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 qw(
10 fail_on_internal_wantarray is_plain_value is_literal_value
13 use Data::Compare (); # no imports!!! guard against insane architecture
15 # not importing first() as it will clash with our own method
19 # De-duplication in _merge_attr() is disabled, but left in for reference
20 # (the merger is used for other things that ought not to be de-duped)
21 *__HM_DEDUP = sub () { 0 };
31 # this is real - CDBICompat overrides it with insanity
32 # yes, prototype won't matter, but that's for now ;)
35 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class result_source/);
39 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
43 my $users_rs = $schema->resultset('User');
44 while( $user = $users_rs->next) {
45 print $user->username;
48 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
49 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
53 A ResultSet is an object which stores a set of conditions representing
54 a query. It is the backbone of DBIx::Class (i.e. the really
55 important/useful bit).
57 No SQL is executed on the database when a ResultSet is created, it
58 just stores all the conditions needed to create the query.
60 A basic ResultSet representing the data of an entire table is returned
61 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
62 L<Source|DBIx::Class::Manual::Glossary/Source> name.
64 my $users_rs = $schema->resultset('User');
66 A new ResultSet is returned from calling L</search> on an existing
67 ResultSet. The new one will contain all the conditions of the
68 original, plus any new conditions added in the C<search> call.
70 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
71 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
74 The query that the ResultSet represents is B<only> executed against
75 the database when these methods are called:
76 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
78 If a resultset is used in a numeric context it returns the L</count>.
79 However, if it is used in a boolean context it is B<always> true. So if
80 you want to check if a resultset has any results, you must use C<if $rs
83 =head1 CUSTOM ResultSet CLASSES THAT USE Moose
85 If you want to make your custom ResultSet classes with L<Moose>, use a template
88 package MyApp::Schema::ResultSet::User;
91 use namespace::autoclean;
93 extends 'DBIx::Class::ResultSet';
95 sub BUILDARGS { $_[2] }
99 __PACKAGE__->meta->make_immutable;
103 The L<MooseX::NonMoose> is necessary so that the L<Moose> constructor does not
104 clash with the regular ResultSet constructor. Alternatively, you can use:
106 __PACKAGE__->meta->make_immutable(inline_constructor => 0);
108 The L<BUILDARGS|Moose::Manual::Construction/BUILDARGS> is necessary because the
109 signature of the ResultSet C<new> is C<< ->new($source, \%args) >>.
113 =head2 Chaining resultsets
115 Let's say you've got a query that needs to be run to return some data
116 to the user. But, you have an authorization system in place that
117 prevents certain users from seeing certain information. So, you want
118 to construct the basic query in one method, but add constraints to it in
123 my $request = $self->get_request; # Get a request object somehow.
124 my $schema = $self->result_source->schema;
126 my $cd_rs = $schema->resultset('CD')->search({
127 title => $request->param('title'),
128 year => $request->param('year'),
131 $cd_rs = $self->apply_security_policy( $cd_rs );
133 return $cd_rs->all();
136 sub apply_security_policy {
145 =head3 Resolving conditions and attributes
147 When a resultset is chained from another resultset (e.g.:
148 C<< my $new_rs = $old_rs->search(\%extra_cond, \%attrs) >>), conditions
149 and attributes with the same keys need resolving.
151 If any of L</columns>, L</select>, L</as> are present, they reset the
152 original selection, and start the selection "clean".
154 The L</join>, L</prefetch>, L</+columns>, L</+select>, L</+as> attributes
155 are merged into the existing ones from the original resultset.
157 The L</where> and L</having> attributes, and any search conditions, are
158 merged with an SQL C<AND> to the existing condition from the original
161 All other attributes are overridden by any new ones supplied in the
164 =head2 Multiple queries
166 Since a resultset just defines a query, you can do all sorts of
167 things with it with the same object.
169 # Don't hit the DB yet.
170 my $cd_rs = $schema->resultset('CD')->search({
171 title => 'something',
175 # Each of these hits the DB individually.
176 my $count = $cd_rs->count;
177 my $most_recent = $cd_rs->get_column('date_released')->max();
178 my @records = $cd_rs->all;
180 And it's not just limited to SELECT statements.
186 $cd_rs->create({ artist => 'Fred' });
188 Which is the same as:
190 $schema->resultset('CD')->create({
191 title => 'something',
196 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
204 =item Arguments: L<$source|DBIx::Class::ResultSource>, L<\%attrs?|/ATTRIBUTES>
206 =item Return Value: L<$resultset|/search>
210 The resultset constructor. Takes a source object (usually a
211 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
212 L</ATTRIBUTES> below). Does not perform any queries -- these are
213 executed as needed by the other methods.
215 Generally you never construct a resultset manually. Instead you get one
217 C<< $schema->L<resultset|DBIx::Class::Schema/resultset>('$source_name') >>
218 or C<< $another_resultset->L<search|/search>(...) >> (the later called in
221 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
227 If called on an object, proxies to L</new_result> instead, so
229 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
231 will return a CD object, not a ResultSet, and is equivalent to:
233 my $cd = $schema->resultset('CD')->new_result({ title => 'Spoon' });
235 Please also keep in mind that many internals call L</new_result> directly,
236 so overloading this method with the idea of intercepting new result object
237 creation B<will not work>. See also warning pertaining to L</create>.
245 return $class->new_result(@_) if ref $class;
247 my ($source, $attrs) = @_;
248 $source = $source->resolve
249 if $source->isa('DBIx::Class::ResultSourceHandle');
251 $attrs = { %{$attrs||{}} };
252 delete @{$attrs}{qw(_last_sqlmaker_alias_map _related_results_construction)};
254 if ($attrs->{page}) {
255 $attrs->{rows} ||= 10;
258 $attrs->{alias} ||= 'me';
261 result_source => $source,
262 cond => $attrs->{where},
267 # if there is a dark selector, this means we are already in a
268 # chain and the cleanup/sanification was taken care of by
270 $self->_normalize_selection($attrs)
271 unless $attrs->{_dark_selector};
274 $attrs->{result_class} || $source->result_class
284 =item Arguments: L<$cond|DBIx::Class::SQLMaker> | undef, L<\%attrs?|/ATTRIBUTES>
286 =item Return Value: $resultset (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
290 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
291 my $new_rs = $cd_rs->search({ year => 2005 });
293 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
294 # year = 2005 OR year = 2004
296 In list context, C<< ->all() >> is called implicitly on the resultset, thus
297 returning a list of L<result|DBIx::Class::Manual::ResultClass> objects instead.
298 To avoid that, use L</search_rs>.
300 If you need to pass in additional attributes but no additional condition,
301 call it as C<search(undef, \%attrs)>.
303 # "SELECT name, artistid FROM $artist_table"
304 my @all_artists = $schema->resultset('Artist')->search(undef, {
305 columns => [qw/name artistid/],
308 For a list of attributes that can be passed to C<search>, see
309 L</ATTRIBUTES>. For more examples of using this function, see
310 L<Searching|DBIx::Class::Manual::Cookbook/SEARCHING>. For a complete
311 documentation for the first argument, see L<SQL::Abstract/"WHERE CLAUSES">
312 and its extension L<DBIx::Class::SQLMaker>.
314 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
318 Note that L</search> does not process/deflate any of the values passed in the
319 L<SQL::Abstract>-compatible search condition structure. This is unlike other
320 condition-bound methods L</new_result>, L</create> and L</find>. The user must ensure
321 manually that any value passed to this method will stringify to something the
322 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
323 objects, for more info see:
324 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
330 my $rs = $self->search_rs( @_ );
333 DBIx::Class::_ENV_::ASSERT_NO_INTERNAL_WANTARRAY and my $sog = fail_on_internal_wantarray($rs);
336 elsif (defined wantarray) {
340 # we can be called by a relationship helper, which in
341 # turn may be called in void context due to some braindead
342 # overload or whatever else the user decided to be clever
343 # at this particular day. Thus limit the exception to
344 # external code calls only
345 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense')
346 if (caller)[0] !~ /^\QDBIx::Class::/;
356 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
358 =item Return Value: L<$resultset|/search>
362 This method does the same exact thing as search() except it will
363 always return a resultset, even in list context.
370 my $rsrc = $self->result_source;
371 my ($call_cond, $call_attrs);
373 # Special-case handling for (undef, undef) or (undef)
374 # Note that (foo => undef) is valid deprecated syntax
375 @_ = () if not scalar grep { defined $_ } @_;
381 # fish out attrs in the ($condref, $attr) case
382 elsif (@_ == 2 and ( ! defined $_[0] or (ref $_[0]) ne '') ) {
383 ($call_cond, $call_attrs) = @_;
386 $self->throw_exception('Odd number of arguments to search')
390 carp_unique 'search( %condition ) is deprecated, use search( \%condition ) instead'
391 unless $rsrc->result_class->isa('DBIx::Class::CDBICompat');
393 for my $i (0 .. $#_) {
395 $self->throw_exception ('All keys in condition key/value pairs must be plain scalars')
396 if (! defined $_[$i] or ref $_[$i] ne '');
402 # see if we can keep the cache (no $rs changes)
404 my %safe = (alias => 1, cache => 1);
405 if ( ! List::Util::first { !$safe{$_} } keys %$call_attrs and (
408 ref $call_cond eq 'HASH' && ! keys %$call_cond
410 ref $call_cond eq 'ARRAY' && ! @$call_cond
412 $cache = $self->get_cache;
415 my $old_attrs = { %{$self->{attrs}} };
416 my ($old_having, $old_where) = delete @{$old_attrs}{qw(having where)};
418 my $new_attrs = { %$old_attrs };
420 # take care of call attrs (only if anything is changing)
421 if ($call_attrs and keys %$call_attrs) {
423 # copy for _normalize_selection
424 $call_attrs = { %$call_attrs };
426 my @selector_attrs = qw/select as columns cols +select +as +columns include_columns/;
428 # reset the current selector list if new selectors are supplied
429 if (List::Util::first { exists $call_attrs->{$_} } qw/columns cols select as/) {
430 delete @{$old_attrs}{(@selector_attrs, '_dark_selector')};
433 # Normalize the new selector list (operates on the passed-in attr structure)
434 # Need to do it on every chain instead of only once on _resolved_attrs, in
435 # order to allow detection of empty vs partial 'as'
436 $call_attrs->{_dark_selector} = $old_attrs->{_dark_selector}
437 if $old_attrs->{_dark_selector};
438 $self->_normalize_selection ($call_attrs);
440 # start with blind overwriting merge, exclude selector attrs
441 $new_attrs = { %{$old_attrs}, %{$call_attrs} };
442 delete @{$new_attrs}{@selector_attrs};
444 for (@selector_attrs) {
445 $new_attrs->{$_} = $self->_merge_attr($old_attrs->{$_}, $call_attrs->{$_})
446 if ( exists $old_attrs->{$_} or exists $call_attrs->{$_} );
449 # older deprecated name, use only if {columns} is not there
450 if (my $c = delete $new_attrs->{cols}) {
451 carp_unique( "Resultset attribute 'cols' is deprecated, use 'columns' instead" );
452 if ($new_attrs->{columns}) {
453 carp "Resultset specifies both the 'columns' and the legacy 'cols' attributes - ignoring 'cols'";
456 $new_attrs->{columns} = $c;
461 # join/prefetch use their own crazy merging heuristics
462 foreach my $key (qw/join prefetch/) {
463 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key})
464 if exists $call_attrs->{$key};
467 # stack binds together
468 $new_attrs->{bind} = [ @{ $old_attrs->{bind} || [] }, @{ $call_attrs->{bind} || [] } ];
472 for ($old_where, $call_cond) {
474 $new_attrs->{where} = $self->_stack_cond (
475 $_, $new_attrs->{where}
480 if (defined $old_having) {
481 $new_attrs->{having} = $self->_stack_cond (
482 $old_having, $new_attrs->{having}
486 my $rs = (ref $self)->new($rsrc, $new_attrs);
488 $rs->set_cache($cache) if ($cache);
494 sub _normalize_selection {
495 my ($self, $attrs) = @_;
498 if ( exists $attrs->{include_columns} ) {
499 carp_unique( "Resultset attribute 'include_columns' is deprecated, use '+columns' instead" );
500 $attrs->{'+columns'} = $self->_merge_attr(
501 $attrs->{'+columns'}, delete $attrs->{include_columns}
505 # columns are always placed first, however
507 # Keep the X vs +X separation until _resolved_attrs time - this allows to
508 # delay the decision on whether to use a default select list ($rsrc->columns)
509 # allowing stuff like the remove_columns helper to work
511 # select/as +select/+as pairs need special handling - the amount of select/as
512 # elements in each pair does *not* have to be equal (think multicolumn
513 # selectors like distinct(foo, bar) ). If the selector is bare (no 'as'
514 # supplied at all) - try to infer the alias, either from the -as parameter
515 # of the selector spec, or use the parameter whole if it looks like a column
516 # name (ugly legacy heuristic). If all fails - leave the selector bare (which
517 # is ok as well), but make sure no more additions to the 'as' chain take place
518 for my $pref ('', '+') {
520 my ($sel, $as) = map {
521 my $key = "${pref}${_}";
523 my $val = [ ref $attrs->{$key} eq 'ARRAY'
525 : $attrs->{$key} || ()
527 delete $attrs->{$key};
531 if (! @$as and ! @$sel ) {
534 elsif (@$as and ! @$sel) {
535 $self->throw_exception(
536 "Unable to handle ${pref}as specification (@$as) without a corresponding ${pref}select"
540 # no as part supplied at all - try to deduce (unless explicit end of named selection is declared)
541 # if any @$as has been supplied we assume the user knows what (s)he is doing
542 # and blindly keep stacking up pieces
543 unless ($attrs->{_dark_selector}) {
546 if ( ref $_ eq 'HASH' and exists $_->{-as} ) {
547 push @$as, $_->{-as};
549 # assume any plain no-space, no-parenthesis string to be a column spec
550 # FIXME - this is retarded but is necessary to support shit like 'count(foo)'
551 elsif ( ! ref $_ and $_ =~ /^ [^\s\(\)]+ $/x) {
554 # if all else fails - raise a flag that no more aliasing will be allowed
556 $attrs->{_dark_selector} = {
558 string => ($dark_sel_dumper ||= do {
559 require Data::Dumper::Concise;
560 Data::Dumper::Concise::DumperObject()->Indent(0);
561 })->Values([$_])->Dump
569 elsif (@$as < @$sel) {
570 $self->throw_exception(
571 "Unable to handle an ${pref}as specification (@$as) with less elements than the corresponding ${pref}select"
574 elsif ($pref and $attrs->{_dark_selector}) {
575 $self->throw_exception(
576 "Unable to process named '+select', resultset contains an unnamed selector $attrs->{_dark_selector}{string}"
582 $attrs->{"${pref}select"} = $self->_merge_attr($attrs->{"${pref}select"}, $sel);
583 $attrs->{"${pref}as"} = $self->_merge_attr($attrs->{"${pref}as"}, $as);
588 my ($self, $left, $right) = @_;
591 (ref $_ eq 'ARRAY' and !@$_)
593 (ref $_ eq 'HASH' and ! keys %$_)
594 ) and $_ = undef for ($left, $right);
596 # either on of the two undef or both undef
597 if ( ( (defined $left) xor (defined $right) ) or ! defined $left ) {
598 return defined $left ? $left : $right;
601 my $cond = $self->result_source->schema->storage->_collapse_cond({ -and => [$left, $right] });
603 for my $c (grep { ref $cond->{$_} eq 'ARRAY' and ($cond->{$_}[0]||'') eq '-and' } keys %$cond) {
605 my @vals = sort @{$cond->{$c}}[ 1..$#{$cond->{$c}} ];
606 my @fin = shift @vals;
609 push @fin, $v unless Data::Compare::Compare( $fin[-1], $v );
612 $cond->{$c} = (@fin == 1) ? $fin[0] : [-and => @fin ];
618 =head2 search_literal
620 B<CAVEAT>: C<search_literal> is provided for Class::DBI compatibility and
621 should only be used in that context. C<search_literal> is a convenience
622 method. It is equivalent to calling C<< $schema->search(\[]) >>, but if you
623 want to ensure columns are bound correctly, use L</search>.
