1 package DBIx::Class::ResultSet;
5 use base qw/DBIx::Class/;
7 use DBIx::Class::Exception;
8 use DBIx::Class::ResultSetColumn;
9 use Scalar::Util qw/blessed weaken/;
11 use Data::Compare (); # no imports!!! guard against insane architecture
13 # not importing first() as it will clash with our own method
17 # De-duplication in _merge_attr() is disabled, but left in for reference
18 # (the merger is used for other things that ought not to be de-duped)
19 *__HM_DEDUP = sub () { 0 };
29 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class result_source/);
33 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
37 my $users_rs = $schema->resultset('User');
38 while( $user = $users_rs->next) {
39 print $user->username;
42 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
43 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
47 A ResultSet is an object which stores a set of conditions representing
48 a query. It is the backbone of DBIx::Class (i.e. the really
49 important/useful bit).
51 No SQL is executed on the database when a ResultSet is created, it
52 just stores all the conditions needed to create the query.
54 A basic ResultSet representing the data of an entire table is returned
55 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
56 L<Source|DBIx::Class::Manual::Glossary/Source> name.
58 my $users_rs = $schema->resultset('User');
60 A new ResultSet is returned from calling L</search> on an existing
61 ResultSet. The new one will contain all the conditions of the
62 original, plus any new conditions added in the C<search> call.
64 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
65 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
68 The query that the ResultSet represents is B<only> executed against
69 the database when these methods are called:
70 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
72 If a resultset is used in a numeric context it returns the L</count>.
73 However, if it is used in a boolean context it is B<always> true. So if
74 you want to check if a resultset has any results, you must use C<if $rs
77 =head1 CUSTOM ResultSet CLASSES THAT USE Moose
79 If you want to make your custom ResultSet classes with L<Moose>, use a template
82 package MyApp::Schema::ResultSet::User;
85 use namespace::autoclean;
87 extends 'DBIx::Class::ResultSet';
89 sub BUILDARGS { $_[2] }
93 __PACKAGE__->meta->make_immutable;
97 The L<MooseX::NonMoose> is necessary so that the L<Moose> constructor does not
98 clash with the regular ResultSet constructor. Alternatively, you can use:
100 __PACKAGE__->meta->make_immutable(inline_constructor => 0);
102 The L<BUILDARGS|Moose::Manual::Construction/BUILDARGS> is necessary because the
103 signature of the ResultSet C<new> is C<< ->new($source, \%args) >>.
107 =head2 Chaining resultsets
109 Let's say you've got a query that needs to be run to return some data
110 to the user. But, you have an authorization system in place that
111 prevents certain users from seeing certain information. So, you want
112 to construct the basic query in one method, but add constraints to it in
117 my $request = $self->get_request; # Get a request object somehow.
118 my $schema = $self->result_source->schema;
120 my $cd_rs = $schema->resultset('CD')->search({
121 title => $request->param('title'),
122 year => $request->param('year'),
125 $cd_rs = $self->apply_security_policy( $cd_rs );
127 return $cd_rs->all();
130 sub apply_security_policy {
139 =head3 Resolving conditions and attributes
141 When a resultset is chained from another resultset, conditions and
142 attributes with the same keys need resolving.
144 L</join>, L</prefetch>, L</+select>, L</+as> attributes are merged
145 into the existing ones from the original resultset.
147 The L</where> and L</having> attributes, and any search conditions, are
148 merged with an SQL C<AND> to the existing condition from the original
151 All other attributes are overridden by any new ones supplied in the
154 =head2 Multiple queries
156 Since a resultset just defines a query, you can do all sorts of
157 things with it with the same object.
159 # Don't hit the DB yet.
160 my $cd_rs = $schema->resultset('CD')->search({
161 title => 'something',
165 # Each of these hits the DB individually.
166 my $count = $cd_rs->count;
167 my $most_recent = $cd_rs->get_column('date_released')->max();
168 my @records = $cd_rs->all;
170 And it's not just limited to SELECT statements.
176 $cd_rs->create({ artist => 'Fred' });
178 Which is the same as:
180 $schema->resultset('CD')->create({
181 title => 'something',
186 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
194 =item Arguments: $source, \%$attrs
196 =item Return Value: $rs
200 The resultset constructor. Takes a source object (usually a
201 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
202 L</ATTRIBUTES> below). Does not perform any queries -- these are
203 executed as needed by the other methods.
205 Generally you won't need to construct a resultset manually. You'll
206 automatically get one from e.g. a L</search> called in scalar context:
208 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
210 IMPORTANT: If called on an object, proxies to new_result instead so
212 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
214 will return a CD object, not a ResultSet.
220 return $class->new_result(@_) if ref $class;
222 my ($source, $attrs) = @_;
223 $source = $source->resolve
224 if $source->isa('DBIx::Class::ResultSourceHandle');
225 $attrs = { %{$attrs||{}} };
227 if ($attrs->{page}) {
228 $attrs->{rows} ||= 10;
231 $attrs->{alias} ||= 'me';
234 result_source => $source,
235 cond => $attrs->{where},
240 # if there is a dark selector, this means we are already in a
241 # chain and the cleanup/sanification was taken care of by
243 $self->_normalize_selection($attrs)
244 unless $attrs->{_dark_selector};
247 $attrs->{result_class} || $source->result_class
257 =item Arguments: $cond, \%attrs?
259 =item Return Value: $resultset (scalar context) || @row_objs (list context)
263 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
264 my $new_rs = $cd_rs->search({ year => 2005 });
266 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
267 # year = 2005 OR year = 2004
269 In list context, C<< ->all() >> is called implicitly on the resultset, thus
270 returning a list of row objects instead. To avoid that, use L</search_rs>.
272 If you need to pass in additional attributes but no additional condition,
273 call it as C<search(undef, \%attrs)>.
275 # "SELECT name, artistid FROM $artist_table"
276 my @all_artists = $schema->resultset('Artist')->search(undef, {
277 columns => [qw/name artistid/],
280 For a list of attributes that can be passed to C<search>, see
281 L</ATTRIBUTES>. For more examples of using this function, see
282 L<Searching|DBIx::Class::Manual::Cookbook/Searching>. For a complete
283 documentation for the first argument, see L<SQL::Abstract>
284 and its extension L<DBIx::Class::SQLMaker>.
286 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
290 Note that L</search> does not process/deflate any of the values passed in the
291 L<SQL::Abstract>-compatible search condition structure. This is unlike other
292 condition-bound methods L</new>, L</create> and L</find>. The user must ensure
293 manually that any value passed to this method will stringify to something the
294 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
295 objects, for more info see:
296 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
302 my $rs = $self->search_rs( @_ );
307 elsif (defined wantarray) {
311 # we can be called by a relationship helper, which in
312 # turn may be called in void context due to some braindead
313 # overload or whatever else the user decided to be clever
314 # at this particular day. Thus limit the exception to
315 # external code calls only
316 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense')
317 if (caller)[0] !~ /^\QDBIx::Class::/;
327 =item Arguments: $cond, \%attrs?
329 =item Return Value: $resultset
333 This method does the same exact thing as search() except it will
334 always return a resultset, even in list context.
341 # Special-case handling for (undef, undef).
342 if ( @_ == 2 && !defined $_[1] && !defined $_[0] ) {
348 if (ref $_[-1] eq 'HASH') {
349 # copy for _normalize_selection
350 $call_attrs = { %{ pop @_ } };
352 elsif (! defined $_[-1] ) {
353 pop @_; # search({}, undef)
357 # see if we can keep the cache (no $rs changes)
359 my %safe = (alias => 1, cache => 1);
360 if ( ! List::Util::first { !$safe{$_} } keys %$call_attrs and (
363 ref $_[0] eq 'HASH' && ! keys %{$_[0]}
365 ref $_[0] eq 'ARRAY' && ! @{$_[0]}
367 $cache = $self->get_cache;
370 my $rsrc = $self->result_source;
372 my $old_attrs = { %{$self->{attrs}} };
373 my $old_having = delete $old_attrs->{having};
374 my $old_where = delete $old_attrs->{where};
376 my $new_attrs = { %$old_attrs };
378 # take care of call attrs (only if anything is changing)
379 if (keys %$call_attrs) {
381 my @selector_attrs = qw/select as columns cols +select +as +columns include_columns/;
383 # reset the current selector list if new selectors are supplied
384 if (List::Util::first { exists $call_attrs->{$_} } qw/columns cols select as/) {
385 delete @{$old_attrs}{(@selector_attrs, '_dark_selector')};
388 # Normalize the new selector list (operates on the passed-in attr structure)
389 # Need to do it on every chain instead of only once on _resolved_attrs, in
390 # order to allow detection of empty vs partial 'as'
391 $call_attrs->{_dark_selector} = $old_attrs->{_dark_selector}
392 if $old_attrs->{_dark_selector};
393 $self->_normalize_selection ($call_attrs);
395 # start with blind overwriting merge, exclude selector attrs
396 $new_attrs = { %{$old_attrs}, %{$call_attrs} };
397 delete @{$new_attrs}{@selector_attrs};
399 for (@selector_attrs) {
400 $new_attrs->{$_} = $self->_merge_attr($old_attrs->{$_}, $call_attrs->{$_})
401 if ( exists $old_attrs->{$_} or exists $call_attrs->{$_} );
404 # older deprecated name, use only if {columns} is not there
