl to incorrectly interpret +columns as a bareword with a
4032 unary plus operator before it.
4034 =head2 include_columns
4038 =item Value: \@columns
4042 Deprecated. Acts as a synonym for L</+columns> for backward compatibility.
4048 =item Value: \@select_columns
4052 Indicates which columns should be selected from the storage. You can use
4053 column names, or in the case of RDBMS back ends, function or stored procedure
4056 $rs = $schema->resultset('Employee')->search(undef, {
4059 { count => 'employeeid' },
4060 { max => { length => 'name' }, -as => 'longest_name' }
4065 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
4067 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
4068 use L</select>, to instruct DBIx::Class how to store the result of the column.
4069 Also note that the L</as> attribute has nothing to do with the SQL-side 'AS'
4070 identifier aliasing. You can however alias a function, so you can use it in
4071 e.g. an C<ORDER BY> clause. This is done via the C<-as> B<select function
4072 attribute> supplied as shown in the example above.
4074 B<NOTE:> You need to explicitly quote '+select'/'+as' when defining the attributes.
4075 Not doing so causes Perl to incorrectly interpret them as a bareword with a
4076 unary plus operator before it.
4082 Indicates additional columns to be selected from storage. Works the same as
4083 L</select> but adds columns to the default selection, instead of specifying
4092 =item Value: \@inflation_names
4096 Indicates column names for object inflation. That is L</as> indicates the
4097 slot name in which the column value will be stored within the
4098 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
4099 identifier by the C<get_column> method (or via the object accessor B<if one
4100 with the same name already exists>) as shown below. The L</as> attribute has
4101 B<nothing to do> with the SQL-side C<AS>. See L</select> for details.
4103 $rs = $schema->resultset('Employee')->search(undef, {
4106 { count => 'employeeid' },
4107 { max => { length => 'name' }, -as => 'longest_name' }
4116 If the object against which the search is performed already has an accessor
4117 matching a column name specified in C<as>, the value can be retrieved using
4118 the accessor as normal:
4120 my $name = $employee->name();
4122 If on the other hand an accessor does not exist in the object, you need to
4123 use C<get_column> instead:
4125 my $employee_count = $employee->get_column('employee_count');
4127 You can create your own accessors if required - see
4128 L<DBIx::Class::Manual::Cookbook> for details.
4134 Indicates additional column names for those added via L</+select>. See L</as>.
4142 =item Value: ($rel_name | \@rel_names | \%rel_names)
4146 Contains a list of relationships that should be joined for this query. For
4149 # Get CDs by Nine Inch Nails
4150 my $rs = $schema->resultset('CD')->search(
4151 { 'artist.name' => 'Nine Inch Nails' },
4152 { join => 'artist' }
4155 Can also contain a hash reference to refer to the other relation's relations.
4158 package MyApp::Schema::Track;
4159 use base qw/DBIx::Class/;
4160 __PACKAGE__->table('track');
4161 __PACKAGE__->add_columns(qw/trackid cd position title/);
4162 __PACKAGE__->set_primary_key('trackid');
4163 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
4166 # In your application
4167 my $rs = $schema->resultset('Artist')->search(
4168 { 'track.title' => 'Teardrop' },
4170 join => { cd => 'track' },
4171 order_by => 'artist.name',
4175 You need to use the relationship (not the table) name in conditions,
4176 because they are aliased as such. The current table is aliased as "me", so
4177 you need to use me.column_name in order to avoid ambiguity. For example:
4179 # Get CDs from 1984 with a 'Foo' track
4180 my $rs = $schema->resultset('CD')->search(
4183 'tracks.name' => 'Foo'
4185 { join => 'tracks' }
4188 If the same join is supplied twice, it will be aliased to <rel>_2 (and
4189 similarly for a third time). For e.g.
4191 my $rs = $schema->resultset('Artist')->search({
4192 'cds.title' => 'Down to Earth',
4193 'cds_2.title' => 'Popular',
4195 join => [ qw/cds cds/ ],
4198 will return a set of all artists that have both a cd with title 'Down
4199 to Earth' and a cd with title 'Popular'.
4201 If you want to fetch related objects from other tables as well, see L</prefetch>
4204 NOTE: An internal join-chain pruner will discard certain joins while
4205 constructing the actual SQL query, as long as the joins in question do not
4206 affect the retrieved result. This for example includes 1:1 left joins
4207 that are not part of the restriction specification (WHERE/HAVING) nor are
4208 a part of the query selection.