625 See L<DBIx::Class::Manual::Cookbook/SEARCHING> and
626 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
627 require C<search_literal>.
631 =item Arguments: $sql_fragment, @standalone_bind_values
633 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
637 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
638 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
640 Pass a literal chunk of SQL to be added to the conditional part of the
643 Example of how to use C<search> instead of C<search_literal>
645 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
646 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
651 my ($self, $sql, @bind) = @_;
653 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
656 return $self->search(\[ $sql, map [ {} => $_ ], @bind ], ($attr || () ));
663 =item Arguments: \%columns_values | @pk_values, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
665 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
669 Finds and returns a single row based on supplied criteria. Takes either a
670 hashref with the same format as L</create> (including inference of foreign
671 keys from related objects), or a list of primary key values in the same
672 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
673 declaration on the L</result_source>.
675 In either case an attempt is made to combine conditions already existing on
676 the resultset with the condition passed to this method.
678 To aid with preparing the correct query for the storage you may supply the
679 C<key> attribute, which is the name of a
680 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
681 unique constraint corresponding to the
682 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
683 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
684 to construct a query that satisfies the named unique constraint fully (
685 non-NULL values for each column member of the constraint) an exception is
688 If no C<key> is specified, the search is carried over all unique constraints
689 which are fully defined by the available condition.
691 If no such constraint is found, C<find> currently defaults to a simple
692 C<< search->(\%column_values) >> which may or may not do what you expect.
693 Note that this fallback behavior may be deprecated in further versions. If
694 you need to search with arbitrary conditions - use L</search>. If the query
695 resulting from this fallback produces more than one row, a warning to the
696 effect is issued, though only the first row is constructed and returned as
699 In addition to C<key>, L</find> recognizes and applies standard
700 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
702 Note that if you have extra concerns about the correctness of the resulting
703 query you need to specify the C<key> attribute and supply the entire condition
704 as an argument to find (since it is not always possible to perform the
705 combination of the resultset condition with the supplied one, especially if
706 the resultset condition contains literal sql).
708 For example, to find a row by its primary key:
710 my $cd = $schema->resultset('CD')->find(5);
712 You can also find a row by a specific unique constraint:
714 my $cd = $schema->resultset('CD')->find(
716 artist => 'Massive Attack',
717 title => 'Mezzanine',
719 { key => 'cd_artist_title' }
722 See also L</find_or_create> and L</update_or_create>.
728 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
730 my $rsrc = $self->result_source;
733 if (exists $attrs->{key}) {
734 $constraint_name = defined $attrs->{key}
736 : $self->throw_exception("An undefined 'key' resultset attribute makes no sense")
740 # Parse out the condition from input
743 if (ref $_[0] eq 'HASH') {
744 $call_cond = { %{$_[0]} };
747 # if only values are supplied we need to default to 'primary'
748 $constraint_name = 'primary' unless defined $constraint_name;
750 my @c_cols = $rsrc->unique_constraint_columns($constraint_name);
752 $self->throw_exception(
753 "No constraint columns, maybe a malformed '$constraint_name' constraint?"
756 $self->throw_exception (
757 'find() expects either a column/value hashref, or a list of values '
758 . "corresponding to the columns of the specified unique constraint '$constraint_name'"
759 ) unless @c_cols == @_;
762 @{$call_cond}{@c_cols} = @_;
766 for my $key (keys %$call_cond) {
768 my $keyref = ref($call_cond->{$key})
770 my $relinfo = $rsrc->relationship_info($key)
772 my $val = delete $call_cond->{$key};
774 next if $keyref eq 'ARRAY'; # has_many for multi_create
776 my ($rel_cond, $crosstable) = $rsrc->_resolve_condition(
777 $relinfo->{cond}, $val, $key, $key
780 $self->throw_exception("Complex condition via relationship '$key' is unsupported in find()")
781 if $crosstable or ref($rel_cond) ne 'HASH';
784 @related{keys %$rel_cond} = values %$rel_cond;
788 # relationship conditions take precedence (?)
789 @{$call_cond}{keys %related} = values %related;
791 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
793 if (defined $constraint_name) {
794 $final_cond = $self->_qualify_cond_columns (
796 $self->_build_unique_cond (
804 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
805 # This means that we got here after a merger of relationship conditions
806 # in ::Relationship::Base::search_related (the row method), and furthermore
807 # the relationship is of the 'single' type. This means that the condition
808 # provided by the relationship (already attached to $self) is sufficient,
809 # as there can be only one row in the database that would satisfy the
813 # no key was specified - fall down to heuristics mode:
814 # run through all unique queries registered on the resultset, and
815 # 'OR' all qualifying queries together
816 my (@unique_queries, %seen_column_combinations);
817 for my $c_name ($rsrc->unique_constraint_names) {
818 next if $seen_column_combinations{
819 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
822 push @unique_queries, try {
823 $self->_build_unique_cond ($c_name, $call_cond, 'croak_on_nulls')
827 $final_cond = @unique_queries
828 ? [ map { $self->_qualify_cond_columns($_, $alias) } @unique_queries ]
829 : $self->_non_unique_find_fallback ($call_cond, $attrs)
833 # Run the query, passing the result_class since it should propagate for find
834 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
835 if ($rs->_resolved_attrs->{collapse}) {
837 carp "Query returned more than one row" if $rs->next;
845 # This is a stop-gap method as agreed during the discussion on find() cleanup:
846 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
848 # It is invoked when find() is called in legacy-mode with insufficiently-unique
849 # condition. It is provided for overrides until a saner way forward is devised
851 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
852 # the road. Please adjust your tests accordingly to catch this situation early
853 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
855 # The method will not be removed without an adequately complete replacement
856 # for strict-mode enforcement
857 sub _non_unique_find_fallback {
858 my ($self, $cond, $attrs) = @_;
860 return $self->_qualify_cond_columns(
862 exists $attrs->{alias}
864 : $self->{attrs}{alias}
869 sub _qualify_cond_columns {
870 my ($self, $cond, $alias) = @_;
872 my %aliased = %$cond;
873 for (keys %aliased) {
874 $aliased{"$alias.$_"} = delete $aliased{$_}
881 sub _build_unique_cond {
882 my ($self, $constraint_name, $extra_cond, $croak_on_null) = @_;
884 my @c_cols = $self->result_source->unique_constraint_columns($constraint_name);
886 # combination may fail if $self->{cond} is non-trivial
887 my ($final_cond) = try {
888 $self->_merge_with_rscond ($extra_cond)
893 # trim out everything not in $columns
894 $final_cond = { map {
895 exists $final_cond->{$_}
896 ? ( $_ => $final_cond->{$_} )
900 if (my @missing = grep
901 { ! ($croak_on_null ? defined $final_cond->{$_} : exists $final_cond->{$_}) }
904 $self->throw_exception( sprintf ( "Unable to satisfy requested constraint '%s', no values for column(s): %s",
906 join (', ', map { "'$_'" } @missing),
913 !$ENV{DBIC_NULLABLE_KEY_NOWARN}
915 my @undefs = sort grep { ! defined $final_cond->{$_} } (keys %$final_cond)
917 carp_unique ( sprintf (
918 "NULL/undef values supplied for requested unique constraint '%s' (NULL "
919 . 'values in column(s): %s). This is almost certainly not what you wanted, '
920 . 'though you can set DBIC_NULLABLE_KEY_NOWARN to disable this warning.',
922 join (', ', map { "'$_'" } @undefs),
929 =head2 search_related
933 =item Arguments: $rel_name, $cond?, L<\%attrs?|/ATTRIBUTES>
935 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
939 $new_rs = $cd_rs->search_related('artist', {
943 Searches the specified relationship, optionally specifying a condition and
944 attributes for matching records. See L</ATTRIBUTES> for more information.
946 In list context, C<< ->all() >> is called implicitly on the resultset, thus
947 returning a list of result objects instead. To avoid that, use L</search_related_rs>.
949 See also L</search_related_rs>.
954 return shift->related_resultset(shift)->search(@_);
957 =head2 search_related_rs
959 This method works exactly the same as search_related, except that
960 it guarantees a resultset, even in list context.
964 sub search_related_rs {
965 return shift->related_resultset(shift)->search_rs(@_);
972 =item Arguments: none
974 =item Return Value: L<$cursor|DBIx::Class::Cursor>
978 Returns a storage-driven cursor to the given resultset. See
979 L<DBIx::Class::Cursor> for more information.
986 return $self->{cursor} ||= do {
987 my $attrs = $self->_resolved_attrs;
988 $self->result_source->storage->select(
989 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
998 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1000 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1004 my $cd = $schema->resultset('CD')->single({ year => 2001 });
1006 Inflates the first result without creating a cursor if the resultset has
1007 any records in it; if not returns C<undef>. Used by L</find> as a lean version
1010 While this method can take an optional search condition (just like L</search>)
1011 being a fast-code-path it does not recognize search attributes. If you need to
1012 add extra joins or similar, call L</search> and then chain-call L</single> on the
1013 L<DBIx::Class::ResultSet> returned.
1019 As of 0.08100, this method enforces the assumption that the preceding
1020 query returns only one row. If more than one row is returned, you will receive
1023 Query returned more than one row
1025 In this case, you should be using L</next> or L</find> instead, or if you really
1026 know what you are doing, use the L</rows> attribute to explicitly limit the size
1029 This method will also throw an exception if it is called on a resultset prefetching
1030 has_many, as such a prefetch implies fetching multiple rows from the database in
1031 order to assemble the resulting object.
1038 my ($self, $where) = @_;
1040 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
1043 my $attrs = { %{$self->_resolved_attrs} };
1045 $self->throw_exception(
1046 'single() can not be used on resultsets collapsing a has_many. Use find( \%cond ) or next() instead'
1047 ) if $attrs->{collapse};
1050 if (defined $attrs->{where}) {
1053 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
1054 $where, delete $attrs->{where} ]
1057 $attrs->{where} = $where;
1061 my $data = [ $self->result_source->storage->select_single(
1062 $attrs->{from}, $attrs->{select},
1063 $attrs->{where}, $attrs
1066 return undef unless @$data;
1067 $self->{_stashed_rows} = [ $data ];
1068 $self->_construct_results->[0];
1075 =item Arguments: L<$cond?|DBIx::Class::SQLMaker>
1077 =item Return Value: L<$resultsetcolumn|DBIx::Class::ResultSetColumn>
1081 my $max_length = $rs->get_column('length')->max;
1083 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
1088 my ($self, $column) = @_;
1089 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
1097 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1099 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1103 # WHERE title LIKE '%blue%'
1104 $cd_rs = $rs->search_like({ title => '%blue%'});
1106 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
1107 that this is simply a convenience method retained for ex Class::DBI users.
1108 You most likely want to use L</search> with specific operators.
1110 For more information, see L<DBIx::Class::Manual::Cookbook>.
1112 This method is deprecated and will be removed in 0.09. Use L</search()>
1113 instead. An example conversion is:
1115 ->search_like({ foo => 'bar' });
1119 ->search({ foo => { like => 'bar' } });
1126 'search_like() is deprecated and will be removed in DBIC version 0.09.'
1127 .' Instead use ->search({ x => { -like => "y%" } })'
1128 .' (note the outer pair of {}s - they are important!)'
1130 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
1131 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
1132 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
1133 return $class->search($query, { %$attrs });
1140 =item Arguments: $first, $last
1142 =item Return Value: L<$resultset|/search> (scalar context) | L<@result_objs|DBIx::Class::Manual::ResultClass> (list context)
1146 Returns a resultset or object list representing a subset of elements from the
1147 resultset slice is called on. Indexes are from 0, i.e., to get the first
1148 three records, call:
1150 my ($one, $two, $three) = $rs->slice(0, 2);
1155 my ($self, $min, $max) = @_;
1156 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
1157 $attrs->{offset} = $self->{attrs}{offset} || 0;
1158 $attrs->{offset} += $min;
1159 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
1160 return $self->search(undef, $attrs);
1167 =item Arguments: none
1169 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1173 Returns the next element in the resultset (C<undef> is there is none).
1175 Can be used to efficiently iterate over records in the resultset:
1177 my $rs = $schema->resultset('CD')->search;
1178 while (my $cd = $rs->next) {
1182 Note that you need to store the resultset object, and call C<next> on it.
1183 Calling C<< resultset('Table')->next >> repeatedly will always return the
1184 first record from the resultset.