405 if (my $c = delete $new_attrs->{cols}) {
406 if ($new_attrs->{columns}) {
407 carp "Resultset specifies both the 'columns' and the legacy 'cols' attributes - ignoring 'cols'";
410 $new_attrs->{columns} = $c;
415 # join/prefetch use their own crazy merging heuristics
416 foreach my $key (qw/join prefetch/) {
417 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key})
418 if exists $call_attrs->{$key};
421 # stack binds together
422 $new_attrs->{bind} = [ @{ $old_attrs->{bind} || [] }, @{ $call_attrs->{bind} || [] } ];
426 # rip apart the rest of @_, parse a condition
429 if (ref $_[0] eq 'HASH') {
430 (keys %{$_[0]}) ? $_[0] : undef
436 $self->throw_exception('Odd number of arguments to search')
444 if( @_ > 1 and ! $rsrc->result_class->isa('DBIx::Class::CDBICompat') ) {
445 carp_unique 'search( %condition ) is deprecated, use search( \%condition ) instead';
448 for ($old_where, $call_cond) {
450 $new_attrs->{where} = $self->_stack_cond (
451 $_, $new_attrs->{where}
456 if (defined $old_having) {
457 $new_attrs->{having} = $self->_stack_cond (
458 $old_having, $new_attrs->{having}
462 my $rs = (ref $self)->new($rsrc, $new_attrs);
464 $rs->set_cache($cache) if ($cache);
470 sub _normalize_selection {
471 my ($self, $attrs) = @_;
474 $attrs->{'+columns'} = $self->_merge_attr($attrs->{'+columns'}, delete $attrs->{include_columns})
475 if exists $attrs->{include_columns};
477 # columns are always placed first, however
479 # Keep the X vs +X separation until _resolved_attrs time - this allows to
480 # delay the decision on whether to use a default select list ($rsrc->columns)
481 # allowing stuff like the remove_columns helper to work
483 # select/as +select/+as pairs need special handling - the amount of select/as
484 # elements in each pair does *not* have to be equal (think multicolumn
485 # selectors like distinct(foo, bar) ). If the selector is bare (no 'as'
486 # supplied at all) - try to infer the alias, either from the -as parameter
487 # of the selector spec, or use the parameter whole if it looks like a column
488 # name (ugly legacy heuristic). If all fails - leave the selector bare (which
489 # is ok as well), but make sure no more additions to the 'as' chain take place
490 for my $pref ('', '+') {
492 my ($sel, $as) = map {
493 my $key = "${pref}${_}";
495 my $val = [ ref $attrs->{$key} eq 'ARRAY'
497 : $attrs->{$key} || ()
499 delete $attrs->{$key};
503 if (! @$as and ! @$sel ) {
506 elsif (@$as and ! @$sel) {
507 $self->throw_exception(
508 "Unable to handle ${pref}as specification (@$as) without a corresponding ${pref}select"
512 # no as part supplied at all - try to deduce (unless explicit end of named selection is declared)
513 # if any @$as has been supplied we assume the user knows what (s)he is doing
514 # and blindly keep stacking up pieces
515 unless ($attrs->{_dark_selector}) {
518 if ( ref $_ eq 'HASH' and exists $_->{-as} ) {
519 push @$as, $_->{-as};
521 # assume any plain no-space, no-parenthesis string to be a column spec
522 # FIXME - this is retarded but is necessary to support shit like 'count(foo)'
523 elsif ( ! ref $_ and $_ =~ /^ [^\s\(\)]+ $/x) {
526 # if all else fails - raise a flag that no more aliasing will be allowed
528 $attrs->{_dark_selector} = {
530 string => ($dark_sel_dumper ||= do {
531 require Data::Dumper::Concise;
532 Data::Dumper::Concise::DumperObject()->Indent(0);
533 })->Values([$_])->Dump
541 elsif (@$as < @$sel) {
542 $self->throw_exception(
543 "Unable to handle an ${pref}as specification (@$as) with less elements than the corresponding ${pref}select"
546 elsif ($pref and $attrs->{_dark_selector}) {
547 $self->throw_exception(
548 "Unable to process named '+select', resultset contains an unnamed selector $attrs->{_dark_selector}{string}"
554 $attrs->{"${pref}select"} = $self->_merge_attr($attrs->{"${pref}select"}, $sel);
555 $attrs->{"${pref}as"} = $self->_merge_attr($attrs->{"${pref}as"}, $as);
560 my ($self, $left, $right) = @_;
562 # collapse single element top-level conditions
563 # (single pass only, unlikely to need recursion)
564 for ($left, $right) {
565 if (ref $_ eq 'ARRAY') {
573 elsif (ref $_ eq 'HASH') {
574 my ($first, $more) = keys %$_;
577 if (! defined $first) {
581 elsif (! defined $more) {
582 if ($first eq '-and' and ref $_->{'-and'} eq 'HASH') {
585 elsif ($first eq '-or' and ref $_->{'-or'} eq 'ARRAY') {
592 # merge hashes with weeding out of duplicates (simple cases only)
593 if (ref $left eq 'HASH' and ref $right eq 'HASH') {
595 # shallow copy to destroy
596 $right = { %$right };
597 for (grep { exists $right->{$_} } keys %$left) {
598 # the use of eq_deeply here is justified - the rhs of an
599 # expression can contain a lot of twisted weird stuff
600 delete $right->{$_} if Data::Compare::Compare( $left->{$_}, $right->{$_} );
603 $right = undef unless keys %$right;
607 if (defined $left xor defined $right) {
608 return defined $left ? $left : $right;
610 elsif (! defined $left) {
614 return { -and => [ $left, $right ] };
618 =head2 search_literal
622 =item Arguments: $sql_fragment, @bind_values
624 =item Return Value: $resultset (scalar context) || @row_objs (list context)
628 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
629 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
631 Pass a literal chunk of SQL to be added to the conditional part of the
634 CAVEAT: C<search_literal> is provided for Class::DBI compatibility and should
635 only be used in that context. C<search_literal> is a convenience method.
636 It is equivalent to calling $schema->search(\[]), but if you want to ensure
637 columns are bound correctly, use C<search>.
639 Example of how to use C<search> instead of C<search_literal>
641 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
642 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
645 See L<DBIx::Class::Manual::Cookbook/Searching> and
646 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
647 require C<search_literal>.
652 my ($self, $sql, @bind) = @_;
654 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
657 return $self->search(\[ $sql, map [ __DUMMY__ => $_ ], @bind ], ($attr || () ));
664 =item Arguments: \%columns_values | @pk_values, \%attrs?
666 =item Return Value: $row_object | undef
670 Finds and returns a single row based on supplied criteria. Takes either a
671 hashref with the same format as L</create> (including inference of foreign
672 keys from related objects), or a list of primary key values in the same
673 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
674 declaration on the L</result_source>.
676 In either case an attempt is made to combine conditions already existing on
677 the resultset with the condition passed to this method.
679 To aid with preparing the correct query for the storage you may supply the
680 C<key> attribute, which is the name of a
681 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
682 unique constraint corresponding to the
683 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
684 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
685 to construct a query that satisfies the named unique constraint fully (
686 non-NULL values for each column member of the constraint) an exception is
689 If no C<key> is specified, the search is carried over all unique constraints
690 which are fully defined by the available condition.
692 If no such constraint is found, C<find> currently defaults to a simple
693 C<< search->(\%column_values) >> which may or may not do what you expect.
694 Note that this fallback behavior may be deprecated in further versions. If
695 you need to search with arbitrary conditions - use L</search>. If the query
696 resulting from this fallback produces more than one row, a warning to the
697 effect is issued, though only the first row is constructed and returned as
700 In addition to C<key>, L</find> recognizes and applies standard
701 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
703 Note that if you have extra concerns about the correctness of the resulting
704 query you need to specify the C<key> attribute and supply the entire condition
705 as an argument to find (since it is not always possible to perform the
706 combination of the resultset condition with the supplied one, especially if
707 the resultset condition contains literal sql).
709 For example, to find a row by its primary key:
711 my $cd = $schema->resultset('CD')->find(5);
713 You can also find a row by a specific unique constraint:
715 my $cd = $schema->resultset('CD')->find(
717 artist => 'Massive Attack',
718 title => 'Mezzanine',
720 { key => 'cd_artist_title' }
723 See also L</find_or_create> and L</update_or_create>.
729 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
731 my $rsrc = $self->result_source;
734 if (exists $attrs->{key}) {
735 $constraint_name = defined $attrs->{key}
737 : $self->throw_exception("An undefined 'key' resultset attribute makes no sense")
741 # Parse out the condition from input
744 if (ref $_[0] eq 'HASH') {
745 $call_cond = { %{$_[0]} };
748 # if only values are supplied we need to default to 'primary'
749 $constraint_name = 'primary' unless defined $constraint_name;
751 my @c_cols = $rsrc->unique_constraint_columns($constraint_name);
753 $self->throw_exception(
754 "No constraint columns, maybe a malformed '$constraint_name' constraint?"
757 $self->throw_exception (
758 'find() expects either a column/value hashref, or a list of values '
759 . "corresponding to the columns of the specified unique constraint '$constraint_name'"
760 ) unless @c_cols == @_;
763 @{$call_cond}{@c_cols} = @_;
767 for my $key (keys %$call_cond) {
769 my $keyref = ref($call_cond->{$key})
771 my $relinfo = $rsrc->relationship_info($key)
773 my $val = delete $call_cond->{$key};
775 next if $keyref eq 'ARRAY'; # has_many for multi_create
777 my $rel_q = $rsrc->_resolve_condition(
778 $relinfo->{cond}, $val, $key, $key
780 die "Can't handle complex relationship conditions in find" if ref($rel_q) ne 'HASH';
781 @related{keys %$rel_q} = values %$rel_q;
785 # relationship conditions take precedence (?)