4210 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
4216 =item Value: (0 | 1)
4220 When set to a true value, indicates that any rows fetched from joined has_many
4221 relationships are to be aggregated into the corresponding "parent" object. For
4222 example, the resultset:
4224 my $rs = $schema->resultset('CD')->search({}, {
4225 '+columns' => [ qw/ tracks.title tracks.position / ],
4230 While executing the following query:
4232 SELECT me.*, tracks.title, tracks.position
4234 LEFT JOIN track tracks
4235 ON tracks.cdid = me.cdid
4237 Will return only as many objects as there are rows in the CD source, even
4238 though the result of the query may span many rows. Each of these CD objects
4239 will in turn have multiple "Track" objects hidden behind the has_many
4240 generated accessor C<tracks>. Without C<< collapse => 1 >>, the return values
4241 of this resultset would be as many CD objects as there are tracks (a "Cartesian
4242 product"), with each CD object containing exactly one of all fetched Track data.
4244 When a collapse is requested on a non-ordered resultset, an order by some
4245 unique part of the main source (the left-most table) is inserted automatically.
4246 This is done so that the resultset is allowed to be "lazy" - calling
4247 L<< $rs->next|/next >> will fetch only as many rows as it needs to build the next
4248 object with all of its related data.
4250 If an L</order_by> is already declared, and orders the resultset in a way that
4251 makes collapsing as described above impossible (e.g. C<< ORDER BY
4252 has_many_rel.column >> or C<ORDER BY RANDOM()>), DBIC will automatically
4253 switch to "eager" mode and slurp the entire resultset before constructing the
4254 first object returned by L</next>.
4256 Setting this attribute on a resultset that does not join any has_many
4257 relations is a no-op.
4259 For a more in-depth discussion, see L</PREFETCHING>.
4265 =item Value: ($rel_name | \@rel_names | \%rel_names)
4269 This attribute is a shorthand for specifying a L</join> spec, adding all
4270 columns from the joined related sources as L</+columns> and setting
4271 L</collapse> to a true value. For example, the following two queries are
4274 my $rs = $schema->resultset('Artist')->search({}, {
4275 prefetch => { cds => ['genre', 'tracks' ] },
4280 my $rs = $schema->resultset('Artist')->search({}, {
4281 join => { cds => ['genre', 'tracks' ] },
4285 { +{ "cds.$_" => "cds.$_" } }
4286 $schema->source('Artist')->related_source('cds')->columns
4289 { +{ "cds.genre.$_" => "genre.$_" } }
4290 $schema->source('Artist')->related_source('cds')->related_source('genre')->columns
4293 { +{ "cds.tracks.$_" => "tracks.$_" } }
4294 $schema->source('Artist')->related_source('cds')->related_source('tracks')->columns
4299 Both producing the following SQL:
4301 SELECT me.artistid, me.name, me.rank, me.charfield,
4302 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track,
4303 genre.genreid, genre.name,
4304 tracks.trackid, tracks.cd, tracks.position, tracks.title, tracks.last_updated_on, tracks.last_updated_at
4307 ON cds.artist = me.artistid
4308 LEFT JOIN genre genre
4309 ON genre.genreid = cds.genreid
4310 LEFT JOIN track tracks
4311 ON tracks.cd = cds.cdid
4312 ORDER BY me.artistid
4314 While L</prefetch> implies a L</join>, it is ok to mix the two together, as
4315 the arguments are properly merged and generally do the right thing. For
4316 example, you may want to do the following:
4318 my $artists_and_cds_without_genre = $schema->resultset('Artist')->search(
4319 { 'genre.genreid' => undef },
4321 join => { cds => 'genre' },
4326 Which generates the following SQL:
4328 SELECT me.artistid, me.name, me.rank, me.charfield,
4329 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track
4332 ON cds.artist = me.artistid
4333 LEFT JOIN genre genre
4334 ON genre.genreid = cds.genreid
4335 WHERE genre.genreid IS NULL
4336 ORDER BY me.artistid
4338 For a more in-depth discussion, see L</PREFETCHING>.
4344 =item Value: $source_alias
4348 Sets the source alias for the query. Normally, this defaults to C<me>, but
4349 nested search queries (sub-SELECTs) might need specific aliases set to
4350 reference inner queries. For example:
4353 ->related_resultset('CDs')
4354 ->related_resultset('Tracks')
4356 'track.id' => { -ident => 'none_search.id' },
4360 my $ids = $self->search({
4363 alias => 'none_search',
4364 group_by => 'none_search.id',
4365 })->get_column('id')->as_query;
4367 $self->search({ id => { -in => $ids } })
4369 This attribute is directly tied to L</current_source_alias>.
4379 Makes the resultset paged and specifies the page to retrieve. Effectively
4380 identical to creating a non-pages resultset and then calling ->page($page)
4383 If L</rows> attribute is not specified it defaults to 10 rows per page.
4385 When you have a paged resultset, L</count> will only return the number
4386 of rows in the page. To get the total, use the L</pager> and call
4387 C<total_entries> on it.
4397 Specifies the maximum number of rows for direct retrieval or the number of
4398 rows per page if the page attribute or method is used.