1191 if (my $cache = $self->get_cache) {
1192 $self->{all_cache_position} ||= 0;
1193 return $cache->[$self->{all_cache_position}++];
1196 if ($self->{attrs}{cache}) {
1197 delete $self->{pager};
1198 $self->{all_cache_position} = 1;
1199 return ($self->all)[0];
1202 return shift(@{$self->{_stashed_results}}) if @{ $self->{_stashed_results}||[] };
1204 $self->{_stashed_results} = $self->_construct_results
1207 return shift @{$self->{_stashed_results}};
1210 # Constructs as many results as it can in one pass while respecting
1211 # cursor laziness. Several modes of operation:
1213 # * Always builds everything present in @{$self->{_stashed_rows}}
1214 # * If called with $fetch_all true - pulls everything off the cursor and
1215 # builds all result structures (or objects) in one pass
1216 # * If $self->_resolved_attrs->{collapse} is true, checks the order_by
1217 # and if the resultset is ordered properly by the left side:
1218 # * Fetches stuff off the cursor until the "master object" changes,
1219 # and saves the last extra row (if any) in @{$self->{_stashed_rows}}
1221 # * Just fetches, and collapses/constructs everything as if $fetch_all
1222 # was requested (there is no other way to collapse except for an
1224 # * If no collapse is requested - just get the next row, construct and
1226 sub _construct_results {
1227 my ($self, $fetch_all) = @_;
1229 my $rsrc = $self->result_source;
1230 my $attrs = $self->_resolved_attrs;
1235 ! $attrs->{order_by}
1239 my @pcols = $rsrc->primary_columns
1241 # default order for collapsing unless the user asked for something
1242 $attrs->{order_by} = [ map { join '.', $attrs->{alias}, $_} @pcols ];
1243 $attrs->{_ordered_for_collapse} = 1;
1244 $attrs->{_order_is_artificial} = 1;
1247 # this will be used as both initial raw-row collector AND as a RV of
1248 # _construct_results. Not regrowing the array twice matters a lot...
1249 # a surprising amount actually
1250 my $rows = delete $self->{_stashed_rows};
1252 my $cursor; # we may not need one at all
1254 my $did_fetch_all = $fetch_all;
1257 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1258 $rows = [ ($rows ? @$rows : ()), $self->cursor->all ];
1260 elsif( $attrs->{collapse} ) {
1262 # a cursor will need to be closed over in case of collapse
1263 $cursor = $self->cursor;
1265 $attrs->{_ordered_for_collapse} = (
1271 ->_main_source_order_by_portion_is_stable($rsrc, $attrs->{order_by}, $attrs->{where})
1273 ) unless defined $attrs->{_ordered_for_collapse};
1275 if (! $attrs->{_ordered_for_collapse}) {
1278 # instead of looping over ->next, use ->all in stealth mode
1279 # *without* calling a ->reset afterwards
1280 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1281 if (! $cursor->{_done}) {
1282 $rows = [ ($rows ? @$rows : ()), $cursor->all ];
1283 $cursor->{_done} = 1;
1288 if (! $did_fetch_all and ! @{$rows||[]} ) {
1289 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1290 $cursor ||= $self->cursor;
1291 if (scalar (my @r = $cursor->next) ) {
1296 return undef unless @{$rows||[]};
1298 # sanity check - people are too clever for their own good
1299 if ($attrs->{collapse} and my $aliastypes = $attrs->{_last_sqlmaker_alias_map} ) {
1301 my $multiplied_selectors;
1302 for my $sel_alias ( grep { $_ ne $attrs->{alias} } keys %{ $aliastypes->{selecting} } ) {
1304 $aliastypes->{multiplying}{$sel_alias}
1306 $aliastypes->{premultiplied}{$sel_alias}
1308 $multiplied_selectors->{$_} = 1 for values %{$aliastypes->{selecting}{$sel_alias}{-seen_columns}}
1312 for my $i (0 .. $#{$attrs->{as}} ) {
1313 my $sel = $attrs->{select}[$i];
1315 if (ref $sel eq 'SCALAR') {
1318 elsif( ref $sel eq 'REF' and ref $$sel eq 'ARRAY' ) {
1322 $self->throw_exception(
1323 'Result collapse not possible - selection from a has_many source redirected to the main object'
1324 ) if ($multiplied_selectors->{$sel} and $attrs->{as}[$i] !~ /\./);
1328 # hotspot - skip the setter
1329 my $res_class = $self->_result_class;
1331 my $inflator_cref = $self->{_result_inflator}{cref} ||= do {
1332 $res_class->can ('inflate_result')
1333 or $self->throw_exception("Inflator $res_class does not provide an inflate_result() method");
1336 my $infmap = $attrs->{as};
1338 $self->{_result_inflator}{is_core_row} = ( (
1341 ( \&DBIx::Class::Row::inflate_result || die "No ::Row::inflate_result() - can't happen" )
1342 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_core_row};
1344 $self->{_result_inflator}{is_hri} = ( (
1345 ! $self->{_result_inflator}{is_core_row}
1348 require DBIx::Class::ResultClass::HashRefInflator
1350 DBIx::Class::ResultClass::HashRefInflator->can('inflate_result')
1352 ) ? 1 : 0 ) unless defined $self->{_result_inflator}{is_hri};
1355 if (! $attrs->{_related_results_construction}) {
1356 # construct a much simpler array->hash folder for the one-table cases right here
1357 if ($self->{_result_inflator}{is_hri}) {
1358 for my $r (@$rows) {
1359 $r = { map { $infmap->[$_] => $r->[$_] } 0..$#$infmap };
1362 # FIXME SUBOPTIMAL this is a very very very hot spot
1363 # while rather optimal we can *still* do much better, by
1364 # building a smarter Row::inflate_result(), and
1365 # switch to feeding it data via a much leaner interface
1367 # crude unscientific benchmarking indicated the shortcut eval is not worth it for
1368 # this particular resultset size
1369 elsif (@$rows < 60) {
1370 for my $r (@$rows) {
1371 $r = $inflator_cref->($res_class, $rsrc, { map { $infmap->[$_] => $r->[$_] } (0..$#$infmap) } );
1376 '$_ = $inflator_cref->($res_class, $rsrc, { %s }) for @$rows',
1377 join (', ', map { "\$infmap->[$_] => \$_->[$_]" } 0..$#$infmap )
1383 $self->{_result_inflator}{is_hri} ? 'hri'
1384 : $self->{_result_inflator}{is_core_row} ? 'classic_pruning'
1385 : 'classic_nonpruning'
1388 # $args and $attrs to _mk_row_parser are separated to delineate what is
1389 # core collapser stuff and what is dbic $rs specific
1390 @{$self->{_row_parser}{$parser_type}}{qw(cref nullcheck)} = $rsrc->_mk_row_parser({
1392 inflate_map => $infmap,
1393 collapse => $attrs->{collapse},
1394 premultiplied => $attrs->{_main_source_premultiplied},
1395 hri_style => $self->{_result_inflator}{is_hri},
1396 prune_null_branches => $self->{_result_inflator}{is_hri} || $self->{_result_inflator}{is_core_row},
1397 }, $attrs) unless $self->{_row_parser}{$parser_type}{cref};
1399 # column_info metadata historically hasn't been too reliable.
1400 # We need to start fixing this somehow (the collapse resolver
1401 # can't work without it). Add an explicit check for the *main*
1402 # result, hopefully this will gradually weed out such errors
1404 # FIXME - this is a temporary kludge that reduces performance
1405 # It is however necessary for the time being
1406 my ($unrolled_non_null_cols_to_check, $err);
1408 if (my $check_non_null_cols = $self->{_row_parser}{$parser_type}{nullcheck} ) {
1411 'Collapse aborted due to invalid ResultSource metadata - the following '
1412 . 'selections are declared non-nullable but NULLs were retrieved: '
1416 COL: for my $i (@$check_non_null_cols) {
1417 ! defined $_->[$i] and push @violating_idx, $i and next COL for @$rows;
1420 $self->throw_exception( $err . join (', ', map { "'$infmap->[$_]'" } @violating_idx ) )
1423 $unrolled_non_null_cols_to_check = join (',', @$check_non_null_cols);
1427 ($did_fetch_all or ! $attrs->{collapse}) ? undef
1428 : defined $unrolled_non_null_cols_to_check ? eval sprintf <<'EOS', $unrolled_non_null_cols_to_check
1430 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1431 my @r = $cursor->next or return;
1432 if (my @violating_idx = grep { ! defined $r[$_] } (%s) ) {
1433 $self->throw_exception( $err . join (', ', map { "'$infmap->[$_]'" } @violating_idx ) )
1439 # FIXME SUBOPTIMAL - we can do better, cursor->next/all (well diff. methods) should return a ref
1440 my @r = $cursor->next or return;
1445 $self->{_row_parser}{$parser_type}{cref}->(
1447 $next_cref ? ( $next_cref, $self->{_stashed_rows} = [] ) : (),
1450 # Special-case multi-object HRI - there is no $inflator_cref pass
1451 unless ($self->{_result_inflator}{is_hri}) {
1452 $_ = $inflator_cref->($res_class, $rsrc, @$_) for @$rows
1456 # The @$rows check seems odd at first - why wouldn't we want to warn
1457 # regardless? The issue is things like find() etc, where the user
1458 # *knows* only one result will come back. In these cases the ->all
1459 # is not a pessimization, but rather something we actually want
1461 'Unable to properly collapse has_many results in iterator mode due '
1462 . 'to order criteria - performed an eager cursor slurp underneath. '
1463 . 'Consider using ->all() instead'
1464 ) if ( ! $fetch_all and @$rows > 1 );
1469 =head2 result_source
1473 =item Arguments: L<$result_source?|DBIx::Class::ResultSource>
1475 =item Return Value: L<$result_source|DBIx::Class::ResultSource>
1479 An accessor for the primary ResultSource object from which this ResultSet
1486 =item Arguments: $result_class?
1488 =item Return Value: $result_class
1492 An accessor for the class to use when creating result objects. Defaults to
1493 C<< result_source->result_class >> - which in most cases is the name of the
1494 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1496 Note that changing the result_class will also remove any components
1497 that were originally loaded in the source class via
1498 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1499 in the original source class will not run.
1504 my ($self, $result_class) = @_;
1505 if ($result_class) {
1507 # don't fire this for an object
1508 $self->ensure_class_loaded($result_class)
1509 unless ref($result_class);
1511 if ($self->get_cache) {
1512 carp_unique('Changing the result_class of a ResultSet instance with cached results is a noop - the cache contents will not be altered');
1514 # FIXME ENCAPSULATION - encapsulation breach, cursor method additions pending
1515 elsif ($self->{cursor} && $self->{cursor}{_pos}) {
1516 $self->throw_exception('Changing the result_class of a ResultSet instance with an active cursor is not supported');
1519 $self->_result_class($result_class);
1521 delete $self->{_result_inflator};
1523 $self->_result_class;
1530 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1532 =item Return Value: $count
1536 Performs an SQL C<COUNT> with the same query as the resultset was built
1537 with to find the number of elements. Passing arguments is equivalent to
1538 C<< $rs->search ($cond, \%attrs)->count >>
1544 return $self->search(@_)->count if @_ and defined $_[0];
1545 return scalar @{ $self->get_cache } if $self->get_cache;
1547 my $attrs = { %{ $self->_resolved_attrs } };
1549 # this is a little optimization - it is faster to do the limit
1550 # adjustments in software, instead of a subquery
1551 my ($rows, $offset) = delete @{$attrs}{qw/rows offset/};
1554 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1555 $crs = $self->_count_subq_rs ($attrs);
1558 $crs = $self->_count_rs ($attrs);
1560 my $count = $crs->next;
1562 $count -= $offset if $offset;
1563 $count = $rows if $rows and $rows < $count;
1564 $count = 0 if ($count < 0);
1573 =item Arguments: L<$cond|DBIx::Class::SQLMaker>, L<\%attrs?|/ATTRIBUTES>
1575 =item Return Value: L<$count_rs|DBIx::Class::ResultSetColumn>
1579 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1580 This can be very handy for subqueries:
1582 ->search( { amount => $some_rs->count_rs->as_query } )
1584 As with regular resultsets the SQL query will be executed only after
1585 the resultset is accessed via L</next> or L</all>. That would return
1586 the same single value obtainable via L</count>.
1592 return $self->search(@_)->count_rs if @_;
1594 # this may look like a lack of abstraction (count() does about the same)
1595 # but in fact an _rs *must* use a subquery for the limits, as the
1596 # software based limiting can not be ported if this $rs is to be used
1597 # in a subquery itself (i.e. ->as_query)
1598 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1599 return $self->_count_subq_rs($self->{_attrs});
1602 return $self->_count_rs($self->{_attrs});
1607 # returns a ResultSetColumn object tied to the count query
1610 my ($self, $attrs) = @_;
1612 my $rsrc = $self->result_source;
1614 my $tmp_attrs = { %$attrs };
1615 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1616 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1618 # overwrite the selector (supplied by the storage)
1619 $rsrc->resultset_class->new($rsrc, {
1621 select => $rsrc->storage->_count_select ($rsrc, $attrs),
1623 })->get_column ('count');
1627 # same as above but uses a subquery
1629 sub _count_subq_rs {
1630 my ($self, $attrs) = @_;
1632 my $rsrc = $self->result_source;
1634 my $sub_attrs = { %$attrs };
1635 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1636 delete @{$sub_attrs}{qw/collapse columns as select order_by for/};
1638 # if we multi-prefetch we group_by something unique, as this is what we would
1639 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1640 if ( $attrs->{collapse} ) {
1641 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } @{
1642 $rsrc->_identifying_column_set || $self->throw_exception(
1643 'Unable to construct a unique group_by criteria properly collapsing the '
1644 . 'has_many prefetch before count()'
1649 # Calculate subquery selector
1650 if (my $g = $sub_attrs->{group_by}) {
1652 my $sql_maker = $rsrc->storage->sql_maker;
1654 # necessary as the group_by may refer to aliased functions
1656 for my $sel (@{$attrs->{select}}) {
1657 $sel_index->{$sel->{-as}} = $sel
1658 if (ref $sel eq 'HASH' and $sel->{-as});
1661 # anything from the original select mentioned on the group-by needs to make it to the inner selector
1662 # also look for named aggregates referred in the having clause
1663 # having often contains scalarrefs - thus parse it out entirely
1665 if ($attrs->{having}) {
1666 local $sql_maker->{having_bind};
1667 local $sql_maker->{quote_char} = $sql_maker->{quote_char};
1668 local $sql_maker->{name_sep} = $sql_maker->{name_sep};
1669 unless (defined $sql_maker->{quote_char} and length $sql_maker->{quote_char}) {
1670 $sql_maker->{quote_char} = [ "\x00", "\xFF" ];
1671 # if we don't unset it we screw up retarded but unfortunately working
1672 # 'MAX(foo.bar)' => { '>', 3 }
1673 $sql_maker->{name_sep} = '';
1676 my ($lquote, $rquote, $sep) = map { quotemeta $_ } ($sql_maker->_quote_chars, $sql_maker->name_sep);
1678 my $having_sql = $sql_maker->_parse_rs_attrs ({ having => $attrs->{having} });
1681 # search for both a proper quoted qualified string, for a naive unquoted scalarref
1682 # and if all fails for an utterly naive quoted scalar-with-function
1683 while ($having_sql =~ /
1684 $rquote $sep $lquote (.+?) $rquote
1686 [\s,] \w+ \. (\w+) [\s,]
1688 [\s,] $lquote (.+?) $rquote [\s,]
1690 my $part = $1 || $2 || $3; # one of them matched if we got here
1691 unless ($seen_having{$part}++) {
1698 my $colpiece = $sel_index->{$_} || $_;
1700 # unqualify join-based group_by's. Arcane but possible query
1701 # also horrible horrible hack to alias a column (not a func.)