786 @{$call_cond}{keys %related} = values %related;
788 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
790 if (defined $constraint_name) {
791 $final_cond = $self->_qualify_cond_columns (
793 $self->_build_unique_cond (
801 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
802 # This means that we got here after a merger of relationship conditions
803 # in ::Relationship::Base::search_related (the row method), and furthermore
804 # the relationship is of the 'single' type. This means that the condition
805 # provided by the relationship (already attached to $self) is sufficient,
806 # as there can be only one row in the database that would satisfy the
810 # no key was specified - fall down to heuristics mode:
811 # run through all unique queries registered on the resultset, and
812 # 'OR' all qualifying queries together
813 my (@unique_queries, %seen_column_combinations);
814 for my $c_name ($rsrc->unique_constraint_names) {
815 next if $seen_column_combinations{
816 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
819 push @unique_queries, try {
820 $self->_build_unique_cond ($c_name, $call_cond, 'croak_on_nulls')
824 $final_cond = @unique_queries
825 ? [ map { $self->_qualify_cond_columns($_, $alias) } @unique_queries ]
826 : $self->_non_unique_find_fallback ($call_cond, $attrs)
830 # Run the query, passing the result_class since it should propagate for find
831 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
832 if ($rs->_resolved_attrs->{collapse}) {
834 carp "Query returned more than one row" if $rs->next;
842 # This is a stop-gap method as agreed during the discussion on find() cleanup:
843 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
845 # It is invoked when find() is called in legacy-mode with insufficiently-unique
846 # condition. It is provided for overrides until a saner way forward is devised
848 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
849 # the road. Please adjust your tests accordingly to catch this situation early
850 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
852 # The method will not be removed without an adequately complete replacement
853 # for strict-mode enforcement
854 sub _non_unique_find_fallback {
855 my ($self, $cond, $attrs) = @_;
857 return $self->_qualify_cond_columns(
859 exists $attrs->{alias}
861 : $self->{attrs}{alias}
866 sub _qualify_cond_columns {
867 my ($self, $cond, $alias) = @_;
869 my %aliased = %$cond;
870 for (keys %aliased) {
871 $aliased{"$alias.$_"} = delete $aliased{$_}
878 sub _build_unique_cond {
879 my ($self, $constraint_name, $extra_cond, $croak_on_null) = @_;
881 my @c_cols = $self->result_source->unique_constraint_columns($constraint_name);
883 # combination may fail if $self->{cond} is non-trivial
884 my ($final_cond) = try {
885 $self->_merge_with_rscond ($extra_cond)
890 # trim out everything not in $columns
891 $final_cond = { map {
892 exists $final_cond->{$_}
893 ? ( $_ => $final_cond->{$_} )
897 if (my @missing = grep
898 { ! ($croak_on_null ? defined $final_cond->{$_} : exists $final_cond->{$_}) }
901 $self->throw_exception( sprintf ( "Unable to satisfy requested constraint '%s', no values for column(s): %s",
903 join (', ', map { "'$_'" } @missing),
910 !$ENV{DBIC_NULLABLE_KEY_NOWARN}
912 my @undefs = grep { ! defined $final_cond->{$_} } (keys %$final_cond)
914 carp_unique ( sprintf (
915 "NULL/undef values supplied for requested unique constraint '%s' (NULL "
916 . 'values in column(s): %s). This is almost certainly not what you wanted, '
917 . 'though you can set DBIC_NULLABLE_KEY_NOWARN to disable this warning.',
919 join (', ', map { "'$_'" } @undefs),
926 =head2 search_related
930 =item Arguments: $rel, $cond?, \%attrs?
932 =item Return Value: $new_resultset (scalar context) || @row_objs (list context)
936 $new_rs = $cd_rs->search_related('artist', {
940 Searches the specified relationship, optionally specifying a condition and
941 attributes for matching records. See L</ATTRIBUTES> for more information.
943 In list context, C<< ->all() >> is called implicitly on the resultset, thus
944 returning a list of row objects instead. To avoid that, use L</search_related_rs>.
946 See also L</search_related_rs>.
951 return shift->related_resultset(shift)->search(@_);
954 =head2 search_related_rs
956 This method works exactly the same as search_related, except that
957 it guarantees a resultset, even in list context.
961 sub search_related_rs {
962 return shift->related_resultset(shift)->search_rs(@_);
969 =item Arguments: none
971 =item Return Value: $cursor
975 Returns a storage-driven cursor to the given resultset. See
976 L<DBIx::Class::Cursor> for more information.
983 my $attrs = $self->_resolved_attrs_copy;
985 return $self->{cursor}
986 ||= $self->result_source->storage->select($attrs->{from}, $attrs->{select},
987 $attrs->{where},$attrs);
994 =item Arguments: $cond?
996 =item Return Value: $row_object | undef
1000 my $cd = $schema->resultset('CD')->single({ year => 2001 });
1002 Inflates the first result without creating a cursor if the resultset has
1003 any records in it; if not returns C<undef>. Used by L</find> as a lean version
1006 While this method can take an optional search condition (just like L</search>)
1007 being a fast-code-path it does not recognize search attributes. If you need to
1008 add extra joins or similar, call L</search> and then chain-call L</single> on the
1009 L<DBIx::Class::ResultSet> returned.
1015 As of 0.08100, this method enforces the assumption that the preceding
1016 query returns only one row. If more than one row is returned, you will receive
1019 Query returned more than one row
1021 In this case, you should be using L</next> or L</find> instead, or if you really
1022 know what you are doing, use the L</rows> attribute to explicitly limit the size
1025 This method will also throw an exception if it is called on a resultset prefetching
1026 has_many, as such a prefetch implies fetching multiple rows from the database in
1027 order to assemble the resulting object.
1034 my ($self, $where) = @_;
1036 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
1039 my $attrs = $self->_resolved_attrs_copy;
1041 $self->throw_exception(
1042 'single() can not be used on resultsets prefetching has_many. Use find( \%cond ) or next() instead'
1043 ) if $attrs->{collapse};
1046 if (defined $attrs->{where}) {
1049 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
1050 $where, delete $attrs->{where} ]
1053 $attrs->{where} = $where;
1057 my $data = [ $self->result_source->storage->select_single(
1058 $attrs->{from}, $attrs->{select},
1059 $attrs->{where}, $attrs
1061 return undef unless @$data;
1062 $self->{stashed_rows} = [ $data ];
1063 $self->_construct_objects->[0];
1069 # Recursively collapse the query, accumulating values for each column.
1071 sub _collapse_query {
1072 my ($self, $query, $collapsed) = @_;
1076 if (ref $query eq 'ARRAY') {
1077 foreach my $subquery (@$query) {
1078 next unless ref $subquery; # -or
1079 $collapsed = $self->_collapse_query($subquery, $collapsed);
1082 elsif (ref $query eq 'HASH') {
1083 if (keys %$query and (keys %$query)[0] eq '-and') {
1084 foreach my $subquery (@{$query->{-and}}) {
1085 $collapsed = $self->_collapse_query($subquery, $collapsed);
1089 foreach my $col (keys %$query) {
1090 my $value = $query->{$col};
1091 $collapsed->{$col}{$value}++;
1103 =item Arguments: $cond?
1105 =item Return Value: $resultsetcolumn
1109 my $max_length = $rs->get_column('length')->max;
1111 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
1116 my ($self, $column) = @_;
1117 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
1125 =item Arguments: $cond, \%attrs?
1127 =item Return Value: $resultset (scalar context) || @row_objs (list context)
1131 # WHERE title LIKE '%blue%'
1132 $cd_rs = $rs->search_like({ title => '%blue%'});
1134 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
1135 that this is simply a convenience method retained for ex Class::DBI users.
1136 You most likely want to use L</search> with specific operators.
1138 For more information, see L<DBIx::Class::Manual::Cookbook>.
1140 This method is deprecated and will be removed in 0.09. Use L</search()>
1141 instead. An example conversion is:
1143 ->search_like({ foo => 'bar' });
1147 ->search({ foo => { like => 'bar' } });
1154 'search_like() is deprecated and will be removed in DBIC version 0.09.'
1155 .' Instead use ->search({ x => { -like => "y%" } })'
1156 .' (note the outer pair of {}s - they are important!)'
1158 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
1159 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
1160 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
1161 return $class->search($query, { %$attrs });
1168 =item Arguments: $first, $last
1170 =item Return Value: $resultset (scalar context) || @row_objs (list context)
1174 Returns a resultset or object list representing a subset of elements from the
1175 resultset slice is called on. Indexes are from 0, i.e., to get the first
1176 three records, call:
1178 my ($one, $two, $three) = $rs->slice(0, 2);
1183 my ($self, $min, $max) = @_;
1184 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
1185 $attrs->{offset} = $self->{attrs}{offset} || 0;
1186 $attrs->{offset} += $min;
1187 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
1188 return $self->search(undef, $attrs);
1189 #my $slice = (ref $self)->new($self->result_source, $attrs);
1190 #return (wantarray ? $slice->all : $slice);
1197 =item Arguments: none
1199 =item Return Value: $result | undef
1203 Returns the next element in the resultset (C<undef> is there is none).
1205 Can be used to efficiently iterate over records in the resultset:
1207 my $rs = $schema->resultset('CD')->search;
1208 while (my $cd = $rs->next) {
1212 Note that you need to store the resultset object, and call C<next> on it.
1213 Calling C<< resultset('Table')->next >> repeatedly will always return the
1214 first record from the resultset.
1221 if (my $cache = $self->get_cache) {
1222 $self->{all_cache_position} ||= 0;
1223 return $cache->[$self->{all_cache_position}++];
1226 if ($self->{attrs}{cache}) {
1227 delete $self->{pager};
1228 $self->{all_cache_position} = 1;
1229 return ($self->all)[0];
1232 return shift(@{$self->{stashed_objects}}) if @{ $self->{stashed_objects}||[] };
1234 $self->{stashed_objects} = $self->_construct_objects
1237 return shift @{$self->{stashed_objects}};
1240 # takes a single DBI-row of data and coinstructs as many objects
1241 # as the resultset attributes call for.
1242 # This can be a bit of an action at a distance - it takes as an argument
1243 # the *current* cursor-row (already taken off the $sth), but if
1244 # collapsing is requested it will keep advancing the cursor either
1245 # until the current row-object is assembled (the collapser was able to
1246 # order the result sensibly) OR until the cursor is exhausted (an
1247 # unordered collapsing resultset effectively triggers ->all)
1248 sub _construct_objects {
1249 my ($self, $fetch_all) = @_;
1251 my $rsrc = $self->result_source;
1252 my $attrs = $self->_resolved_attrs;
1253 my $cursor = $self->cursor;
1255 # this will be used as both initial raw-row collector AND as a RV of
1256 # _construct_objects. Not regrowing the array twice matters a lot...