4404 =item Value: $offset
4408 Specifies the (zero-based) row number for the first row to be returned, or the
4409 of the first row of the first page if paging is used.
4411 =head2 software_limit
4415 =item Value: (0 | 1)
4419 When combined with L</rows> and/or L</offset> the generated SQL will not
4420 include any limit dialect stanzas. Instead the entire result will be selected
4421 as if no limits were specified, and DBIC will perform the limit locally, by
4422 artificially advancing and finishing the resulting L</cursor>.
4424 This is the recommended way of performing resultset limiting when no sane RDBMS
4425 implementation is available (e.g.
4426 L<Sybase ASE|DBIx::Class::Storage::DBI::Sybase::ASE> using the
4427 L<Generic Sub Query|DBIx::Class::SQLMaker::LimitDialects/GenericSubQ> hack)
4433 =item Value: \@columns
4437 A arrayref of columns to group by. Can include columns of joined tables.
4439 group_by => [qw/ column1 column2 ... /]
4445 =item Value: $condition
4449 HAVING is a select statement attribute that is applied between GROUP BY and
4450 ORDER BY. It is applied to the after the grouping calculations have been
4453 having => { 'count_employee' => { '>=', 100 } }
4455 or with an in-place function in which case literal SQL is required:
4457 having => \[ 'count(employee) >= ?', [ count => 100 ] ]
4463 =item Value: (0 | 1)
4467 Set to 1 to automatically generate a L</group_by> clause based on the selection
4468 (including intelligent handling of L</order_by> contents). Note that the group
4469 criteria calculation takes place over the B<final> selection. This includes
4470 any L</+columns>, L</+select> or L</order_by> additions in subsequent
4471 L</search> calls, and standalone columns selected via
4472 L<DBIx::Class::ResultSetColumn> (L</get_column>). A notable exception are the
4473 extra selections specified via L</prefetch> - such selections are explicitly
4474 excluded from group criteria calculations.
4476 If the final ResultSet also explicitly defines a L</group_by> attribute, this
4477 setting is ignored and an appropriate warning is issued.
4483 Adds to the WHERE clause.
4485 # only return rows WHERE deleted IS NULL for all searches
4486 __PACKAGE__->resultset_attributes({ where => { deleted => undef } });
4488 Can be overridden by passing C<< { where => undef } >> as an attribute
4491 For more complicated where clauses see L<SQL::Abstract/WHERE CLAUSES>.
4497 Set to 1 to cache search results. This prevents extra SQL queries if you
4498 revisit rows in your ResultSet:
4500 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
4502 while( my $artist = $resultset->next ) {
4506 $rs->first; # without cache, this would issue a query
4508 By default, searches are not cached.
4510 For more examples of using these attributes, see
4511 L<DBIx::Class::Manual::Cookbook>.
4517 =item Value: ( 'update' | 'shared' | \$scalar )
4521 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT
4522 ... FOR SHARED. If \$scalar is passed, this is taken directly and embedded in the
4527 DBIx::Class supports arbitrary related data prefetching from multiple related
4528 sources. Any combination of relationship types and column sets are supported.
4529 If L<collapsing|/collapse> is requested, there is an additional requirement of
4530 selecting enough data to make every individual object uniquely identifiable.
4532 Here are some more involved examples, based on the following relationship map:
4535 My::Schema::CD->belongs_to( artist => 'My::Schema::Artist' );
4536 My::Schema::CD->might_have( liner_note => 'My::Schema::LinerNotes' );
4537 My::Schema::CD->has_many( tracks => 'My::Schema::Track' );
4539 My::Schema::Artist->belongs_to( record_label => 'My::Schema::RecordLabel' );
4541 My::Schema::Track->has_many( guests => 'My::Schema::Guest' );
4545 my $rs = $schema->resultset('Tag')->search(
4554 The initial search results in SQL like the following:
4556 SELECT tag.*, cd.*, artist.* FROM tag
4557 JOIN cd ON tag.cd = cd.cdid
4558 JOIN artist ON cd.artist = artist.artistid
4560 L<DBIx::Class> has no need to go back to the database when we access the
4561 C<cd> or C<artist> relationships, which saves us two SQL statements in this
4564 Simple prefetches will be joined automatically, so there is no need
4565 for a C<join> attribute in the above search.