1702 # (probably need to introduce SQLA syntax)
1703 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1706 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1708 push @{$sub_attrs->{select}}, $colpiece;
1712 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1713 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1716 return $rsrc->resultset_class
1717 ->new ($rsrc, $sub_attrs)
1719 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1720 ->get_column ('count');
1724 =head2 count_literal
1726 B<CAVEAT>: C<count_literal> is provided for Class::DBI compatibility and
1727 should only be used in that context. See L</search_literal> for further info.
1731 =item Arguments: $sql_fragment, @standalone_bind_values
1733 =item Return Value: $count
1737 Counts the results in a literal query. Equivalent to calling L</search_literal>
1738 with the passed arguments, then L</count>.
1742 sub count_literal { shift->search_literal(@_)->count; }
1748 =item Arguments: none
1750 =item Return Value: L<@result_objs|DBIx::Class::Manual::ResultClass>
1754 Returns all elements in the resultset.
1761 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1764 delete @{$self}{qw/_stashed_rows _stashed_results/};
1766 if (my $c = $self->get_cache) {
1770 $self->cursor->reset;
1772 my $objs = $self->_construct_results('fetch_all') || [];
1774 $self->set_cache($objs) if $self->{attrs}{cache};
1783 =item Arguments: none
1785 =item Return Value: $self
1789 Resets the resultset's cursor, so you can iterate through the elements again.
1790 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1798 delete @{$self}{qw/_stashed_rows _stashed_results/};
1799 $self->{all_cache_position} = 0;
1800 $self->cursor->reset;
1808 =item Arguments: none
1810 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass> | undef
1814 L<Resets|/reset> the resultset (causing a fresh query to storage) and returns
1815 an object for the first result (or C<undef> if the resultset is empty).
1820 return $_[0]->reset->next;
1826 # Determines whether and what type of subquery is required for the $rs operation.
1827 # If grouping is necessary either supplies its own, or verifies the current one
1828 # After all is done delegates to the proper storage method.
1830 sub _rs_update_delete {
1831 my ($self, $op, $values) = @_;
1833 my $rsrc = $self->result_source;
1834 my $storage = $rsrc->schema->storage;
1836 my $attrs = { %{$self->_resolved_attrs} };
1838 my $join_classifications;
1839 my ($existing_group_by) = delete @{$attrs}{qw(group_by _grouped_by_distinct)};
1841 # do we need a subquery for any reason?
1843 defined $existing_group_by
1845 # if {from} is unparseable wrap a subq
1846 ref($attrs->{from}) ne 'ARRAY'
1848 # limits call for a subq
1849 $self->_has_resolved_attr(qw/rows offset/)
1852 # simplify the joinmap, so we can further decide if a subq is necessary
1853 if (!$needs_subq and @{$attrs->{from}} > 1) {
1855 ($attrs->{from}, $join_classifications) =
1856 $storage->_prune_unused_joins ($attrs);
1858 # any non-pruneable non-local restricting joins imply subq
1859 $needs_subq = defined List::Util::first { $_ ne $attrs->{alias} } keys %{ $join_classifications->{restricting} || {} };
1862 # check if the head is composite (by now all joins are thrown out unless $needs_subq)
1864 (ref $attrs->{from}[0]) ne 'HASH'
1866 ref $attrs->{from}[0]{ $attrs->{from}[0]{-alias} }
1870 # do we need anything like a subquery?
1871 if (! $needs_subq) {
1872 # Most databases do not allow aliasing of tables in UPDATE/DELETE. Thus
1873 # a condition containing 'me' or other table prefixes will not work
1874 # at all. Tell SQLMaker to dequalify idents via a gross hack.
1876 my $sqla = $rsrc->storage->sql_maker;
1877 local $sqla->{_dequalify_idents} = 1;
1878 \[ $sqla->_recurse_where($self->{cond}) ];
1882 # we got this far - means it is time to wrap a subquery
1883 my $idcols = $rsrc->_identifying_column_set || $self->throw_exception(
1885 "Unable to perform complex resultset %s() without an identifying set of columns on source '%s'",
1891 # make a new $rs selecting only the PKs (that's all we really need for the subq)
1892 delete $attrs->{$_} for qw/select as collapse/;
1893 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } @$idcols ];
1895 # this will be consumed by the pruner waaaaay down the stack
1896 $attrs->{_force_prune_multiplying_joins} = 1;
1898 my $subrs = (ref $self)->new($rsrc, $attrs);
1900 if (@$idcols == 1) {
1901 $cond = { $idcols->[0] => { -in => $subrs->as_query } };
1903 elsif ($storage->_use_multicolumn_in) {
1904 # no syntax for calling this properly yet
1905 # !!! EXPERIMENTAL API !!! WILL CHANGE !!!
1906 $cond = $storage->sql_maker->_where_op_multicolumn_in (
1907 $idcols, # how do I convey a list of idents...? can binds reside on lhs?
1912 # if all else fails - get all primary keys and operate over a ORed set
1913 # wrap in a transaction for consistency
1914 # this is where the group_by/multiplication starts to matter
1918 # we do not need to check pre-multipliers, since if the premulti is there, its
1919 # parent (who is multi) will be there too
1920 keys %{ $join_classifications->{multiplying} || {} }
1922 # make sure if there is a supplied group_by it matches the columns compiled above
1923 # perfectly. Anything else can not be sanely executed on most databases so croak
1924 # right then and there
1925 if ($existing_group_by) {
1926 my @current_group_by = map
1927 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1932 join ("\x00", sort @current_group_by)
1934 join ("\x00", sort @{$attrs->{columns}} )
1936 $self->throw_exception (
1937 "You have just attempted a $op operation on a resultset which does group_by"
1938 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1939 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1940 . ' kind of queries. Please retry the operation with a modified group_by or'
1941 . ' without using one at all.'
1946 $subrs = $subrs->search({}, { group_by => $attrs->{columns} });
1949 $guard = $storage->txn_scope_guard;
1951 for my $row ($subrs->cursor->all) {
1953 { $idcols->[$_] => $row->[$_] }
1960 my $res = $cond ? $storage->$op (
1962 $op eq 'update' ? $values : (),
1966 $guard->commit if $guard;
1975 =item Arguments: \%values
1977 =item Return Value: $underlying_storage_rv
1981 Sets the specified columns in the resultset to the supplied values in a
1982 single query. Note that this will not run any accessor/set_column/update
1983 triggers, nor will it update any result object instances derived from this
1984 resultset (this includes the contents of the L<resultset cache|/set_cache>
1985 if any). See L</update_all> if you need to execute any on-update
1986 triggers or cascades defined either by you or a
1987 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
1989 The return value is a pass through of what the underlying
1990 storage backend returned, and may vary. See L<DBI/execute> for the most
1995 Note that L</update> does not process/deflate any of the values passed in.
1996 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
1997 ensure manually that any value passed to this method will stringify to
1998 something the RDBMS knows how to deal with. A notable example is the
1999 handling of L<DateTime> objects, for more info see:
2000 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
2005 my ($self, $values) = @_;
2006 $self->throw_exception('Values for update must be a hash')
2007 unless ref $values eq 'HASH';
2009 return $self->_rs_update_delete ('update', $values);
2016 =item Arguments: \%values
2018 =item Return Value: 1
2022 Fetches all objects and updates them one at a time via
2023 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
2024 triggers, while L</update> will not.
2029 my ($self, $values) = @_;
2030 $self->throw_exception('Values for update_all must be a hash')
2031 unless ref $values eq 'HASH';
2033 my $guard = $self->result_source->schema->txn_scope_guard;
2034 $_->update({%$values}) for $self->all; # shallow copy - update will mangle it
2043 =item Arguments: none
2045 =item Return Value: $underlying_storage_rv
2049 Deletes the rows matching this resultset in a single query. Note that this
2050 will not run any delete triggers, nor will it alter the
2051 L<in_storage|DBIx::Class::Row/in_storage> status of any result object instances
2052 derived from this resultset (this includes the contents of the
2053 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
2054 execute any on-delete triggers or cascades defined either by you or a
2055 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
2057 The return value is a pass through of what the underlying storage backend
2058 returned, and may vary. See L<DBI/execute> for the most common case.
2064 $self->throw_exception('delete does not accept any arguments')
2067 return $self->_rs_update_delete ('delete');
2074 =item Arguments: none
2076 =item Return Value: 1
2080 Fetches all objects and deletes them one at a time via
2081 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
2082 triggers, while L</delete> will not.
2088 $self->throw_exception('delete_all does not accept any arguments')
2091 my $guard = $self->result_source->schema->txn_scope_guard;
2092 $_->delete for $self->all;
2101 =item Arguments: [ \@column_list, \@row_values+ ] | [ \%col_data+ ]
2103 =item Return Value: L<\@result_objects|DBIx::Class::Manual::ResultClass> (scalar context) | L<@result_objects|DBIx::Class::Manual::ResultClass> (list context)
2107 Accepts either an arrayref of hashrefs or alternatively an arrayref of
2114 The context of this method call has an important effect on what is
2115 submitted to storage. In void context data is fed directly to fastpath
2116 insertion routines provided by the underlying storage (most often
2117 L<DBI/execute_for_fetch>), bypassing the L<new|DBIx::Class::Row/new> and
2118 L<insert|DBIx::Class::Row/insert> calls on the
2119 L<Result|DBIx::Class::Manual::ResultClass> class, including any
2120 augmentation of these methods provided by components. For example if you
2121 are using something like L<DBIx::Class::UUIDColumns> to create primary
2122 keys for you, you will find that your PKs are empty. In this case you
2123 will have to explicitly force scalar or list context in order to create
2128 In non-void (scalar or list) context, this method is simply a wrapper
2129 for L</create>. Depending on list or scalar context either a list of
2130 L<Result|DBIx::Class::Manual::ResultClass> objects or an arrayref
2131 containing these objects is returned.
2133 When supplying data in "arrayref of arrayrefs" invocation style, the
2134 first element should be a list of column names and each subsequent
2135 element should be a data value in the earlier specified column order.
2138 $schema->resultset("Artist")->populate([
2139 [ qw( artistid name ) ],
2140 [ 100, 'A Formally Unknown Singer' ],
2141 [ 101, 'A singer that jumped the shark two albums ago' ],
2142 [ 102, 'An actually cool singer' ],
2145 For the arrayref of hashrefs style each hashref should be a structure
2146 suitable for passing to L</create>. Multi-create is also permitted with
2149 $schema->resultset("Artist")->populate([
2150 { artistid => 4, name => 'Manufactured Crap', cds => [
2151 { title => 'My First CD', year => 2006 },
2152 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2155 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
2156 { title => 'My parents sold me to a record company', year => 2005 },
2157 { title => 'Why Am I So Ugly?', year => 2006 },
2158 { title => 'I Got Surgery and am now Popular', year => 2007 }
2163 If you attempt a void-context multi-create as in the example above (each
2164 Artist also has the related list of CDs), and B<do not> supply the
2165 necessary autoinc foreign key information, this method will proxy to the
2166 less efficient L</create>, and then throw the Result objects away. In this
2167 case there are obviously no benefits to using this method over L</create>.
2174 # cruft placed in standalone method
2175 my $data = $self->_normalize_populate_args(@_);
2177 return unless @$data;
2179 if(defined wantarray) {
2180 my @created = map { $self->create($_) } @$data;
2181 return wantarray ? @created : \@created;
2184 my $first = $data->[0];
2186 # if a column is a registered relationship, and is a non-blessed hash/array, consider
2187 # it relationship data
2188 my (@rels, @columns);
2189 my $rsrc = $self->result_source;
2190 my $rels = { map { $_ => $rsrc->relationship_info($_) } $rsrc->relationships };
2191 for (keys %$first) {
2192 my $ref = ref $first->{$_};
2193 $rels->{$_} && ($ref eq 'ARRAY' or $ref eq 'HASH')
2199 my @pks = $rsrc->primary_columns;
2201 ## do the belongs_to relationships
2202 foreach my $index (0..$#$data) {
2204 # delegate to create() for any dataset without primary keys with specified relationships
2205 if (grep { !defined $data->[$index]->{$_} } @pks ) {
2207 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
2208 my @ret = $self->populate($data);
2214 foreach my $rel (@rels) {
2215 next unless ref $data->[$index]->{$rel} eq "HASH";
2216 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
2217 my (undef, $reverse_relinfo) = %{$rsrc->reverse_relationship_info($rel)};
2218 my $related = $result->result_source->_resolve_condition(
2219 $reverse_relinfo->{cond},
2225 delete $data->[$index]->{$rel};
2226 $data->[$index] = {%{$data->[$index]}, %$related};
2228 push @columns, keys %$related if $index == 0;
2232 ## inherit the data locked in the conditions of the resultset
2233 my ($rs_data) = $self->_merge_with_rscond({});
2234 delete @{$rs_data}{@columns};
2236 ## do bulk insert on current row
2237 $rsrc->storage->insert_bulk(
2239 [@columns, keys %$rs_data],
2240 [ map { [ @$_{@columns}, values %$rs_data ] } @$data ],
2243 ## do the has_many relationships
2244 foreach my $item (@$data) {
2248 foreach my $rel (@rels) {
2249 next unless ref $item->{$rel} eq "ARRAY" && @{ $item->{$rel} };
2251 $main_row ||= $self->new_result({map { $_ => $item->{$_} } @pks});
2253 my $child = $main_row->$rel;
2255 my $related = $child->result_source->_resolve_condition(
2256 $rels->{$rel}{cond},
2262 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
2263 my @populate = map { {%$_, %$related} } @rows_to_add;
2265 $child->populate( \@populate );
2272 # populate() arguments went over several incarnations
2273 # What we ultimately support is AoH
2274 sub _normalize_populate_args {
2275 my ($self, $arg) = @_;
2277 if (ref $arg eq 'ARRAY') {
2281 elsif (ref $arg->[0] eq 'HASH') {
2284 elsif (ref $arg->[0] eq 'ARRAY') {
2286 my @colnames = @{$arg->[0]};
2287 foreach my $values (@{$arg}[1 .. $#$arg]) {
2288 push @ret, { map { $colnames[$_] => $values->[$_] } (0 .. $#colnames) };
2294 $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
2301 =item Arguments: none
2303 =item Return Value: L<$pager|Data::Page>
2307 Returns a L<Data::Page> object for the current resultset. Only makes
2308 sense for queries with a C<page> attribute.