1257 # a suprising amount actually
1258 my $rows = (delete $self->{stashed_rows}) || [];
1260 # FIXME - we can do better, cursor->next/all (well diff. methods) should return a ref
1261 $rows = [ @$rows, $cursor->all ];
1263 elsif (!$attrs->{collapse}) {
1264 push @$rows, do { my @r = $cursor->next; @r ? \@r : () }
1268 $attrs->{_ordered_for_collapse} ||= (!$attrs->{order_by}) ? undef : do {
1269 my $st = $rsrc->schema->storage;
1272 ( $st->_extract_order_criteria($attrs->{order_by}) )
1275 my $colinfos = $st->_resolve_column_info($attrs->{from}, \@ord_cols);
1277 for (0 .. $#ord_cols) {
1279 ! $colinfos->{$ord_cols[$_]}
1281 $colinfos->{$ord_cols[$_]}{-result_source} != $rsrc
1283 splice @ord_cols, $_;
1288 # since all we check here are the start of the order_by belonging to the
1289 # top level $rsrc, the order stability check will fail unless the whole
1290 # thing is ordered as we need it
1291 (@ord_cols and $rsrc->_identifying_column_set({ map
1292 { $colinfos->{$_}{-colname} => $colinfos->{$_} }
1297 if ($attrs->{_ordered_for_collapse}) {
1298 push @$rows, do { my @r = $cursor->next; @r ? \@r : () };
1300 # instead of looping over ->next, use ->all in stealth mode
1301 elsif (! $cursor->{done}) {
1302 push @$rows, $cursor->all;
1303 $cursor->{done} = 1;
1308 return undef unless @$rows;
1310 my $res_class = $self->result_class;
1311 my $inflator = $res_class->can ('inflate_result')
1312 or $self->throw_exception("Inflator $res_class does not provide an inflate_result() method");
1314 my $infmap = $attrs->{as};
1316 if (!$attrs->{collapse} and $attrs->{_single_object_inflation}) {
1317 # construct a much simpler array->hash folder for the one-table cases right here
1319 # FIXME SUBOPTIMAL this is a very very very hot spot
1320 # while rather optimal we can *still* do much better, by
1321 # building a smarter [Row|HRI]::inflate_result(), and
1322 # switch to feeding it data via a much leaner interface
1324 # crude unscientific benchmarking indicated the shortcut eval is not worth it for
1325 # this particular resultset size
1327 my @as_idx = 0..$#$infmap;
1328 for my $r (@$rows) {
1329 $r = $inflator->($res_class, $rsrc, { map { $infmap->[$_] => $r->[$_] } @as_idx } );
1334 '$_ = $inflator->($res_class, $rsrc, { %s }) for @$rows',
1335 join (', ', map { "\$infmap->[$_] => \$_->[$_]" } 0..$#$infmap )
1340 ($self->{_row_parser} ||= eval sprintf 'sub { %s }', $rsrc->_mk_row_parser({
1341 inflate_map => $infmap,
1342 selection => $attrs->{select},
1343 collapse => $attrs->{collapse},
1344 }) or die $@)->($rows, $fetch_all ? () : (
1345 sub { my @r = $cursor->next or return; \@r },
1346 ($self->{stashed_rows} = []),
1347 )); # modify $rows in-place, shrinking/extending as necessary
1349 $_ = $inflator->($res_class, $rsrc, @$_) for @$rows;
1354 if ($attrs->{record_filter}) {
1355 $_ = $attrs->{record_filter}->($_) for @$rows;
1361 =head2 result_source
1365 =item Arguments: $result_source?
1367 =item Return Value: $result_source
1371 An accessor for the primary ResultSource object from which this ResultSet
1378 =item Arguments: $result_class?
1380 =item Return Value: $result_class
1384 An accessor for the class to use when creating row objects. Defaults to
1385 C<< result_source->result_class >> - which in most cases is the name of the
1386 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1388 Note that changing the result_class will also remove any components
1389 that were originally loaded in the source class via
1390 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1391 in the original source class will not run.
1396 my ($self, $result_class) = @_;
1397 if ($result_class) {
1398 unless (ref $result_class) { # don't fire this for an object
1399 $self->ensure_class_loaded($result_class);
1401 $self->_result_class($result_class);
1402 # THIS LINE WOULD BE A BUG - this accessor specifically exists to
1403 # permit the user to set result class on one result set only; it only
1404 # chains if provided to search()
1405 #$self->{attrs}{result_class} = $result_class if ref $self;
1407 $self->_result_class;
1414 =item Arguments: $cond, \%attrs??
1416 =item Return Value: $count
1420 Performs an SQL C<COUNT> with the same query as the resultset was built
1421 with to find the number of elements. Passing arguments is equivalent to
1422 C<< $rs->search ($cond, \%attrs)->count >>
1428 return $self->search(@_)->count if @_ and defined $_[0];
1429 return scalar @{ $self->get_cache } if $self->get_cache;
1431 my $attrs = $self->_resolved_attrs_copy;
1433 # this is a little optimization - it is faster to do the limit
1434 # adjustments in software, instead of a subquery
1435 my $rows = delete $attrs->{rows};
1436 my $offset = delete $attrs->{offset};
1439 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1440 $crs = $self->_count_subq_rs ($attrs);
1443 $crs = $self->_count_rs ($attrs);
1445 my $count = $crs->next;
1447 $count -= $offset if $offset;
1448 $count = $rows if $rows and $rows < $count;
1449 $count = 0 if ($count < 0);
1458 =item Arguments: $cond, \%attrs??
1460 =item Return Value: $count_rs
1464 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1465 This can be very handy for subqueries:
1467 ->search( { amount => $some_rs->count_rs->as_query } )
1469 As with regular resultsets the SQL query will be executed only after
1470 the resultset is accessed via L</next> or L</all>. That would return
1471 the same single value obtainable via L</count>.
1477 return $self->search(@_)->count_rs if @_;
1479 # this may look like a lack of abstraction (count() does about the same)
1480 # but in fact an _rs *must* use a subquery for the limits, as the
1481 # software based limiting can not be ported if this $rs is to be used
1482 # in a subquery itself (i.e. ->as_query)
1483 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1484 return $self->_count_subq_rs;
1487 return $self->_count_rs;
1492 # returns a ResultSetColumn object tied to the count query
1495 my ($self, $attrs) = @_;
1497 my $rsrc = $self->result_source;
1498 $attrs ||= $self->_resolved_attrs;
1500 my $tmp_attrs = { %$attrs };
1501 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1502 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1504 # overwrite the selector (supplied by the storage)
1505 $tmp_attrs->{select} = $rsrc->storage->_count_select ($rsrc, $attrs);
1506 $tmp_attrs->{as} = 'count';
1507 delete @{$tmp_attrs}{qw/columns/};
1509 my $tmp_rs = $rsrc->resultset_class->new($rsrc, $tmp_attrs)->get_column ('count');
1515 # same as above but uses a subquery
1517 sub _count_subq_rs {
1518 my ($self, $attrs) = @_;
1520 my $rsrc = $self->result_source;
1521 $attrs ||= $self->_resolved_attrs;
1523 my $sub_attrs = { %$attrs };
1524 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1525 delete @{$sub_attrs}{qw/collapse columns as select _prefetch_selector_range order_by for/};
1527 # if we multi-prefetch we group_by something unique, as this is what we would
1528 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1529 if ( $attrs->{collapse} ) {
1530 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } @{
1531 $rsrc->_identifying_column_set || $self->throw_exception(
1532 'Unable to construct a unique group_by criteria properly collapsing the '
1533 . 'has_many prefetch before count()'
1538 # Calculate subquery selector
1539 if (my $g = $sub_attrs->{group_by}) {
1541 my $sql_maker = $rsrc->storage->sql_maker;
1543 # necessary as the group_by may refer to aliased functions
1545 for my $sel (@{$attrs->{select}}) {
1546 $sel_index->{$sel->{-as}} = $sel
1547 if (ref $sel eq 'HASH' and $sel->{-as});
1550 # anything from the original select mentioned on the group-by needs to make it to the inner selector
1551 # also look for named aggregates referred in the having clause
1552 # having often contains scalarrefs - thus parse it out entirely
1554 if ($attrs->{having}) {
1555 local $sql_maker->{having_bind};
1556 local $sql_maker->{quote_char} = $sql_maker->{quote_char};
1557 local $sql_maker->{name_sep} = $sql_maker->{name_sep};
1558 unless (defined $sql_maker->{quote_char} and length $sql_maker->{quote_char}) {
1559 $sql_maker->{quote_char} = [ "\x00", "\xFF" ];
1560 # if we don't unset it we screw up retarded but unfortunately working
1561 # 'MAX(foo.bar)' => { '>', 3 }
1562 $sql_maker->{name_sep} = '';
1565 my ($lquote, $rquote, $sep) = map { quotemeta $_ } ($sql_maker->_quote_chars, $sql_maker->name_sep);
1567 my $sql = $sql_maker->_parse_rs_attrs ({ having => $attrs->{having} });
1569 # search for both a proper quoted qualified string, for a naive unquoted scalarref
1570 # and if all fails for an utterly naive quoted scalar-with-function
1572 $rquote $sep $lquote (.+?) $rquote
1574 [\s,] \w+ \. (\w+) [\s,]
1576 [\s,] $lquote (.+?) $rquote [\s,]
1578 push @parts, ($1 || $2 || $3); # one of them matched if we got here
1583 my $colpiece = $sel_index->{$_} || $_;
1585 # unqualify join-based group_by's. Arcane but possible query
1586 # also horrible horrible hack to alias a column (not a func.)
1587 # (probably need to introduce SQLA syntax)
1588 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1591 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1593 push @{$sub_attrs->{select}}, $colpiece;
1597 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1598 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1601 return $rsrc->resultset_class
1602 ->new ($rsrc, $sub_attrs)
1604 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1605 ->get_column ('count');
1612 =head2 count_literal
1616 =item Arguments: $sql_fragment, @bind_values
1618 =item Return Value: $count
1622 Counts the results in a literal query. Equivalent to calling L</search_literal>
1623 with the passed arguments, then L</count>.
1627 sub count_literal { shift->search_literal(@_)->count; }
1633 =item Arguments: none
1635 =item Return Value: @objects
1639 Returns all elements in the resultset.
1646 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1649 delete $self->{stashed_rows};
1650 delete $self->{stashed_objects};
1652 if (my $c = $self->get_cache) {
1656 $self->cursor->reset;
1658 my $objs = $self->_construct_objects('fetch_all') || [];
1660 $self->set_cache($objs) if $self->{attrs}{cache};
1669 =item Arguments: none
1671 =item Return Value: $self
1675 Resets the resultset's cursor, so you can iterate through the elements again.
1676 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1683 delete $self->{_attrs};
1684 delete $self->{stashed_rows};
1685 delete $self->{stashed_objects};
1687 $self->{all_cache_position} = 0;
1688 $self->cursor->reset;
1696 =item Arguments: none
1698 =item Return Value: $object | undef
1702 Resets the resultset and returns an object for the first result (or C<undef>
1703 if the resultset is empty).
1708 return $_[0]->reset->next;
1714 # Determines whether and what type of subquery is required for the $rs operation.
1715 # If grouping is necessary either supplies its own, or verifies the current one
1716 # After all is done delegates to the proper storage method.