4567 The L</prefetch> attribute can be used with any of the relationship types
4568 and multiple prefetches can be specified together. Below is a more complex
4569 example that prefetches a CD's artist, its liner notes (if present),
4570 the cover image, the tracks on that CD, and the guests on those
4573 my $rs = $schema->resultset('CD')->search(
4577 { artist => 'record_label'}, # belongs_to => belongs_to
4578 'liner_note', # might_have
4579 'cover_image', # has_one
4580 { tracks => 'guests' }, # has_many => has_many
4585 This will produce SQL like the following:
4587 SELECT cd.*, artist.*, record_label.*, liner_note.*, cover_image.*,
4591 ON artist.artistid = me.artistid
4592 JOIN record_label record_label
4593 ON record_label.labelid = artist.labelid
4594 LEFT JOIN track tracks
4595 ON tracks.cdid = me.cdid
4596 LEFT JOIN guest guests
4597 ON guests.trackid = track.trackid
4598 LEFT JOIN liner_notes liner_note
4599 ON liner_note.cdid = me.cdid
4600 JOIN cd_artwork cover_image
4601 ON cover_image.cdid = me.cdid
4604 Now the C<artist>, C<record_label>, C<liner_note>, C<cover_image>,
4605 C<tracks>, and C<guests> of the CD will all be available through the
4606 relationship accessors without the need for additional queries to the
4611 Prefetch does a lot of deep magic. As such, it may not behave exactly
4612 as you might expect.
4618 Prefetch uses the L</cache> to populate the prefetched relationships. This
4619 may or may not be what you want.
4623 If you specify a condition on a prefetched relationship, ONLY those
4624 rows that match the prefetched condition will be fetched into that relationship.
4625 This means that adding prefetch to a search() B<may alter> what is returned by
4626 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
4628 my $artist_rs = $schema->resultset('Artist')->search({
4634 my $count = $artist_rs->first->cds->count;
4636 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
4638 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
4640 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
4642 That cmp_ok() may or may not pass depending on the datasets involved. In other
4643 words the C<WHERE> condition would apply to the entire dataset, just like
4644 it would in regular SQL. If you want to add a condition only to the "right side"
4645 of a C<LEFT JOIN> - consider declaring and using a L<relationship with a custom
4646 condition|DBIx::Class::Relationship::Base/condition>
4650 =head1 DBIC BIND VALUES
4652 Because DBIC may need more information to bind values than just the column name
4653 and value itself, it uses a special format for both passing and receiving bind
4654 values. Each bind value should be composed of an arrayref of
4655 C<< [ \%args => $val ] >>. The format of C<< \%args >> is currently:
4661 If present (in any form), this is what is being passed directly to bind_param.
4662 Note that different DBD's expect different bind args. (e.g. DBD::SQLite takes
4663 a single numerical type, while DBD::Pg takes a hashref if bind options.)
4665 If this is specified, all other bind options described below are ignored.
4669 If present, this is used to infer the actual bind attribute by passing to
4670 C<< $resolved_storage->bind_attribute_by_data_type() >>. Defaults to the
4671 "data_type" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>.
4673 Note that the data type is somewhat freeform (hence the sqlt_ prefix);
4674 currently drivers are expected to "Do the Right Thing" when given a common
4675 datatype name. (Not ideal, but that's what we got at this point.)
4679 Currently used to correctly allocate buffers for bind_param_inout().
4680 Defaults to "size" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>,
4681 or to a sensible value based on the "data_type".
4685 Used to fill in missing sqlt_datatype and sqlt_size attributes (if they are
4686 explicitly specified they are never overridden). Also used by some weird DBDs,
4687 where the column name should be available at bind_param time (e.g. Oracle).
4691 For backwards compatibility and convenience, the following shortcuts are
4694 [ $name => $val ] === [ { dbic_colname => $name }, $val ]
4695 [ \$dt => $val ] === [ { sqlt_datatype => $dt }, $val ]
4696 [ undef, $val ] === [ {}, $val ]
4697 $val === [ {}, $val ]
4699 =head1 AUTHOR AND CONTRIBUTORS
4701 See L<AUTHOR|DBIx::Class/AUTHOR> and L<CONTRIBUTORS|DBIx::Class/CONTRIBUTORS> in DBIx::Class
4705 You may distribute this code under the same terms as Perl itself.