2310 To get the full count of entries for a paged resultset, call
2311 C<total_entries> on the L<Data::Page> object.
2318 return $self->{pager} if $self->{pager};
2320 my $attrs = $self->{attrs};
2321 if (!defined $attrs->{page}) {
2322 $self->throw_exception("Can't create pager for non-paged rs");
2324 elsif ($attrs->{page} <= 0) {
2325 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
2327 $attrs->{rows} ||= 10;
2329 # throw away the paging flags and re-run the count (possibly
2330 # with a subselect) to get the real total count
2331 my $count_attrs = { %$attrs };
2332 delete @{$count_attrs}{qw/rows offset page pager/};
2334 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
2336 require DBIx::Class::ResultSet::Pager;
2337 return $self->{pager} = DBIx::Class::ResultSet::Pager->new(
2338 sub { $total_rs->count }, #lazy-get the total
2340 $self->{attrs}{page},
2348 =item Arguments: $page_number
2350 =item Return Value: L<$resultset|/search>
2354 Returns a resultset for the $page_number page of the resultset on which page
2355 is called, where each page contains a number of rows equal to the 'rows'
2356 attribute set on the resultset (10 by default).
2361 my ($self, $page) = @_;
2362 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2369 =item Arguments: \%col_data
2371 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2375 Creates a new result object in the resultset's result class and returns
2376 it. The row is not inserted into the database at this point, call
2377 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2378 will tell you whether the result object has been inserted or not.
2380 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2385 my ($self, $values) = @_;
2387 $self->throw_exception( "new_result takes only one argument - a hashref of values" )
2390 $self->throw_exception( "new_result expects a hashref" )
2391 unless (ref $values eq 'HASH');
2393 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2395 my $new = $self->result_class->new({
2397 ( @$cols_from_relations
2398 ? (-cols_from_relations => $cols_from_relations)
2401 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2405 reftype($new) eq 'HASH'
2411 carp_unique (sprintf (
2412 "%s->new returned a blessed empty hashref - a strong indicator something is wrong with its inheritance chain",
2413 $self->result_class,
2420 # _merge_with_rscond
2422 # Takes a simple hash of K/V data and returns its copy merged with the
2423 # condition already present on the resultset. Additionally returns an
2424 # arrayref of value/condition names, which were inferred from related
2425 # objects (this is needed for in-memory related objects)
2426 sub _merge_with_rscond {
2427 my ($self, $data) = @_;
2429 my (%new_data, @cols_from_relations);
2431 my $alias = $self->{attrs}{alias};
2433 if (! defined $self->{cond}) {
2434 # just massage $data below
2436 elsif ($self->{cond} eq $DBIx::Class::ResultSource::UNRESOLVABLE_CONDITION) {
2437 %new_data = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
2438 @cols_from_relations = keys %new_data;
2440 elsif (ref $self->{cond} ne 'HASH') {
2441 $self->throw_exception(
2442 "Can't abstract implicit construct, resultset condition not a hash"
2446 if ($self->{cond}) {
2447 my $implied = $self->_remove_alias(
2448 $self->result_source->schema->storage->_collapse_cond($self->{cond}),
2452 for my $c (keys %$implied) {
2453 my $v = $implied->{$c};
2454 if ( ! length ref $v or is_plain_value($v) ) {
2458 ref $v eq 'HASH' and keys %$v == 1 and exists $v->{'='} and is_literal_value($v->{'='})
2460 $new_data{$c} = $v->{'='};
2466 # precedence must be given to passed values over values inherited from
2467 # the cond, so the order here is important.
2470 %{ $self->_remove_alias($data, $alias) },
2473 return (\%new_data, \@cols_from_relations);
2476 # _has_resolved_attr
2478 # determines if the resultset defines at least one
2479 # of the attributes supplied
2481 # used to determine if a subquery is necessary
2483 # supports some virtual attributes:
2485 # This will scan for any joins being present on the resultset.
2486 # It is not a mere key-search but a deep inspection of {from}
2489 sub _has_resolved_attr {
2490 my ($self, @attr_names) = @_;
2492 my $attrs = $self->_resolved_attrs;
2496 for my $n (@attr_names) {
2497 if (grep { $n eq $_ } (qw/-join/) ) {
2498 $extra_checks{$n}++;
2502 my $attr = $attrs->{$n};
2504 next if not defined $attr;
2506 if (ref $attr eq 'HASH') {
2507 return 1 if keys %$attr;
2509 elsif (ref $attr eq 'ARRAY') {
2517 # a resolved join is expressed as a multi-level from
2519 $extra_checks{-join}
2521 ref $attrs->{from} eq 'ARRAY'
2523 @{$attrs->{from}} > 1
2531 # Remove the specified alias from the specified query hash. A copy is made so
2532 # the original query is not modified.
2535 my ($self, $query, $alias) = @_;
2537 my %orig = %{ $query || {} };
2540 foreach my $key (keys %orig) {
2542 $unaliased{$key} = $orig{$key};
2545 $unaliased{$1} = $orig{$key}
2546 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2556 =item Arguments: none
2558 =item Return Value: \[ $sql, L<@bind_values|/DBIC BIND VALUES> ]
2562 Returns the SQL query and bind vars associated with the invocant.
2564 This is generally used as the RHS for a subquery.
2571 my $attrs = { %{ $self->_resolved_attrs } };
2573 my $aq = $self->result_source->storage->_select_args_to_query (
2574 $attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs
2584 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2586 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2590 my $artist = $schema->resultset('Artist')->find_or_new(
2591 { artist => 'fred' }, { key => 'artists' });
2593 $cd->cd_to_producer->find_or_new({ producer => $producer },
2594 { key => 'primary' });
2596 Find an existing record from this resultset using L</find>. if none exists,
2597 instantiate a new result object and return it. The object will not be saved
2598 into your storage until you call L<DBIx::Class::Row/insert> on it.
2600 You most likely want this method when looking for existing rows using a unique
2601 constraint that is not the primary key, or looking for related rows.
2603 If you want objects to be saved immediately, use L</find_or_create> instead.
2605 B<Note>: Make sure to read the documentation of L</find> and understand the
2606 significance of the C<key> attribute, as its lack may skew your search, and
2607 subsequently result in spurious new objects.
2609 B<Note>: Take care when using C<find_or_new> with a table having
2610 columns with default values that you intend to be automatically
2611 supplied by the database (e.g. an auto_increment primary key column).
2612 In normal usage, the value of such columns should NOT be included at
2613 all in the call to C<find_or_new>, even when set to C<undef>.
2619 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2620 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2621 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2624 return $self->new_result($hash);
2631 =item Arguments: \%col_data
2633 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2637 Attempt to create a single new row or a row with multiple related rows
2638 in the table represented by the resultset (and related tables). This
2639 will not check for duplicate rows before inserting, use
2640 L</find_or_create> to do that.
2642 To create one row for this resultset, pass a hashref of key/value
2643 pairs representing the columns of the table and the values you wish to
2644 store. If the appropriate relationships are set up, foreign key fields
2645 can also be passed an object representing the foreign row, and the
2646 value will be set to its primary key.
2648 To create related objects, pass a hashref of related-object column values
2649 B<keyed on the relationship name>. If the relationship is of type C<multi>
2650 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2651 The process will correctly identify columns holding foreign keys, and will
2652 transparently populate them from the keys of the corresponding relation.
2653 This can be applied recursively, and will work correctly for a structure
2654 with an arbitrary depth and width, as long as the relationships actually
2655 exists and the correct column data has been supplied.
2657 Instead of hashrefs of plain related data (key/value pairs), you may
2658 also pass new or inserted objects. New objects (not inserted yet, see
2659 L</new_result>), will be inserted into their appropriate tables.
2661 Effectively a shortcut for C<< ->new_result(\%col_data)->insert >>.
2663 Example of creating a new row.
2665 $person_rs->create({
2666 name=>"Some Person",
2667 email=>"somebody@someplace.com"
2670 Example of creating a new row and also creating rows in a related C<has_many>
2671 or C<has_one> resultset. Note Arrayref.
2674 { artistid => 4, name => 'Manufactured Crap', cds => [
2675 { title => 'My First CD', year => 2006 },
2676 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2681 Example of creating a new row and also creating a row in a related
2682 C<belongs_to> resultset. Note Hashref.
2685 title=>"Music for Silly Walks",
2688 name=>"Silly Musician",
2696 When subclassing ResultSet never attempt to override this method. Since
2697 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2698 lot of the internals simply never call it, so your override will be
2699 bypassed more often than not. Override either L<DBIx::Class::Row/new>
2700 or L<DBIx::Class::Row/insert> depending on how early in the
2701 L</create> process you need to intervene. See also warning pertaining to
2709 my ($self, $col_data) = @_;
2710 $self->throw_exception( "create needs a hashref" )
2711 unless ref $col_data eq 'HASH';
2712 return $self->new_result($col_data)->insert;
2715 =head2 find_or_create
2719 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2721 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2725 $cd->cd_to_producer->find_or_create({ producer => $producer },
2726 { key => 'primary' });
2728 Tries to find a record based on its primary key or unique constraints; if none
2729 is found, creates one and returns that instead.
2731 my $cd = $schema->resultset('CD')->find_or_create({
2733 artist => 'Massive Attack',
2734 title => 'Mezzanine',
2738 Also takes an optional C<key> attribute, to search by a specific key or unique
2739 constraint. For example:
2741 my $cd = $schema->resultset('CD')->find_or_create(
2743 artist => 'Massive Attack',
2744 title => 'Mezzanine',
2746 { key => 'cd_artist_title' }
2749 B<Note>: Make sure to read the documentation of L</find> and understand the
2750 significance of the C<key> attribute, as its lack may skew your search, and
2751 subsequently result in spurious row creation.
2753 B<Note>: Because find_or_create() reads from the database and then
2754 possibly inserts based on the result, this method is subject to a race
2755 condition. Another process could create a record in the table after
2756 the find has completed and before the create has started. To avoid
2757 this problem, use find_or_create() inside a transaction.
2759 B<Note>: Take care when using C<find_or_create> with a table having
2760 columns with default values that you intend to be automatically
2761 supplied by the database (e.g. an auto_increment primary key column).
2762 In normal usage, the value of such columns should NOT be included at
2763 all in the call to C<find_or_create>, even when set to C<undef>.
2765 See also L</find> and L</update_or_create>. For information on how to declare
2766 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2768 If you need to know if an existing row was found or a new one created use
2769 L</find_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2770 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2773 my $cd = $schema->resultset('CD')->find_or_new({
2775 artist => 'Massive Attack',
2776 title => 'Mezzanine',
2780 if( !$cd->in_storage ) {
2787 sub find_or_create {
2789 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2790 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2791 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2794 return $self->create($hash);
2797 =head2 update_or_create
2801 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2803 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2807 $resultset->update_or_create({ col => $val, ... });
2809 Like L</find_or_create>, but if a row is found it is immediately updated via
2810 C<< $found_row->update (\%col_data) >>.
2813 Takes an optional C<key> attribute to search on a specific unique constraint.
2816 # In your application
2817 my $cd = $schema->resultset('CD')->update_or_create(
2819 artist => 'Massive Attack',
2820 title => 'Mezzanine',
2823 { key => 'cd_artist_title' }
2826 $cd->cd_to_producer->update_or_create({
2827 producer => $producer,
2833 B<Note>: Make sure to read the documentation of L</find> and understand the
2834 significance of the C<key> attribute, as its lack may skew your search, and
2835 subsequently result in spurious row creation.
2837 B<Note>: Take care when using C<update_or_create> with a table having
2838 columns with default values that you intend to be automatically
2839 supplied by the database (e.g. an auto_increment primary key column).
2840 In normal usage, the value of such columns should NOT be included at
2841 all in the call to C<update_or_create>, even when set to C<undef>.
2843 See also L</find> and L</find_or_create>. For information on how to declare
2844 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2846 If you need to know if an existing row was updated or a new one created use
2847 L</update_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2848 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2853 sub update_or_create {
2855 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2856 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2858 my $row = $self->find($cond, $attrs);
2860 $row->update($cond);
2864 return $self->create($cond);
2867 =head2 update_or_new
2871 =item Arguments: \%col_data, { key => $unique_constraint, L<%attrs|/ATTRIBUTES> }?
2873 =item Return Value: L<$result|DBIx::Class::Manual::ResultClass>
2877 $resultset->update_or_new({ col => $val, ... });
2879 Like L</find_or_new> but if a row is found it is immediately updated via
2880 C<< $found_row->update (\%col_data) >>.
2884 # In your application
2885 my $cd = $schema->resultset('CD')->update_or_new(
2887 artist => 'Massive Attack',
2888 title => 'Mezzanine',
2891 { key => 'cd_artist_title' }
2894 if ($cd->in_storage) {
2895 # the cd was updated
2898 # the cd is not yet in the database, let's insert it
2902 B<Note>: Make sure to read the documentation of L</find> and understand the
2903 significance of the C<key> attribute, as its lack may skew your search, and
2904 subsequently result in spurious new objects.
2906 B<Note>: Take care when using C<update_or_new> with a table having
2907 columns with default values that you intend to be automatically
2908 supplied by the database (e.g. an auto_increment primary key column).
2909 In normal usage, the value of such columns should NOT be included at
2910 all in the call to C<update_or_new>, even when set to C<undef>.