1718 sub _rs_update_delete {
1719 my ($self, $op, $values) = @_;
1721 my $cond = $self->{cond};
1722 my $rsrc = $self->result_source;
1723 my $storage = $rsrc->schema->storage;
1725 my $attrs = { %{$self->_resolved_attrs} };
1727 # "needs" is a strong word here - if the subquery is part of an IN clause - no point of
1728 # even adding the group_by. It will really be used only when composing a poor-man's
1729 # multicolumn-IN equivalent OR set
1730 my $needs_group_by_subq = defined $attrs->{group_by};
1732 # simplify the joinmap and maybe decide if a grouping (and thus subquery) is necessary
1733 my $relation_classifications;
1734 if (ref($attrs->{from}) eq 'ARRAY') {
1735 $attrs->{from} = $storage->_prune_unused_joins ($attrs->{from}, $attrs->{select}, $cond, $attrs);
1737 $relation_classifications = $storage->_resolve_aliastypes_from_select_args (
1738 [ @{$attrs->{from}}[1 .. $#{$attrs->{from}}] ],
1742 ) unless $needs_group_by_subq; # we already know we need a group, no point of resolving them
1745 $needs_group_by_subq ||= 1; # if {from} is unparseable assume the worst
1748 $needs_group_by_subq ||= exists $relation_classifications->{multiplying};
1750 # if no subquery - life is easy-ish
1752 $needs_group_by_subq
1754 keys %$relation_classifications # if any joins at all - need to wrap a subq
1756 $self->_has_resolved_attr(qw/rows offset/) # limits call for a subq
1758 # Most databases do not allow aliasing of tables in UPDATE/DELETE. Thus
1759 # a condition containing 'me' or other table prefixes will not work
1760 # at all. What this code tries to do (badly) is to generate a condition
1761 # with the qualifiers removed, by exploiting the quote mechanism of sqla
1763 # this is atrocious and should be replaced by normal sqla introspection
1765 my ($sql, @bind) = do {
1766 my $sqla = $rsrc->storage->sql_maker;
1767 local $sqla->{_dequalify_idents} = 1;
1768 $sqla->_recurse_where($self->{cond});
1771 return $rsrc->storage->$op(
1773 $op eq 'update' ? $values : (),
1774 $self->{cond} ? \[$sql, @bind] : (),
1778 # we got this far - means it is time to wrap a subquery
1779 my $idcols = $rsrc->_identifying_column_set || $self->throw_exception(
1781 "Unable to perform complex resultset %s() without an identifying set of columns on source '%s'",
1786 my $existing_group_by = delete $attrs->{group_by};
1788 # make a new $rs selecting only the PKs (that's all we really need for the subq)
1789 delete $attrs->{$_} for qw/collapse select _prefetch_selector_range as/;
1790 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } @$idcols ];
1791 $attrs->{group_by} = \ ''; # FIXME - this is an evil hack, it causes the optimiser to kick in and throw away the LEFT joins
1792 my $subrs = (ref $self)->new($rsrc, $attrs);
1794 if (@$idcols == 1) {
1795 return $storage->$op (
1797 $op eq 'update' ? $values : (),
1798 { $idcols->[0] => { -in => $subrs->as_query } },
1801 elsif ($storage->_use_multicolumn_in) {
1802 # This is hideously ugly, but SQLA does not understand multicol IN expressions
1803 my $sql_maker = $storage->sql_maker;
1804 my ($sql, @bind) = @${$subrs->as_query};
1805 $sql = sprintf ('(%s) IN %s', # the as_query already comes with a set of parenthesis
1806 join (', ', map { $sql_maker->_quote ($_) } @$idcols),
1810 return $storage->$op (
1812 $op eq 'update' ? $values : (),
1817 # if all else fails - get all primary keys and operate over a ORed set
1818 # wrap in a transaction for consistency
1819 # this is where the group_by starts to matter
1821 if ($needs_group_by_subq) {
1822 $subq_group_by = $attrs->{columns};
1824 # make sure if there is a supplied group_by it matches the columns compiled above
1825 # perfectly. Anything else can not be sanely executed on most databases so croak
1826 # right then and there
1827 if ($existing_group_by) {
1828 my @current_group_by = map
1829 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1834 join ("\x00", sort @current_group_by)
1836 join ("\x00", sort @$subq_group_by )
1838 $self->throw_exception (
1839 "You have just attempted a $op operation on a resultset which does group_by"
1840 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1841 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1842 . ' kind of queries. Please retry the operation with a modified group_by or'
1843 . ' without using one at all.'
1849 my $guard = $storage->txn_scope_guard;
1852 for my $row ($subrs->search({}, { group_by => $subq_group_by })->cursor->all) {
1853 push @op_condition, { map
1854 { $idcols->[$_] => $row->[$_] }
1859 my $res = $storage->$op (
1861 $op eq 'update' ? $values : (),
1875 =item Arguments: \%values
1877 =item Return Value: $storage_rv
1881 Sets the specified columns in the resultset to the supplied values in a
1882 single query. Note that this will not run any accessor/set_column/update
1883 triggers, nor will it update any row object instances derived from this
1884 resultset (this includes the contents of the L<resultset cache|/set_cache>
1885 if any). See L</update_all> if you need to execute any on-update
1886 triggers or cascades defined either by you or a
1887 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
1889 The return value is a pass through of what the underlying
1890 storage backend returned, and may vary. See L<DBI/execute> for the most
1895 Note that L</update> does not process/deflate any of the values passed in.
1896 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
1897 ensure manually that any value passed to this method will stringify to
1898 something the RDBMS knows how to deal with. A notable example is the
1899 handling of L<DateTime> objects, for more info see:
1900 L<DBIx::Class::Manual::Cookbook/Formatting DateTime objects in queries>.
1905 my ($self, $values) = @_;
1906 $self->throw_exception('Values for update must be a hash')
1907 unless ref $values eq 'HASH';
1909 return $self->_rs_update_delete ('update', $values);
1916 =item Arguments: \%values
1918 =item Return Value: 1
1922 Fetches all objects and updates them one at a time via
1923 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
1924 triggers, while L</update> will not.
1929 my ($self, $values) = @_;
1930 $self->throw_exception('Values for update_all must be a hash')
1931 unless ref $values eq 'HASH';
1933 my $guard = $self->result_source->schema->txn_scope_guard;
1934 $_->update({%$values}) for $self->all; # shallow copy - update will mangle it
1943 =item Arguments: none
1945 =item Return Value: $storage_rv
1949 Deletes the rows matching this resultset in a single query. Note that this
1950 will not run any delete triggers, nor will it alter the
1951 L<in_storage|DBIx::Class::Row/in_storage> status of any row object instances
1952 derived from this resultset (this includes the contents of the
1953 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
1954 execute any on-delete triggers or cascades defined either by you or a
1955 L<result component|DBIx::Class::Manual::Component/WHAT IS A COMPONENT>.
1957 The return value is a pass through of what the underlying storage backend
1958 returned, and may vary. See L<DBI/execute> for the most common case.
1964 $self->throw_exception('delete does not accept any arguments')
1967 return $self->_rs_update_delete ('delete');
1974 =item Arguments: none
1976 =item Return Value: 1
1980 Fetches all objects and deletes them one at a time via
1981 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
1982 triggers, while L</delete> will not.
1988 $self->throw_exception('delete_all does not accept any arguments')
1991 my $guard = $self->result_source->schema->txn_scope_guard;
1992 $_->delete for $self->all;
2001 =item Arguments: \@data;
2005 Accepts either an arrayref of hashrefs or alternatively an arrayref of arrayrefs.
2006 For the arrayref of hashrefs style each hashref should be a structure suitable
2007 for submitting to a $resultset->create(...) method.
2009 In void context, C<insert_bulk> in L<DBIx::Class::Storage::DBI> is used
2010 to insert the data, as this is a faster method.
2012 Otherwise, each set of data is inserted into the database using
2013 L<DBIx::Class::ResultSet/create>, and the resulting objects are
2014 accumulated into an array. The array itself, or an array reference
2015 is returned depending on scalar or list context.
2017 Example: Assuming an Artist Class that has many CDs Classes relating:
2019 my $Artist_rs = $schema->resultset("Artist");
2021 ## Void Context Example
2022 $Artist_rs->populate([
2023 { artistid => 4, name => 'Manufactured Crap', cds => [
2024 { title => 'My First CD', year => 2006 },
2025 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2028 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
2029 { title => 'My parents sold me to a record company', year => 2005 },
2030 { title => 'Why Am I So Ugly?', year => 2006 },
2031 { title => 'I Got Surgery and am now Popular', year => 2007 }
2036 ## Array Context Example
2037 my ($ArtistOne, $ArtistTwo, $ArtistThree) = $Artist_rs->populate([
2038 { name => "Artist One"},
2039 { name => "Artist Two"},
2040 { name => "Artist Three", cds=> [
2041 { title => "First CD", year => 2007},
2042 { title => "Second CD", year => 2008},
2046 print $ArtistOne->name; ## response is 'Artist One'
2047 print $ArtistThree->cds->count ## reponse is '2'
2049 For the arrayref of arrayrefs style, the first element should be a list of the
2050 fieldsnames to which the remaining elements are rows being inserted. For
2053 $Arstist_rs->populate([
2054 [qw/artistid name/],
2055 [100, 'A Formally Unknown Singer'],
2056 [101, 'A singer that jumped the shark two albums ago'],
2057 [102, 'An actually cool singer'],
2060 Please note an important effect on your data when choosing between void and
2061 wantarray context. Since void context goes straight to C<insert_bulk> in
2062 L<DBIx::Class::Storage::DBI> this will skip any component that is overriding
2063 C<insert>. So if you are using something like L<DBIx-Class-UUIDColumns> to
2064 create primary keys for you, you will find that your PKs are empty. In this
2065 case you will have to use the wantarray context in order to create those
2073 # cruft placed in standalone method
2074 my $data = $self->_normalize_populate_args(@_);
2076 return unless @$data;
2078 if(defined wantarray) {
2080 foreach my $item (@$data) {
2081 push(@created, $self->create($item));
2083 return wantarray ? @created : \@created;
2086 my $first = $data->[0];
2088 # if a column is a registered relationship, and is a non-blessed hash/array, consider
2089 # it relationship data
2090 my (@rels, @columns);
2091 my $rsrc = $self->result_source;
2092 my $rels = { map { $_ => $rsrc->relationship_info($_) } $rsrc->relationships };
2093 for (keys %$first) {
2094 my $ref = ref $first->{$_};
2095 $rels->{$_} && ($ref eq 'ARRAY' or $ref eq 'HASH')
2101 my @pks = $rsrc->primary_columns;
2103 ## do the belongs_to relationships
2104 foreach my $index (0..$#$data) {
2106 # delegate to create() for any dataset without primary keys with specified relationships
2107 if (grep { !defined $data->[$index]->{$_} } @pks ) {
2109 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
2110 my @ret = $self->populate($data);
2116 foreach my $rel (@rels) {
2117 next unless ref $data->[$index]->{$rel} eq "HASH";
2118 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
2119 my ($reverse_relname, $reverse_relinfo) = %{$rsrc->reverse_relationship_info($rel)};
2120 my $related = $result->result_source->_resolve_condition(
2121 $reverse_relinfo->{cond},
2127 delete $data->[$index]->{$rel};
2128 $data->[$index] = {%{$data->[$index]}, %$related};
2130 push @columns, keys %$related if $index == 0;
2134 ## inherit the data locked in the conditions of the resultset
2135 my ($rs_data) = $self->_merge_with_rscond({});
2136 delete @{$rs_data}{@columns};
2137 my @inherit_cols = keys %$rs_data;
2138 my @inherit_data = values %$rs_data;
2140 ## do bulk insert on current row
2141 $rsrc->storage->insert_bulk(
2143 [@columns, @inherit_cols],
2144 [ map { [ @$_{@columns}, @inherit_data ] } @$data ],
2147 ## do the has_many relationships
2148 foreach my $item (@$data) {
2152 foreach my $rel (@rels) {
2153 next unless ref $item->{$rel} eq "ARRAY" && @{ $item->{$rel} };
2155 $main_row ||= $self->new_result({map { $_ => $item->{$_} } @pks});
2157 my $child = $main_row->$rel;
2159 my $related = $child->result_source->_resolve_condition(
2160 $rels->{$rel}{cond},
2166 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
2167 my @populate = map { {%$_, %$related} } @rows_to_add;
2169 $child->populate( \@populate );
2176 # populate() argumnets went over several incarnations
2177 # What we ultimately support is AoH
2178 sub _normalize_populate_args {
2179 my ($self, $arg) = @_;
2181 if (ref $arg eq 'ARRAY') {
2185 elsif (ref $arg->[0] eq 'HASH') {
2188 elsif (ref $arg->[0] eq 'ARRAY') {
2190 my @colnames = @{$arg->[0]};
2191 foreach my $values (@{$arg}[1 .. $#$arg]) {
2192 push @ret, { map { $colnames[$_] => $values->[$_] } (0 .. $#colnames) };
2198 $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
2205 =item Arguments: none
2207 =item Return Value: $pager
2211 Return Value a L<Data::Page> object for the current resultset. Only makes
2212 sense for queries with a C<page> attribute.
2214 To get the full count of entries for a paged resultset, call
2215 C<total_entries> on the L<Data::Page> object.
2222 return $self->{pager} if $self->{pager};
2224 my $attrs = $self->{attrs};
2225 if (!defined $attrs->{page}) {
2226 $self->throw_exception("Can't create pager for non-paged rs");
2228 elsif ($attrs->{page} <= 0) {
2229 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
2231 $attrs->{rows} ||= 10;
2233 # throw away the paging flags and re-run the count (possibly
2234 # with a subselect) to get the real total count
2235 my $count_attrs = { %$attrs };
2236 delete $count_attrs->{$_} for qw/rows offset page pager/;
2238 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
2240 require DBIx::Class::ResultSet::Pager;
2241 return $self->{pager} = DBIx::Class::ResultSet::Pager->new(
2242 sub { $total_rs->count }, #lazy-get the total
2244 $self->{attrs}{page},
2252 =item Arguments: $page_number
2254 =item Return Value: $rs
2258 Returns a resultset for the $page_number page of the resultset on which page
2259 is called, where each page contains a number of rows equal to the 'rows'
2260 attribute set on the resultset (10 by default).
2265 my ($self, $page) = @_;
2266 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2273 =item Arguments: \%vals
2275 =item Return Value: $rowobject
2279 Creates a new row object in the resultset's result class and returns
2280 it. The row is not inserted into the database at this point, call
2281 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2282 will tell you whether the row object has been inserted or not.
2284 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2289 my ($self, $values) = @_;
2290 $self->throw_exception( "new_result needs a hash" )
2291 unless (ref $values eq 'HASH');
2293 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2297 @$cols_from_relations
2298 ? (-cols_from_relations => $cols_from_relations)
2300 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2303 return $self->result_class->new(\%new);
2306 # _merge_with_rscond
2308 # Takes a simple hash of K/V data and returns its copy merged with the
2309 # condition already present on the resultset. Additionally returns an
2310 # arrayref of value/condition names, which were inferred from related
2311 # objects (this is needed for in-memory related objects)
2312 sub _merge_with_rscond {
2313 my ($self, $data) = @_;
2315 my (%new_data, @cols_from_relations);
2317 my $alias = $self->{attrs}{alias};
2319 if (! defined $self->{cond}) {
2320 # just massage $data below
2322 elsif ($self->{cond} eq $DBIx::Class::ResultSource::UNRESOLVABLE_CONDITION) {
2323 %new_data = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
2324 @cols_from_relations = keys %new_data;
2326 elsif (ref $self->{cond} ne 'HASH') {
2327 $self->throw_exception(
2328 "Can't abstract implicit construct, resultset condition not a hash"
2332 # precendence must be given to passed values over values inherited from
2333 # the cond, so the order here is important.
2334 my $collapsed_cond = $self->_collapse_cond($self->{cond});
2335 my %implied = %{$self->_remove_alias($collapsed_cond, $alias)};
2337 while ( my($col, $value) = each %implied ) {
2338 my $vref = ref $value;
2344 (keys %$value)[0] eq '='
2346 $new_data{$col} = $value->{'='};
2348 elsif( !$vref or $vref eq 'SCALAR' or blessed($value) ) {
2349 $new_data{$col} = $value;
2356 %{ $self->_remove_alias($data, $alias) },
2359 return (\%new_data, \@cols_from_relations);
2362 # _has_resolved_attr
2364 # determines if the resultset defines at least one
2365 # of the attributes supplied
2367 # used to determine if a subquery is neccessary
2369 # supports some virtual attributes:
2371 # This will scan for any joins being present on the resultset.
2372 # It is not a mere key-search but a deep inspection of {from}
2375 sub _has_resolved_attr {
2376 my ($self, @attr_names) = @_;
2378 my $attrs = $self->_resolved_attrs;
2382 for my $n (@attr_names) {
2383 if (grep { $n eq $_ } (qw/-join/) ) {
2384 $extra_checks{$n}++;
2388 my $attr = $attrs->{$n};
2390 next if not defined $attr;
2392 if (ref $attr eq 'HASH') {
2393 return 1 if keys %$attr;
2395 elsif (ref $attr eq 'ARRAY') {
2403 # a resolved join is expressed as a multi-level from
2405 $extra_checks{-join}
2407 ref $attrs->{from} eq 'ARRAY'
2409 @{$attrs->{from}} > 1
2417 # Recursively collapse the condition.
2419 sub _collapse_cond {
2420 my ($self, $cond, $collapsed) = @_;
2424 if (ref $cond eq 'ARRAY') {
2425 foreach my $subcond (@$cond) {
2426 next unless ref $subcond; # -or
2427 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2430 elsif (ref $cond eq 'HASH') {
2431 if (keys %$cond and (keys %$cond)[0] eq '-and') {
2432 foreach my $subcond (@{$cond->{-and}}) {
2433 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2437 foreach my $col (keys %$cond) {
2438 my $value = $cond->{$col};
2439 $collapsed->{$col} = $value;
2449 # Remove the specified alias from the specified query hash. A copy is made so
2450 # the original query is not modified.
2453 my ($self, $query, $alias) = @_;
2455 my %orig = %{ $query || {} };
2458 foreach my $key (keys %orig) {
2460 $unaliased{$key} = $orig{$key};
2463 $unaliased{$1} = $orig{$key}
2464 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2474 =item Arguments: none
2476 =item Return Value: \[ $sql, @bind ]
2480 Returns the SQL query and bind vars associated with the invocant.
2482 This is generally used as the RHS for a subquery.
2489 my $attrs = $self->_resolved_attrs_copy;
2494 # my ($sql, \@bind, \%dbi_bind_attrs) = _select_args_to_query (...)
2495 # $sql also has no wrapping parenthesis in list ctx
2497 my $sqlbind = $self->result_source->storage
2498 ->_select_args_to_query ($attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs);
2507 =item Arguments: \%vals, \%attrs?
2509 =item Return Value: $rowobject
2513 my $artist = $schema->resultset('Artist')->find_or_new(
2514 { artist => 'fred' }, { key => 'artists' });
2516 $cd->cd_to_producer->find_or_new({ producer => $producer },
2517 { key => 'primary });
2519 Find an existing record from this resultset using L</find>. if none exists,
2520 instantiate a new result object and return it. The object will not be saved
2521 into your storage until you call L<DBIx::Class::Row/insert> on it.
2523 You most likely want this method when looking for existing rows using a unique
2524 constraint that is not the primary key, or looking for related rows.
2526 If you want objects to be saved immediately, use L</find_or_create> instead.
2528 B<Note>: Make sure to read the documentation of L</find> and understand the
2529 significance of the C<key> attribute, as its lack may skew your search, and
2530 subsequently result in spurious new objects.
2532 B<Note>: Take care when using C<find_or_new> with a table having
2533 columns with default values that you intend to be automatically
2534 supplied by the database (e.g. an auto_increment primary key column).
2535 In normal usage, the value of such columns should NOT be included at
2536 all in the call to C<find_or_new>, even when set to C<undef>.
2542 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2543 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2544 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2547 return $self->new_result($hash);
2554 =item Arguments: \%vals
2556 =item Return Value: a L<DBIx::Class::Row> $object
2560 Attempt to create a single new row or a row with multiple related rows
2561 in the table represented by the resultset (and related tables). This
2562 will not check for duplicate rows before inserting, use
2563 L</find_or_create> to do that.
2565 To create one row for this resultset, pass a hashref of key/value
2566 pairs representing the columns of the table and the values you wish to
2567 store. If the appropriate relationships are set up, foreign key fields
2568 can also be passed an object representing the foreign row, and the
2569 value will be set to its primary key.
2571 To create related objects, pass a hashref of related-object column values
2572 B<keyed on the relationship name>. If the relationship is of type C<multi>
2573 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2574 The process will correctly identify columns holding foreign keys, and will
2575 transparently populate them from the keys of the corresponding relation.
2576 This can be applied recursively, and will work correctly for a structure
2577 with an arbitrary depth and width, as long as the relationships actually
2578 exists and the correct column data has been supplied.