2912 See also L</find>, L</find_or_create> and L</find_or_new>.
2918 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
2919 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2921 my $row = $self->find( $cond, $attrs );
2922 if ( defined $row ) {
2923 $row->update($cond);
2927 return $self->new_result($cond);
2934 =item Arguments: none
2936 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass> | undef
2940 Gets the contents of the cache for the resultset, if the cache is set.
2942 The cache is populated either by using the L</prefetch> attribute to
2943 L</search> or by calling L</set_cache>.
2955 =item Arguments: L<\@result_objs|DBIx::Class::Manual::ResultClass>
2957 =item Return Value: L<\@result_objs|DBIx::Class::Manual::ResultClass>
2961 Sets the contents of the cache for the resultset. Expects an arrayref
2962 of objects of the same class as those produced by the resultset. Note that
2963 if the cache is set, the resultset will return the cached objects rather
2964 than re-querying the database even if the cache attr is not set.
2966 The contents of the cache can also be populated by using the
2967 L</prefetch> attribute to L</search>.
2972 my ( $self, $data ) = @_;
2973 $self->throw_exception("set_cache requires an arrayref")
2974 if defined($data) && (ref $data ne 'ARRAY');
2975 $self->{all_cache} = $data;
2982 =item Arguments: none
2984 =item Return Value: undef
2988 Clears the cache for the resultset.
2993 shift->set_cache(undef);
3000 =item Arguments: none
3002 =item Return Value: true, if the resultset has been paginated
3010 return !!$self->{attrs}{page};
3017 =item Arguments: none
3019 =item Return Value: true, if the resultset has been ordered with C<order_by>.
3027 return scalar $self->result_source->storage->_extract_order_criteria($self->{attrs}{order_by});
3030 =head2 related_resultset
3034 =item Arguments: $rel_name
3036 =item Return Value: L<$resultset|/search>
3040 Returns a related resultset for the supplied relationship name.
3042 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
3046 sub related_resultset {
3047 my ($self, $rel) = @_;
3049 return $self->{related_resultsets}{$rel}
3050 if defined $self->{related_resultsets}{$rel};
3052 return $self->{related_resultsets}{$rel} = do {
3053 my $rsrc = $self->result_source;
3054 my $rel_info = $rsrc->relationship_info($rel);
3056 $self->throw_exception(
3057 "search_related: result source '" . $rsrc->source_name .
3058 "' has no such relationship $rel")
3061 my $attrs = $self->_chain_relationship($rel);
3063 my $join_count = $attrs->{seen_join}{$rel};
3065 my $alias = $self->result_source->storage
3066 ->relname_to_table_alias($rel, $join_count);
3068 # since this is search_related, and we already slid the select window inwards
3069 # (the select/as attrs were deleted in the beginning), we need to flip all
3070 # left joins to inner, so we get the expected results
3071 # read the comment on top of the actual function to see what this does
3072 $attrs->{from} = $rsrc->schema->storage->_inner_join_to_node ($attrs->{from}, $alias);
3075 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
3076 delete @{$attrs}{qw(result_class alias)};
3078 my $rel_source = $rsrc->related_source($rel);
3082 # The reason we do this now instead of passing the alias to the
3083 # search_rs below is that if you wrap/overload resultset on the
3084 # source you need to know what alias it's -going- to have for things
3085 # to work sanely (e.g. RestrictWithObject wants to be able to add
3086 # extra query restrictions, and these may need to be $alias.)
3088 my $rel_attrs = $rel_source->resultset_attributes;
3089 local $rel_attrs->{alias} = $alias;
3091 $rel_source->resultset
3095 where => $attrs->{where},
3099 if (my $cache = $self->get_cache) {
3100 my @related_cache = map
3101 { $_->related_resultset($rel)->get_cache || () }
3105 $new->set_cache([ map @$_, @related_cache ]) if @related_cache == @$cache;
3112 =head2 current_source_alias
3116 =item Arguments: none
3118 =item Return Value: $source_alias
3122 Returns the current table alias for the result source this resultset is built
3123 on, that will be used in the SQL query. Usually it is C<me>.
3125 Currently the source alias that refers to the result set returned by a
3126 L</search>/L</find> family method depends on how you got to the resultset: it's
3127 C<me> by default, but eg. L</search_related> aliases it to the related result
3128 source name (and keeps C<me> referring to the original result set). The long
3129 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
3130 (and make this method unnecessary).
3132 Thus it's currently necessary to use this method in predefined queries (see
3133 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
3134 source alias of the current result set:
3136 # in a result set class
3138 my ($self, $user) = @_;
3140 my $me = $self->current_source_alias;
3142 return $self->search({
3143 "$me.modified" => $user->id,
3149 sub current_source_alias {
3150 return (shift->{attrs} || {})->{alias} || 'me';
3153 =head2 as_subselect_rs
3157 =item Arguments: none
3159 =item Return Value: L<$resultset|/search>
3163 Act as a barrier to SQL symbols. The resultset provided will be made into a
3164 "virtual view" by including it as a subquery within the from clause. From this
3165 point on, any joined tables are inaccessible to ->search on the resultset (as if
3166 it were simply where-filtered without joins). For example:
3168 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
3170 # 'x' now pollutes the query namespace
3172 # So the following works as expected
3173 my $ok_rs = $rs->search({'x.other' => 1});
3175 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
3176 # def) we look for one row with contradictory terms and join in another table
3177 # (aliased 'x_2') which we never use
3178 my $broken_rs = $rs->search({'x.name' => 'def'});
3180 my $rs2 = $rs->as_subselect_rs;
3182 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
3183 my $not_joined_rs = $rs2->search({'x.other' => 1});
3185 # works as expected: finds a 'table' row related to two x rows (abc and def)
3186 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
3188 Another example of when one might use this would be to select a subset of
3189 columns in a group by clause:
3191 my $rs = $schema->resultset('Bar')->search(undef, {
3192 group_by => [qw{ id foo_id baz_id }],
3193 })->as_subselect_rs->search(undef, {
3194 columns => [qw{ id foo_id }]
3197 In the above example normally columns would have to be equal to the group by,
3198 but because we isolated the group by into a subselect the above works.
3202 sub as_subselect_rs {
3205 my $attrs = $self->_resolved_attrs;
3207 my $fresh_rs = (ref $self)->new (
3208 $self->result_source
3211 # these pieces will be locked in the subquery
3212 delete $fresh_rs->{cond};
3213 delete @{$fresh_rs->{attrs}}{qw/where bind/};
3215 return $fresh_rs->search( {}, {
3217 $attrs->{alias} => $self->as_query,
3218 -alias => $attrs->{alias},
3219 -rsrc => $self->result_source,
3221 alias => $attrs->{alias},
3225 # This code is called by search_related, and makes sure there
3226 # is clear separation between the joins before, during, and
3227 # after the relationship. This information is needed later
3228 # in order to properly resolve prefetch aliases (any alias
3229 # with a relation_chain_depth less than the depth of the
3230 # current prefetch is not considered)
3232 # The increments happen twice per join. An even number means a
3233 # relationship specified via a search_related, whereas an odd
3234 # number indicates a join/prefetch added via attributes
3236 # Also this code will wrap the current resultset (the one we
3237 # chain to) in a subselect IFF it contains limiting attributes
3238 sub _chain_relationship {
3239 my ($self, $rel) = @_;
3240 my $source = $self->result_source;
3241 my $attrs = { %{$self->{attrs}||{}} };
3243 # we need to take the prefetch the attrs into account before we
3244 # ->_resolve_join as otherwise they get lost - captainL
3245 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
3247 delete @{$attrs}{qw/join prefetch collapse group_by distinct _grouped_by_distinct select as columns +select +as +columns/};
3249 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
3252 my @force_subq_attrs = qw/offset rows group_by having/;
3255 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
3257 $self->_has_resolved_attr (@force_subq_attrs)
3259 # Nuke the prefetch (if any) before the new $rs attrs
3260 # are resolved (prefetch is useless - we are wrapping
3261 # a subquery anyway).
3262 my $rs_copy = $self->search;
3263 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
3264 $rs_copy->{attrs}{join},
3265 delete $rs_copy->{attrs}{prefetch},
3270 -alias => $attrs->{alias},
3271 $attrs->{alias} => $rs_copy->as_query,
3273 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
3274 $seen->{-relation_chain_depth} = 0;
3276 elsif ($attrs->{from}) { #shallow copy suffices
3277 $from = [ @{$attrs->{from}} ];
3282 -alias => $attrs->{alias},
3283 $attrs->{alias} => $source->from,
3287 my $jpath = ($seen->{-relation_chain_depth})
3288 ? $from->[-1][0]{-join_path}
3291 my @requested_joins = $source->_resolve_join(
3298 push @$from, @requested_joins;
3300 $seen->{-relation_chain_depth}++;
3302 # if $self already had a join/prefetch specified on it, the requested
3303 # $rel might very well be already included. What we do in this case
3304 # is effectively a no-op (except that we bump up the chain_depth on
3305 # the join in question so we could tell it *is* the search_related)
3308 # we consider the last one thus reverse
3309 for my $j (reverse @requested_joins) {
3310 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
3311 if ($rel eq $last_j) {
3312 $j->[0]{-relation_chain_depth}++;
3318 unless ($already_joined) {
3319 push @$from, $source->_resolve_join(
3327 $seen->{-relation_chain_depth}++;
3329 return {%$attrs, from => $from, seen_join => $seen};
3332 sub _resolved_attrs {
3334 return $self->{_attrs} if $self->{_attrs};
3336 my $attrs = { %{ $self->{attrs} || {} } };
3337 my $source = $self->result_source;
3338 my $alias = $attrs->{alias};
3340 $self->throw_exception("Specifying distinct => 1 in conjunction with collapse => 1 is unsupported")
3341 if $attrs->{collapse} and $attrs->{distinct};
3343 # default selection list
3344 $attrs->{columns} = [ $source->columns ]
3345 unless List::Util::first { exists $attrs->{$_} } qw/columns cols select as/;
3347 # merge selectors together
3348 for (qw/columns select as/) {
3349 $attrs->{$_} = $self->_merge_attr($attrs->{$_}, delete $attrs->{"+$_"})
3350 if $attrs->{$_} or $attrs->{"+$_"};
3353 # disassemble columns
3355 if (my $cols = delete $attrs->{columns}) {
3356 for my $c (ref $cols eq 'ARRAY' ? @$cols : $cols) {
3357 if (ref $c eq 'HASH') {
3358 for my $as (sort keys %$c) {
3359 push @sel, $c->{$as};
3370 # when trying to weed off duplicates later do not go past this point -
3371 # everything added from here on is unbalanced "anyone's guess" stuff
3372 my $dedup_stop_idx = $#as;
3374 push @as, @{ ref $attrs->{as} eq 'ARRAY' ? $attrs->{as} : [ $attrs->{as} ] }
3376 push @sel, @{ ref $attrs->{select} eq 'ARRAY' ? $attrs->{select} : [ $attrs->{select} ] }
3377 if $attrs->{select};
3379 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3380 $_ = (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" for @sel;
3382 # disqualify all $alias.col as-bits (inflate-map mandated)
3383 $_ = ($_ =~ /^\Q$alias.\E(.+)$/) ? $1 : $_ for @as;
3385 # de-duplicate the result (remove *identical* select/as pairs)
3386 # and also die on duplicate {as} pointing to different {select}s
3387 # not using a c-style for as the condition is prone to shrinkage
3390 while ($i <= $dedup_stop_idx) {
3391 if ($seen->{"$sel[$i] \x00\x00 $as[$i]"}++) {
3396 elsif ($seen->{$as[$i]}++) {
3397 $self->throw_exception(
3398 "inflate_result() alias '$as[$i]' specified twice with different SQL-side {select}-ors"
3406 $attrs->{select} = \@sel;
3407 $attrs->{as} = \@as;
3409 $attrs->{from} ||= [{
3411 -alias => $self->{attrs}{alias},
3412 $self->{attrs}{alias} => $source->from,
3415 if ( $attrs->{join} || $attrs->{prefetch} ) {
3417 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3418 if ref $attrs->{from} ne 'ARRAY';
3420 my $join = (delete $attrs->{join}) || {};
3422 if ( defined $attrs->{prefetch} ) {
3423 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3426 $attrs->{from} = # have to copy here to avoid corrupting the original
3428 @{ $attrs->{from} },
3429 $source->_resolve_join(
3432 { %{ $attrs->{seen_join} || {} } },
3433 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3434 ? $attrs->{from}[-1][0]{-join_path}
3441 if ( defined $attrs->{order_by} ) {
3442 $attrs->{order_by} = (
3443 ref( $attrs->{order_by} ) eq 'ARRAY'
3444 ? [ @{ $attrs->{order_by} } ]
3445 : [ $attrs->{order_by} || () ]
3449 if ($attrs->{group_by} and ref $attrs->{group_by} ne 'ARRAY') {
3450 $attrs->{group_by} = [ $attrs->{group_by} ];
3454 # generate selections based on the prefetch helper
3455 my ($prefetch, @prefetch_select, @prefetch_as);
3456 $prefetch = $self->_merge_joinpref_attr( {}, delete $attrs->{prefetch} )
3457 if defined $attrs->{prefetch};
3461 $self->throw_exception("Unable to prefetch, resultset contains an unnamed selector $attrs->{_dark_selector}{string}")
3462 if $attrs->{_dark_selector};
3464 $self->throw_exception("Specifying prefetch in conjunction with an explicit collapse => 0 is unsupported")
3465 if defined $attrs->{collapse} and ! $attrs->{collapse};
3467 $attrs->{collapse} = 1;
3469 # this is a separate structure (we don't look in {from} directly)
3470 # as the resolver needs to shift things off the lists to work
3471 # properly (identical-prefetches on different branches)
3473 if (ref $attrs->{from} eq 'ARRAY') {
3475 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3477 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3478 next unless $j->[0]{-alias};
3479 next unless $j->[0]{-join_path};
3480 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3482 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3485 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3486 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3490 my @prefetch = $source->_resolve_prefetch( $prefetch, $alias, $join_map );
3492 # save these for after distinct resolution
3493 @prefetch_select = map { $_->[0] } @prefetch;
3494 @prefetch_as = map { $_->[1] } @prefetch;
3497 # run through the resulting joinstructure (starting from our current slot)
3498 # and unset collapse if proven unnecessary
3500 # also while we are at it find out if the current root source has