2581 Instead of hashrefs of plain related data (key/value pairs), you may
2582 also pass new or inserted objects. New objects (not inserted yet, see
2583 L</new>), will be inserted into their appropriate tables.
2585 Effectively a shortcut for C<< ->new_result(\%vals)->insert >>.
2587 Example of creating a new row.
2589 $person_rs->create({
2590 name=>"Some Person",
2591 email=>"somebody@someplace.com"
2594 Example of creating a new row and also creating rows in a related C<has_many>
2595 or C<has_one> resultset. Note Arrayref.
2598 { artistid => 4, name => 'Manufactured Crap', cds => [
2599 { title => 'My First CD', year => 2006 },
2600 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2605 Example of creating a new row and also creating a row in a related
2606 C<belongs_to> resultset. Note Hashref.
2609 title=>"Music for Silly Walks",
2612 name=>"Silly Musician",
2620 When subclassing ResultSet never attempt to override this method. Since
2621 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2622 lot of the internals simply never call it, so your override will be
2623 bypassed more often than not. Override either L<new|DBIx::Class::Row/new>
2624 or L<insert|DBIx::Class::Row/insert> depending on how early in the
2625 L</create> process you need to intervene.
2632 my ($self, $attrs) = @_;
2633 $self->throw_exception( "create needs a hashref" )
2634 unless ref $attrs eq 'HASH';
2635 return $self->new_result($attrs)->insert;
2638 =head2 find_or_create
2642 =item Arguments: \%vals, \%attrs?
2644 =item Return Value: $rowobject
2648 $cd->cd_to_producer->find_or_create({ producer => $producer },
2649 { key => 'primary' });
2651 Tries to find a record based on its primary key or unique constraints; if none
2652 is found, creates one and returns that instead.
2654 my $cd = $schema->resultset('CD')->find_or_create({
2656 artist => 'Massive Attack',
2657 title => 'Mezzanine',
2661 Also takes an optional C<key> attribute, to search by a specific key or unique
2662 constraint. For example:
2664 my $cd = $schema->resultset('CD')->find_or_create(
2666 artist => 'Massive Attack',
2667 title => 'Mezzanine',
2669 { key => 'cd_artist_title' }
2672 B<Note>: Make sure to read the documentation of L</find> and understand the
2673 significance of the C<key> attribute, as its lack may skew your search, and
2674 subsequently result in spurious row creation.
2676 B<Note>: Because find_or_create() reads from the database and then
2677 possibly inserts based on the result, this method is subject to a race
2678 condition. Another process could create a record in the table after
2679 the find has completed and before the create has started. To avoid
2680 this problem, use find_or_create() inside a transaction.
2682 B<Note>: Take care when using C<find_or_create> with a table having
2683 columns with default values that you intend to be automatically
2684 supplied by the database (e.g. an auto_increment primary key column).
2685 In normal usage, the value of such columns should NOT be included at
2686 all in the call to C<find_or_create>, even when set to C<undef>.
2688 See also L</find> and L</update_or_create>. For information on how to declare
2689 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2691 If you need to know if an existing row was found or a new one created use
2692 L</find_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2693 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2696 my $cd = $schema->resultset('CD')->find_or_new({
2698 artist => 'Massive Attack',
2699 title => 'Mezzanine',
2703 if( $cd->in_storage ) {
2710 sub find_or_create {
2712 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2713 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2714 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2717 return $self->create($hash);
2720 =head2 update_or_create
2724 =item Arguments: \%col_values, { key => $unique_constraint }?
2726 =item Return Value: $row_object
2730 $resultset->update_or_create({ col => $val, ... });
2732 Like L</find_or_create>, but if a row is found it is immediately updated via
2733 C<< $found_row->update (\%col_values) >>.
2736 Takes an optional C<key> attribute to search on a specific unique constraint.
2739 # In your application
2740 my $cd = $schema->resultset('CD')->update_or_create(
2742 artist => 'Massive Attack',
2743 title => 'Mezzanine',
2746 { key => 'cd_artist_title' }
2749 $cd->cd_to_producer->update_or_create({
2750 producer => $producer,
2756 B<Note>: Make sure to read the documentation of L</find> and understand the
2757 significance of the C<key> attribute, as its lack may skew your search, and
2758 subsequently result in spurious row creation.
2760 B<Note>: Take care when using C<update_or_create> with a table having
2761 columns with default values that you intend to be automatically
2762 supplied by the database (e.g. an auto_increment primary key column).
2763 In normal usage, the value of such columns should NOT be included at
2764 all in the call to C<update_or_create>, even when set to C<undef>.
2766 See also L</find> and L</find_or_create>. For information on how to declare
2767 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2769 If you need to know if an existing row was updated or a new one created use
2770 L</update_or_new> and L<DBIx::Class::Row/in_storage> instead. Don't forget
2771 to call L<DBIx::Class::Row/insert> to save the newly created row to the
2774 my $cd = $schema->resultset('CD')->update_or_new(
2776 artist => 'Massive Attack',
2777 title => 'Mezzanine',
2780 { key => 'cd_artist_title' }
2783 if( $cd->in_storage ) {
2790 sub update_or_create {
2792 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2793 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2795 my $row = $self->find($cond, $attrs);
2797 $row->update($cond);
2801 return $self->create($cond);
2804 =head2 update_or_new
2808 =item Arguments: \%col_values, { key => $unique_constraint }?
2810 =item Return Value: $rowobject
2814 $resultset->update_or_new({ col => $val, ... });
2816 Like L</find_or_new> but if a row is found it is immediately updated via
2817 C<< $found_row->update (\%col_values) >>.
2821 # In your application
2822 my $cd = $schema->resultset('CD')->update_or_new(
2824 artist => 'Massive Attack',
2825 title => 'Mezzanine',
2828 { key => 'cd_artist_title' }
2831 if ($cd->in_storage) {
2832 # the cd was updated
2835 # the cd is not yet in the database, let's insert it
2839 B<Note>: Make sure to read the documentation of L</find> and understand the
2840 significance of the C<key> attribute, as its lack may skew your search, and
2841 subsequently result in spurious new objects.
2843 B<Note>: Take care when using C<update_or_new> with a table having
2844 columns with default values that you intend to be automatically
2845 supplied by the database (e.g. an auto_increment primary key column).
2846 In normal usage, the value of such columns should NOT be included at
2847 all in the call to C<update_or_new>, even when set to C<undef>.
2849 See also L</find>, L</find_or_create> and L</find_or_new>.
2855 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
2856 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2858 my $row = $self->find( $cond, $attrs );
2859 if ( defined $row ) {
2860 $row->update($cond);
2864 return $self->new_result($cond);
2871 =item Arguments: none
2873 =item Return Value: \@cache_objects | undef
2877 Gets the contents of the cache for the resultset, if the cache is set.
2879 The cache is populated either by using the L</prefetch> attribute to
2880 L</search> or by calling L</set_cache>.
2892 =item Arguments: \@cache_objects
2894 =item Return Value: \@cache_objects
2898 Sets the contents of the cache for the resultset. Expects an arrayref
2899 of objects of the same class as those produced by the resultset. Note that
2900 if the cache is set the resultset will return the cached objects rather
2901 than re-querying the database even if the cache attr is not set.
2903 The contents of the cache can also be populated by using the
2904 L</prefetch> attribute to L</search>.
2909 my ( $self, $data ) = @_;
2910 $self->throw_exception("set_cache requires an arrayref")
2911 if defined($data) && (ref $data ne 'ARRAY');
2912 $self->{all_cache} = $data;
2919 =item Arguments: none
2921 =item Return Value: undef
2925 Clears the cache for the resultset.
2930 shift->set_cache(undef);
2937 =item Arguments: none
2939 =item Return Value: true, if the resultset has been paginated
2947 return !!$self->{attrs}{page};
2954 =item Arguments: none
2956 =item Return Value: true, if the resultset has been ordered with C<order_by>.
2964 return scalar $self->result_source->storage->_extract_order_criteria($self->{attrs}{order_by});
2967 =head2 related_resultset
2971 =item Arguments: $relationship_name
2973 =item Return Value: $resultset
2977 Returns a related resultset for the supplied relationship name.
2979 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
2983 sub related_resultset {
2984 my ($self, $rel) = @_;
2986 $self->{related_resultsets} ||= {};
2987 return $self->{related_resultsets}{$rel} ||= do {
2988 my $rsrc = $self->result_source;
2989 my $rel_info = $rsrc->relationship_info($rel);
2991 $self->throw_exception(
2992 "search_related: result source '" . $rsrc->source_name .
2993 "' has no such relationship $rel")
2996 my $attrs = $self->_chain_relationship($rel);
2998 my $join_count = $attrs->{seen_join}{$rel};
3000 my $alias = $self->result_source->storage
3001 ->relname_to_table_alias($rel, $join_count);
3003 # since this is search_related, and we already slid the select window inwards
3004 # (the select/as attrs were deleted in the beginning), we need to flip all
3005 # left joins to inner, so we get the expected results
3006 # read the comment on top of the actual function to see what this does
3007 $attrs->{from} = $rsrc->schema->storage->_inner_join_to_node ($attrs->{from}, $alias);
3010 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
3011 delete @{$attrs}{qw(result_class alias)};
3015 if (my $cache = $self->get_cache) {
3016 if ($cache->[0] && $cache->[0]->related_resultset($rel)->get_cache) {
3017 $new_cache = [ map { @{$_->related_resultset($rel)->get_cache||[]} }
3022 my $rel_source = $rsrc->related_source($rel);
3026 # The reason we do this now instead of passing the alias to the
3027 # search_rs below is that if you wrap/overload resultset on the
3028 # source you need to know what alias it's -going- to have for things
3029 # to work sanely (e.g. RestrictWithObject wants to be able to add
3030 # extra query restrictions, and these may need to be $alias.)
3032 my $rel_attrs = $rel_source->resultset_attributes;
3033 local $rel_attrs->{alias} = $alias;
3035 $rel_source->resultset
3039 where => $attrs->{where},
3042 $new->set_cache($new_cache) if $new_cache;
3047 =head2 current_source_alias
3051 =item Arguments: none
3053 =item Return Value: $source_alias
3057 Returns the current table alias for the result source this resultset is built
3058 on, that will be used in the SQL query. Usually it is C<me>.
3060 Currently the source alias that refers to the result set returned by a
3061 L</search>/L</find> family method depends on how you got to the resultset: it's
3062 C<me> by default, but eg. L</search_related> aliases it to the related result
3063 source name (and keeps C<me> referring to the original result set). The long
3064 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
3065 (and make this method unnecessary).