3501 # been premultiplied by previous related_source chaining
3503 # this allows to predict whether a root object with all other relation
3504 # data set to NULL is in fact unique
3505 if ($attrs->{collapse}) {
3507 if (ref $attrs->{from} eq 'ARRAY') {
3509 if (@{$attrs->{from}} == 1) {
3510 # no joins - no collapse
3511 $attrs->{collapse} = 0;
3514 # find where our table-spec starts
3515 my @fromlist = @{$attrs->{from}};
3517 my $t = shift @fromlist;
3520 # me vs join from-spec distinction - a ref means non-root
3521 if (ref $t eq 'ARRAY') {
3523 $is_multi ||= ! $t->{-is_single};
3525 last if ($t->{-alias} && $t->{-alias} eq $alias);
3526 $attrs->{_main_source_premultiplied} ||= $is_multi;
3529 # no non-singles remaining, nor any premultiplication - nothing to collapse
3531 ! $attrs->{_main_source_premultiplied}
3533 ! List::Util::first { ! $_->[0]{-is_single} } @fromlist
3535 $attrs->{collapse} = 0;
3541 # if we can not analyze the from - err on the side of safety
3542 $attrs->{_main_source_premultiplied} = 1;
3546 # generate the distinct induced group_by before injecting the prefetched select/as parts
3547 if (delete $attrs->{distinct}) {
3548 if ($attrs->{group_by}) {
3549 carp_unique ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3552 $attrs->{_grouped_by_distinct} = 1;
3553 # distinct affects only the main selection part, not what prefetch may add below
3554 ($attrs->{group_by}, my $new_order) = $source->storage->_group_over_selection($attrs);
3556 # FIXME possibly ignore a rewritten order_by (may turn out to be an issue)
3557 # The thinking is: if we are collapsing the subquerying prefetch engine will
3558 # rip stuff apart for us anyway, and we do not want to have a potentially
3559 # function-converted external order_by
3560 # ( there is an explicit if ( collapse && _grouped_by_distinct ) check in DBIHacks )
3561 $attrs->{order_by} = $new_order unless $attrs->{collapse};
3565 # inject prefetch-bound selection (if any)
3566 push @{$attrs->{select}}, @prefetch_select;
3567 push @{$attrs->{as}}, @prefetch_as;
3569 # whether we can get away with the dumbest (possibly DBI-internal) collapser
3570 if ( List::Util::first { $_ =~ /\./ } @{$attrs->{as}} ) {
3571 $attrs->{_related_results_construction} = 1;
3574 # if both page and offset are specified, produce a combined offset
3575 # even though it doesn't make much sense, this is what pre 081xx has
3577 if (my $page = delete $attrs->{page}) {
3579 ($attrs->{rows} * ($page - 1))
3581 ($attrs->{offset} || 0)
3585 return $self->{_attrs} = $attrs;
3589 my ($self, $attr) = @_;
3591 if (ref $attr eq 'HASH') {
3592 return $self->_rollout_hash($attr);
3593 } elsif (ref $attr eq 'ARRAY') {
3594 return $self->_rollout_array($attr);
3600 sub _rollout_array {
3601 my ($self, $attr) = @_;
3604 foreach my $element (@{$attr}) {
3605 if (ref $element eq 'HASH') {
3606 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3607 } elsif (ref $element eq 'ARRAY') {
3608 # XXX - should probably recurse here
3609 push( @rolled_array, @{$self->_rollout_array($element)} );
3611 push( @rolled_array, $element );
3614 return \@rolled_array;
3618 my ($self, $attr) = @_;
3621 foreach my $key (keys %{$attr}) {
3622 push( @rolled_array, { $key => $attr->{$key} } );
3624 return \@rolled_array;
3627 sub _calculate_score {
3628 my ($self, $a, $b) = @_;
3630 if (defined $a xor defined $b) {
3633 elsif (not defined $a) {
3637 if (ref $b eq 'HASH') {
3638 my ($b_key) = keys %{$b};
3639 if (ref $a eq 'HASH') {
3640 my ($a_key) = keys %{$a};
3641 if ($a_key eq $b_key) {
3642 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3647 return ($a eq $b_key) ? 1 : 0;
3650 if (ref $a eq 'HASH') {
3651 my ($a_key) = keys %{$a};
3652 return ($b eq $a_key) ? 1 : 0;
3654 return ($b eq $a) ? 1 : 0;
3659 sub _merge_joinpref_attr {
3660 my ($self, $orig, $import) = @_;
3662 return $import unless defined($orig);
3663 return $orig unless defined($import);
3665 $orig = $self->_rollout_attr($orig);
3666 $import = $self->_rollout_attr($import);
3669 foreach my $import_element ( @{$import} ) {
3670 # find best candidate from $orig to merge $b_element into
3671 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3672 foreach my $orig_element ( @{$orig} ) {
3673 my $score = $self->_calculate_score( $orig_element, $import_element );
3674 if ($score > $best_candidate->{score}) {
3675 $best_candidate->{position} = $position;
3676 $best_candidate->{score} = $score;
3680 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3681 $import_key = '' if not defined $import_key;
3683 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3684 push( @{$orig}, $import_element );
3686 my $orig_best = $orig->[$best_candidate->{position}];
3687 # merge orig_best and b_element together and replace original with merged
3688 if (ref $orig_best ne 'HASH') {
3689 $orig->[$best_candidate->{position}] = $import_element;
3690 } elsif (ref $import_element eq 'HASH') {
3691 my ($key) = keys %{$orig_best};
3692 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3695 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3698 return @$orig ? $orig : ();
3706 require Hash::Merge;
3707 my $hm = Hash::Merge->new;
3709 $hm->specify_behavior({
3712 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
3714 if ($defl xor $defr) {
3715 return [ $defl ? $_[0] : $_[1] ];
3720 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
3724 return [$_[0], $_[1]];
3728 return $_[1] if !defined $_[0];
3729 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3730 return [$_[0], @{$_[1]}]
3733 return [] if !defined $_[0] and !keys %{$_[1]};
3734 return [ $_[1] ] if !defined $_[0];
3735 return [ $_[0] ] if !keys %{$_[1]};
3736 return [$_[0], $_[1]]
3741 return $_[0] if !defined $_[1];
3742 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3743 return [@{$_[0]}, $_[1]]
3746 my @ret = @{$_[0]} or return $_[1];
3747 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
3748 my %idx = map { $_ => 1 } @ret;
3749 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
3753 return [ $_[1] ] if ! @{$_[0]};
3754 return $_[0] if !keys %{$_[1]};
3755 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3756 return [ @{$_[0]}, $_[1] ];
3761 return [] if !keys %{$_[0]} and !defined $_[1];
3762 return [ $_[0] ] if !defined $_[1];
3763 return [ $_[1] ] if !keys %{$_[0]};
3764 return [$_[0], $_[1]]
3767 return [] if !keys %{$_[0]} and !@{$_[1]};
3768 return [ $_[0] ] if !@{$_[1]};
3769 return $_[1] if !keys %{$_[0]};
3770 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3771 return [ $_[0], @{$_[1]} ];
3774 return [] if !keys %{$_[0]} and !keys %{$_[1]};
3775 return [ $_[0] ] if !keys %{$_[1]};
3776 return [ $_[1] ] if !keys %{$_[0]};
3777 return [ $_[0] ] if $_[0] eq $_[1];
3778 return [ $_[0], $_[1] ];
3781 } => 'DBIC_RS_ATTR_MERGER');
3785 return $hm->merge ($_[1], $_[2]);
3789 sub STORABLE_freeze {
3790 my ($self, $cloning) = @_;
3791 my $to_serialize = { %$self };
3793 # A cursor in progress can't be serialized (and would make little sense anyway)
3794 # the parser can be regenerated (and can't be serialized)
3795 delete @{$to_serialize}{qw/cursor _row_parser _result_inflator/};
3797 # nor is it sensical to store a not-yet-fired-count pager
3798 if ($to_serialize->{pager} and ref $to_serialize->{pager}{total_entries} eq 'CODE') {
3799 delete $to_serialize->{pager};
3802 Storable::nfreeze($to_serialize);
3805 # need this hook for symmetry
3807 my ($self, $cloning, $serialized) = @_;
3809 %$self = %{ Storable::thaw($serialized) };
3815 =head2 throw_exception
3817 See L<DBIx::Class::Schema/throw_exception> for details.
3821 sub throw_exception {
3824 if (ref $self and my $rsrc = $self->result_source) {
3825 $rsrc->throw_exception(@_)
3828 DBIx::Class::Exception->throw(@_);
3836 # XXX: FIXME: Attributes docs need clearing up
3840 Attributes are used to refine a ResultSet in various ways when
3841 searching for data. They can be passed to any method which takes an
3842 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3845 Default attributes can be set on the result class using
3846 L<DBIx::Class::ResultSource/resultset_attributes>. (Please read
3847 the CAVEATS on that feature before using it!)
3849 These are in no particular order:
3855 =item Value: ( $order_by | \@order_by | \%order_by )
3859 Which column(s) to order the results by.
3861 [The full list of suitable values is documented in
3862 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3865 If a single column name, or an arrayref of names is supplied, the
3866 argument is passed through directly to SQL. The hashref syntax allows
3867 for connection-agnostic specification of ordering direction:
3869 For descending order:
3871 order_by => { -desc => [qw/col1 col2 col3/] }
3873 For explicit ascending order:
3875 order_by => { -asc => 'col' }
3877 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3878 supported, although you are strongly encouraged to use the hashref
3879 syntax as outlined above.
3885 =item Value: \@columns | \%columns | $column
3889 Shortcut to request a particular set of columns to be retrieved. Each
3890 column spec may be a string (a table column name), or a hash (in which
3891 case the key is the C<as> value, and the value is used as the C<select>
3892 expression). Adds C<me.> onto the start of any column without a C<.> in
3893 it and sets C<select> from that, then auto-populates C<as> from
3894 C<select> as normal. (You may also use the C<cols> attribute, as in
3895 earlier versions of DBIC, but this is deprecated)
3897 Essentially C<columns> does the same as L</select> and L</as>.
3899 columns => [ 'foo', { bar => 'baz' } ]
3903 select => [qw/foo baz/],
3908 B<NOTE:> You B<MUST> explicitly quote C<'+columns'> when using this attribute.
3909 Not doing so causes Perl to incorrectly interpret C<+columns> as a bareword
3910 with a unary plus operator before it, which is the same as simply C<columns>.
3914 =item Value: \@extra_columns
3918 Indicates additional columns to be selected from storage. Works the same as
3919 L</columns> but adds columns to the current selection. (You may also use the
3920 C<include_columns> attribute, as in earlier versions of DBIC, but this is
3923 $schema->resultset('CD')->search(undef, {
3924 '+columns' => ['artist.name'],
3928 would return all CDs and include a 'name' column to the information
3929 passed to object inflation. Note that the 'artist' is the name of the
3930 column (or relationship) accessor, and 'name' is the name of the column
3931 accessor in the related table.
3937 =item Value: \@select_columns
3941 Indicates which columns should be selected from the storage. You can use
3942 column names, or in the case of RDBMS back ends, function or stored procedure
3945 $rs = $schema->resultset('Employee')->search(undef, {
3948 { count => 'employeeid' },
3949 { max => { length => 'name' }, -as => 'longest_name' }
3954 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
3956 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
3957 use L</select>, to instruct DBIx::Class how to store the result of the column.
3958 Also note that the L</as> attribute has nothing to do with the SQL-side 'AS'
3959 identifier aliasing. You can however alias a function, so you can use it in
3960 e.g. an C<ORDER BY> clause. This is done via the C<-as> B<select function
3961 attribute> supplied as shown in the example above.
3965 B<NOTE:> You B<MUST> explicitly quote C<'+select'> when using this attribute.
3966 Not doing so causes Perl to incorrectly interpret C<+select> as a bareword
3967 with a unary plus operator before it, which is the same as simply C<select>.
3971 =item Value: \@extra_select_columns
3975 Indicates additional columns to be selected from storage. Works the same as
3976 L</select> but adds columns to the current selection, instead of specifying
3977 a new explicit list.
3983 =item Value: \@inflation_names
3987 Indicates DBIC-side names for object inflation. That is L</as> indicates the
3988 slot name in which the column value will be stored within the
3989 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
3990 identifier by the C<get_column> method (or via the object accessor B<if one
3991 with the same name already exists>) as shown below. The L</as> attribute has
3992 B<nothing to do> with the SQL-side C<AS>. See L</select> for details.
3994 $rs = $schema->resultset('Employee')->search(undef, {
3997 { count => 'employeeid' },
3998 { max => { length => 'name' }, -as => 'longest_name' }
4007 If the object against which the search is performed already has an accessor
4008 matching a column name specified in C<as>, the value can be retrieved using
4009 the accessor as normal:
4011 my $name = $employee->name();
4013 If on the other hand an accessor does not exist in the object, you need to
4014 use C<get_column> instead:
4016 my $employee_count = $employee->get_column('employee_count');
4018 You can create your own accessors if required - see
4019 L<DBIx::Class::Manual::Cookbook> for details.
4023 B<NOTE:> You B<MUST> explicitly quote C<'+as'> when using this attribute.
4024 Not doing so causes Perl to incorrectly interpret C<+as> as a bareword
4025 with a unary plus operator before it, which is the same as simply C<as>.
4029 =item Value: \@extra_inflation_names
4033 Indicates additional inflation names for selectors added via L</+select>. See L</as>.
4039 =item Value: ($rel_name | \@rel_names | \%rel_names)
4043 Contains a list of relationships that should be joined for this query. For
4046 # Get CDs by Nine Inch Nails
4047 my $rs = $schema->resultset('CD')->search(
4048 { 'artist.name' => 'Nine Inch Nails' },
4049 { join => 'artist' }
4052 Can also contain a hash reference to refer to the other relation's relations.
4055 package MyApp::Schema::Track;
4056 use base qw/DBIx::Class/;
4057 __PACKAGE__->table('track');
4058 __PACKAGE__->add_columns(qw/trackid cd position title/);
4059 __PACKAGE__->set_primary_key('trackid');
4060 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
4063 # In your application
4064 my $rs = $schema->resultset('Artist')->search(
4065 { 'track.title' => 'Teardrop' },
4067 join => { cd => 'track' },
4068 order_by => 'artist.name',
4072 You need to use the relationship (not the table) name in conditions,
4073 because they are aliased as such. The current table is aliased as "me", so
4074 you need to use me.column_name in order to avoid ambiguity. For example:
4076 # Get CDs from 1984 with a 'Foo' track
4077 my $rs = $schema->resultset('CD')->search(
4080 'tracks.name' => 'Foo'
4082 { join => 'tracks' }
4085 If the same join is supplied twice, it will be aliased to <rel>_2 (and
4086 similarly for a third time). For e.g.
4088 my $rs = $schema->resultset('Artist')->search({
4089 'cds.title' => 'Down to Earth',
4090 'cds_2.title' => 'Popular',
4092 join => [ qw/cds cds/ ],
4095 will return a set of all artists that have both a cd with title 'Down
4096 to Earth' and a cd with title 'Popular'.
4098 If you want to fetch related objects from other tables as well, see L</prefetch>
4101 NOTE: An internal join-chain pruner will discard certain joins while
4102 constructing the actual SQL query, as long as the joins in question do not
4103 affect the retrieved result. This for example includes 1:1 left joins
4104 that are not part of the restriction specification (WHERE/HAVING) nor are
4105 a part of the query selection.
4107 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
4113 =item Value: (0 | 1)
4117 When set to a true value, indicates that any rows fetched from joined has_many
4118 relationships are to be aggregated into the corresponding "parent" object. For
4119 example, the resultset:
4121 my $rs = $schema->resultset('CD')->search({}, {
4122 '+columns' => [ qw/ tracks.title tracks.position / ],
4127 While executing the following query:
4129 SELECT me.*, tracks.title, tracks.position
4131 LEFT JOIN track tracks
4132 ON tracks.cdid = me.cdid
4134 Will return only as many objects as there are rows in the CD source, even
4135 though the result of the query may span many rows. Each of these CD objects
4136 will in turn have multiple "Track" objects hidden behind the has_many
4137 generated accessor C<tracks>. Without C<< collapse => 1 >>, the return values
4138 of this resultset would be as many CD objects as there are tracks (a "Cartesian
4139 product"), with each CD object containing exactly one of all fetched Track data.
4141 When a collapse is requested on a non-ordered resultset, an order by some
4142 unique part of the main source (the left-most table) is inserted automatically.
4143 This is done so that the resultset is allowed to be "lazy" - calling
4144 L<< $rs->next|/next >> will fetch only as many rows as it needs to build the next
4145 object with all of its related data.
4147 If an L</order_by> is already declared, and orders the resultset in a way that
4148 makes collapsing as described above impossible (e.g. C<< ORDER BY
4149 has_many_rel.column >> or C<ORDER BY RANDOM()>), DBIC will automatically
4150 switch to "eager" mode and slurp the entire resultset before constructing the
4151 first object returned by L</next>.
4153 Setting this attribute on a resultset that does not join any has_many
4154 relations is a no-op.
4156 For a more in-depth discussion, see L</PREFETCHING>.
4162 =item Value: ($rel_name | \@rel_names | \%rel_names)
4166 This attribute is a shorthand for specifying a L</join> spec, adding all
4167 columns from the joined related sources as L</+columns> and setting
4168 L</collapse> to a true value. For example, the following two queries are
4171 my $rs = $schema->resultset('Artist')->search({}, {
4172 prefetch => { cds => ['genre', 'tracks' ] },
4177 my $rs = $schema->resultset('Artist')->search({}, {
4178 join => { cds => ['genre', 'tracks' ] },
4182 { +{ "cds.$_" => "cds.$_" } }
4183 $schema->source('Artist')->related_source('cds')->columns
4186 { +{ "cds.genre.$_" => "genre.$_" } }
4187 $schema->source('Artist')->related_source('cds')->related_source('genre')->columns
4190 { +{ "cds.tracks.$_" => "tracks.$_" } }
4191 $schema->source('Artist')->related_source('cds')->related_source('tracks')->columns
4196 Both producing the following SQL:
4198 SELECT me.artistid, me.name, me.rank, me.charfield,
4199 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track,
4200 genre.genreid, genre.name,
4201 tracks.trackid, tracks.cd, tracks.position, tracks.title, tracks.last_updated_on, tracks.last_updated_at
4204 ON cds.artist = me.artistid
4205 LEFT JOIN genre genre
4206 ON genre.genreid = cds.genreid
4207 LEFT JOIN track tracks
4208 ON tracks.cd = cds.cdid
4209 ORDER BY me.artistid
4211 While L</prefetch> implies a L</join>, it is ok to mix the two together, as
4212 the arguments are properly merged and generally do the right thing. For
4213 example, you may want to do the following:
4215 my $artists_and_cds_without_genre = $schema->resultset('Artist')->search(
4216 { 'genre.genreid' => undef },
4218 join => { cds => 'genre' },
4223 Which generates the following SQL:
4225 SELECT me.artistid, me.name, me.rank, me.charfield,
4226 cds.cdid, cds.artist, cds.title, cds.year, cds.genreid, cds.single_track
4229 ON cds.artist = me.artistid
4230 LEFT JOIN genre genre
4231 ON genre.genreid = cds.genreid
4232 WHERE genre.genreid IS NULL
4233 ORDER BY me.artistid
4235 For a more in-depth discussion, see L</PREFETCHING>.
4241 =item Value: $source_alias
4245 Sets the source alias for the query. Normally, this defaults to C<me>, but
4246 nested search queries (sub-SELECTs) might need specific aliases set to
4247 reference inner queries. For example:
4250 ->related_resultset('CDs')
4251 ->related_resultset('Tracks')
4253 'track.id' => { -ident => 'none_search.id' },
4257 my $ids = $self->search({
4260 alias => 'none_search',
4261 group_by => 'none_search.id',
4262 })->get_column('id')->as_query;
4264 $self->search({ id => { -in => $ids } })
4266 This attribute is directly tied to L</current_source_alias>.
4276 Makes the resultset paged and specifies the page to retrieve. Effectively
4277 identical to creating a non-pages resultset and then calling ->page($page)
4280 If L</rows> attribute is not specified it defaults to 10 rows per page.
4282 When you have a paged resultset, L</count> will only return the number
4283 of rows in the page. To get the total, use the L</pager> and call
4284 C<total_entries> on it.
4294 Specifies the maximum number of rows for direct retrieval or the number of
4295 rows per page if the page attribute or method is used.
4301 =item Value: $offset
4305 Specifies the (zero-based) row number for the first row to be returned, or the
4306 of the first row of the first page if paging is used.
4308 =head2 software_limit
4312 =item Value: (0 | 1)
4316 When combined with L</rows> and/or L</offset> the generated SQL will not
4317 include any limit dialect stanzas. Instead the entire result will be selected
4318 as if no limits were specified, and DBIC will perform the limit locally, by
4319 artificially advancing and finishing the resulting L</cursor>.
4321 This is the recommended way of performing resultset limiting when no sane RDBMS
4322 implementation is available (e.g.
4323 L<Sybase ASE|DBIx::Class::Storage::DBI::Sybase::ASE> using the
4324 L<Generic Sub Query|DBIx::Class::SQLMaker::LimitDialects/GenericSubQ> hack)
4330 =item Value: \@columns
4334 A arrayref of columns to group by. Can include columns of joined tables.
4336 group_by => [qw/ column1 column2 ... /]
4342 =item Value: $condition
4346 HAVING is a select statement attribute that is applied between GROUP BY and
4347 ORDER BY. It is applied to the after the grouping calculations have been
4350 having => { 'count_employee' => { '>=', 100 } }
4352 or with an in-place function in which case literal SQL is required:
4354 having => \[ 'count(employee) >= ?', [ count => 100 ] ]
4360 =item Value: (0 | 1)
4364 Set to 1 to automatically generate a L</group_by> clause based on the selection
4365 (including intelligent handling of L</order_by> contents). Note that the group
4366 criteria calculation takes place over the B<final> selection. This includes
4367 any L</+columns>, L</+select> or L</order_by> additions in subsequent
4368 L</search> calls, and standalone columns selected via
4369 L<DBIx::Class::ResultSetColumn> (L</get_column>). A notable exception are the
4370 extra selections specified via L</prefetch> - such selections are explicitly
4371 excluded from group criteria calculations.
4373 If the final ResultSet also explicitly defines a L</group_by> attribute, this
4374 setting is ignored and an appropriate warning is issued.
4380 Adds to the WHERE clause.
4382 # only return rows WHERE deleted IS NULL for all searches
4383 __PACKAGE__->resultset_attributes({ where => { deleted => undef } });
4385 Can be overridden by passing C<< { where => undef } >> as an attribute
4388 For more complicated where clauses see L<SQL::Abstract/WHERE CLAUSES>.
4394 Set to 1 to cache search results. This prevents extra SQL queries if you
4395 revisit rows in your ResultSet:
4397 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
4399 while( my $artist = $resultset->next ) {
4403 $rs->first; # without cache, this would issue a query
4405 By default, searches are not cached.
4407 For more examples of using these attributes, see
4408 L<DBIx::Class::Manual::Cookbook>.
4414 =item Value: ( 'update' | 'shared' | \$scalar )
4418 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT
4419 ... FOR SHARED. If \$scalar is passed, this is taken directly and embedded in the
4424 DBIx::Class supports arbitrary related data prefetching from multiple related
4425 sources. Any combination of relationship types and column sets are supported.
4426 If L<collapsing|/collapse> is requested, there is an additional requirement of
4427 selecting enough data to make every individual object uniquely identifiable.
4429 Here are some more involved examples, based on the following relationship map:
4432 My::Schema::CD->belongs_to( artist => 'My::Schema::Artist' );
4433 My::Schema::CD->might_have( liner_note => 'My::Schema::LinerNotes' );
4434 My::Schema::CD->has_many( tracks => 'My::Schema::Track' );
4436 My::Schema::Artist->belongs_to( record_label => 'My::Schema::RecordLabel' );
4438 My::Schema::Track->has_many( guests => 'My::Schema::Guest' );
4442 my $rs = $schema->resultset('Tag')->search(
4451 The initial search results in SQL like the following:
4453 SELECT tag.*, cd.*, artist.* FROM tag
4454 JOIN cd ON tag.cd = cd.cdid
4455 JOIN artist ON cd.artist = artist.artistid
4457 L<DBIx::Class> has no need to go back to the database when we access the
4458 C<cd> or C<artist> relationships, which saves us two SQL statements in this
4461 Simple prefetches will be joined automatically, so there is no need
4462 for a C<join> attribute in the above search.
4464 The L</prefetch> attribute can be used with any of the relationship types
4465 and multiple prefetches can be specified together. Below is a more complex
4466 example that prefetches a CD's artist, its liner notes (if present),
4467 the cover image, the tracks on that CD, and the guests on those
4470 my $rs = $schema->resultset('CD')->search(
4474 { artist => 'record_label'}, # belongs_to => belongs_to
4475 'liner_note', # might_have
4476 'cover_image', # has_one
4477 { tracks => 'guests' }, # has_many => has_many
4482 This will produce SQL like the following:
4484 SELECT cd.*, artist.*, record_label.*, liner_note.*, cover_image.*,
4488 ON artist.artistid = me.artistid
4489 JOIN record_label record_label
4490 ON record_label.labelid = artist.labelid
4491 LEFT JOIN track tracks
4492 ON tracks.cdid = me.cdid
4493 LEFT JOIN guest guests
4494 ON guests.trackid = track.trackid
4495 LEFT JOIN liner_notes liner_note
4496 ON liner_note.cdid = me.cdid
4497 JOIN cd_artwork cover_image
4498 ON cover_image.cdid = me.cdid
4501 Now the C<artist>, C<record_label>, C<liner_note>, C<cover_image>,
4502 C<tracks>, and C<guests> of the CD will all be available through the
4503 relationship accessors without the need for additional queries to the
4508 Prefetch does a lot of deep magic. As such, it may not behave exactly
4509 as you might expect.
4515 Prefetch uses the L</cache> to populate the prefetched relationships. This
4516 may or may not be what you want.
4520 If you specify a condition on a prefetched relationship, ONLY those
4521 rows that match the prefetched condition will be fetched into that relationship.
4522 This means that adding prefetch to a search() B<may alter> what is returned by
4523 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
4525 my $artist_rs = $schema->resultset('Artist')->search({
4531 my $count = $artist_rs->first->cds->count;
4533 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
4535 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
4537 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
4539 That cmp_ok() may or may not pass depending on the datasets involved. In other
4540 words the C<WHERE> condition would apply to the entire dataset, just like
4541 it would in regular SQL. If you want to add a condition only to the "right side"
4542 of a C<LEFT JOIN> - consider declaring and using a L<relationship with a custom
4543 condition|DBIx::Class::Relationship::Base/condition>
4547 =head1 DBIC BIND VALUES
4549 Because DBIC may need more information to bind values than just the column name
4550 and value itself, it uses a special format for both passing and receiving bind
4551 values. Each bind value should be composed of an arrayref of
4552 C<< [ \%args => $val ] >>. The format of C<< \%args >> is currently:
4558 If present (in any form), this is what is being passed directly to bind_param.
4559 Note that different DBD's expect different bind args. (e.g. DBD::SQLite takes
4560 a single numerical type, while DBD::Pg takes a hashref if bind options.)
4562 If this is specified, all other bind options described below are ignored.
4566 If present, this is used to infer the actual bind attribute by passing to
4567 C<< $resolved_storage->bind_attribute_by_data_type() >>. Defaults to the
4568 "data_type" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>.
4570 Note that the data type is somewhat freeform (hence the sqlt_ prefix);
4571 currently drivers are expected to "Do the Right Thing" when given a common
4572 datatype name. (Not ideal, but that's what we got at this point.)
4576 Currently used to correctly allocate buffers for bind_param_inout().
4577 Defaults to "size" from the L<add_columns column info|DBIx::Class::ResultSource/add_columns>,
4578 or to a sensible value based on the "data_type".
4582 Used to fill in missing sqlt_datatype and sqlt_size attributes (if they are
4583 explicitly specified they are never overridden). Also used by some weird DBDs,
4584 where the column name should be available at bind_param time (e.g. Oracle).
4588 For backwards compatibility and convenience, the following shortcuts are
4591 [ $name => $val ] === [ { dbic_colname => $name }, $val ]
4592 [ \$dt => $val ] === [ { sqlt_datatype => $dt }, $val ]
4593 [ undef, $val ] === [ {}, $val ]
4594 $val === [ {}, $val ]
4596 =head1 AUTHOR AND CONTRIBUTORS
4598 See L<AUTHOR|DBIx::Class/AUTHOR> and L<CONTRIBUTORS|DBIx::Class/CONTRIBUTORS> in DBIx::Class
4602 You may distribute this code under the same terms as Perl itself.