3067 Thus it's currently necessary to use this method in predefined queries (see
3068 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
3069 source alias of the current result set:
3071 # in a result set class
3073 my ($self, $user) = @_;
3075 my $me = $self->current_source_alias;
3077 return $self->search({
3078 "$me.modified" => $user->id,
3084 sub current_source_alias {
3087 return ($self->{attrs} || {})->{alias} || 'me';
3090 =head2 as_subselect_rs
3094 =item Arguments: none
3096 =item Return Value: $resultset
3100 Act as a barrier to SQL symbols. The resultset provided will be made into a
3101 "virtual view" by including it as a subquery within the from clause. From this
3102 point on, any joined tables are inaccessible to ->search on the resultset (as if
3103 it were simply where-filtered without joins). For example:
3105 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
3107 # 'x' now pollutes the query namespace
3109 # So the following works as expected
3110 my $ok_rs = $rs->search({'x.other' => 1});
3112 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
3113 # def) we look for one row with contradictory terms and join in another table
3114 # (aliased 'x_2') which we never use
3115 my $broken_rs = $rs->search({'x.name' => 'def'});
3117 my $rs2 = $rs->as_subselect_rs;
3119 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
3120 my $not_joined_rs = $rs2->search({'x.other' => 1});
3122 # works as expected: finds a 'table' row related to two x rows (abc and def)
3123 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
3125 Another example of when one might use this would be to select a subset of
3126 columns in a group by clause:
3128 my $rs = $schema->resultset('Bar')->search(undef, {
3129 group_by => [qw{ id foo_id baz_id }],
3130 })->as_subselect_rs->search(undef, {
3131 columns => [qw{ id foo_id }]
3134 In the above example normally columns would have to be equal to the group by,
3135 but because we isolated the group by into a subselect the above works.
3139 sub as_subselect_rs {
3142 my $attrs = $self->_resolved_attrs;
3144 my $fresh_rs = (ref $self)->new (
3145 $self->result_source
3148 # these pieces will be locked in the subquery
3149 delete $fresh_rs->{cond};
3150 delete @{$fresh_rs->{attrs}}{qw/where bind/};
3152 return $fresh_rs->search( {}, {
3154 $attrs->{alias} => $self->as_query,
3155 -alias => $attrs->{alias},
3156 -rsrc => $self->result_source,
3158 alias => $attrs->{alias},
3162 # This code is called by search_related, and makes sure there
3163 # is clear separation between the joins before, during, and
3164 # after the relationship. This information is needed later
3165 # in order to properly resolve prefetch aliases (any alias
3166 # with a relation_chain_depth less than the depth of the
3167 # current prefetch is not considered)
3169 # The increments happen twice per join. An even number means a
3170 # relationship specified via a search_related, whereas an odd
3171 # number indicates a join/prefetch added via attributes
3173 # Also this code will wrap the current resultset (the one we
3174 # chain to) in a subselect IFF it contains limiting attributes
3175 sub _chain_relationship {
3176 my ($self, $rel) = @_;
3177 my $source = $self->result_source;
3178 my $attrs = { %{$self->{attrs}||{}} };
3180 # we need to take the prefetch the attrs into account before we
3181 # ->_resolve_join as otherwise they get lost - captainL
3182 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
3184 delete @{$attrs}{qw/join prefetch collapse group_by distinct select as columns +select +as +columns/};
3186 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
3189 my @force_subq_attrs = qw/offset rows group_by having/;
3192 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
3194 $self->_has_resolved_attr (@force_subq_attrs)
3196 # Nuke the prefetch (if any) before the new $rs attrs
3197 # are resolved (prefetch is useless - we are wrapping
3198 # a subquery anyway).
3199 my $rs_copy = $self->search;
3200 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
3201 $rs_copy->{attrs}{join},
3202 delete $rs_copy->{attrs}{prefetch},
3207 -alias => $attrs->{alias},
3208 $attrs->{alias} => $rs_copy->as_query,
3210 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
3211 $seen->{-relation_chain_depth} = 0;
3213 elsif ($attrs->{from}) { #shallow copy suffices
3214 $from = [ @{$attrs->{from}} ];
3219 -alias => $attrs->{alias},
3220 $attrs->{alias} => $source->from,
3224 my $jpath = ($seen->{-relation_chain_depth})
3225 ? $from->[-1][0]{-join_path}
3228 my @requested_joins = $source->_resolve_join(
3235 push @$from, @requested_joins;
3237 $seen->{-relation_chain_depth}++;
3239 # if $self already had a join/prefetch specified on it, the requested
3240 # $rel might very well be already included. What we do in this case
3241 # is effectively a no-op (except that we bump up the chain_depth on
3242 # the join in question so we could tell it *is* the search_related)
3245 # we consider the last one thus reverse
3246 for my $j (reverse @requested_joins) {
3247 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
3248 if ($rel eq $last_j) {
3249 $j->[0]{-relation_chain_depth}++;
3255 unless ($already_joined) {
3256 push @$from, $source->_resolve_join(
3264 $seen->{-relation_chain_depth}++;
3266 return {%$attrs, from => $from, seen_join => $seen};
3269 # too many times we have to do $attrs = { %{$self->_resolved_attrs} }
3270 sub _resolved_attrs_copy {
3272 return { %{$self->_resolved_attrs (@_)} };
3275 sub _resolved_attrs {
3277 return $self->{_attrs} if $self->{_attrs};
3279 my $attrs = { %{ $self->{attrs} || {} } };
3280 my $source = $self->result_source;
3281 my $alias = $attrs->{alias};
3283 # default selection list
3284 $attrs->{columns} = [ $source->columns ]
3285 unless List::Util::first { exists $attrs->{$_} } qw/columns cols select as/;
3287 # merge selectors together
3288 for (qw/columns select as/) {
3289 $attrs->{$_} = $self->_merge_attr($attrs->{$_}, delete $attrs->{"+$_"})
3290 if $attrs->{$_} or $attrs->{"+$_"};
3293 # disassemble columns
3295 if (my $cols = delete $attrs->{columns}) {
3296 for my $c (ref $cols eq 'ARRAY' ? @$cols : $cols) {
3297 if (ref $c eq 'HASH') {
3298 for my $as (keys %$c) {
3299 push @sel, $c->{$as};
3310 # when trying to weed off duplicates later do not go past this point -
3311 # everything added from here on is unbalanced "anyone's guess" stuff
3312 my $dedup_stop_idx = $#as;
3314 push @as, @{ ref $attrs->{as} eq 'ARRAY' ? $attrs->{as} : [ $attrs->{as} ] }
3316 push @sel, @{ ref $attrs->{select} eq 'ARRAY' ? $attrs->{select} : [ $attrs->{select} ] }
3317 if $attrs->{select};
3319 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3320 $_ = (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" for @sel;
3322 # disqualify all $alias.col as-bits (inflate-map mandated)
3323 $_ = ($_ =~ /^\Q$alias.\E(.+)$/) ? $1 : $_ for @as;
3325 # de-duplicate the result (remove *identical* select/as pairs)
3326 # and also die on duplicate {as} pointing to different {select}s
3327 # not using a c-style for as the condition is prone to shrinkage
3330 while ($i <= $dedup_stop_idx) {
3331 if ($seen->{"$sel[$i] \x00\x00 $as[$i]"}++) {
3336 elsif ($seen->{$as[$i]}++) {
3337 $self->throw_exception(
3338 "inflate_result() alias '$as[$i]' specified twice with different SQL-side {select}-ors"
3346 $attrs->{select} = \@sel;
3347 $attrs->{as} = \@as;
3349 $attrs->{from} ||= [{
3351 -alias => $self->{attrs}{alias},
3352 $self->{attrs}{alias} => $source->from,
3355 if ( $attrs->{join} || $attrs->{prefetch} ) {
3357 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3358 if ref $attrs->{from} ne 'ARRAY';
3360 my $join = (delete $attrs->{join}) || {};
3362 if ( defined $attrs->{prefetch} ) {
3363 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3366 $attrs->{from} = # have to copy here to avoid corrupting the original
3368 @{ $attrs->{from} },
3369 $source->_resolve_join(
3372 { %{ $attrs->{seen_join} || {} } },
3373 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3374 ? $attrs->{from}[-1][0]{-join_path}
3381 if ( defined $attrs->{order_by} ) {
3382 $attrs->{order_by} = (
3383 ref( $attrs->{order_by} ) eq 'ARRAY'
3384 ? [ @{ $attrs->{order_by} } ]
3385 : [ $attrs->{order_by} || () ]
3389 if ($attrs->{group_by} and ref $attrs->{group_by} ne 'ARRAY') {
3390 $attrs->{group_by} = [ $attrs->{group_by} ];
3393 # generate the distinct induced group_by early, as prefetch will be carried via a
3394 # subquery (since a group_by is present)
3395 if (delete $attrs->{distinct}) {
3396 if ($attrs->{group_by}) {
3397 carp_unique ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3400 # distinct affects only the main selection part, not what prefetch may
3402 $attrs->{group_by} = $source->storage->_group_over_selection (
3410 # generate selections based on the prefetch helper
3412 $prefetch = $self->_merge_joinpref_attr( {}, delete $attrs->{prefetch} )
3413 if defined $attrs->{prefetch};
3417 $self->throw_exception("Unable to prefetch, resultset contains an unnamed selector $attrs->{_dark_selector}{string}")
3418 if $attrs->{_dark_selector};
3420 $attrs->{collapse} = 1;
3422 # this is a separate structure (we don't look in {from} directly)
3423 # as the resolver needs to shift things off the lists to work
3424 # properly (identical-prefetches on different branches)
3426 if (ref $attrs->{from} eq 'ARRAY') {
3428 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3430 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3431 next unless $j->[0]{-alias};
3432 next unless $j->[0]{-join_path};
3433 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3435 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3438 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3439 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3443 my @prefetch = $source->_resolve_prefetch( $prefetch, $alias, $join_map );
3445 # we need to somehow mark which columns came from prefetch
3447 my $sel_end = $#{$attrs->{select}};
3448 $attrs->{_prefetch_selector_range} = [ $sel_end + 1, $sel_end + @prefetch ];
3451 push @{ $attrs->{select} }, (map { $_->[0] } @prefetch);
3452 push @{ $attrs->{as} }, (map { $_->[1] } @prefetch);
3455 $attrs->{_single_object_inflation} = ! List::Util::first { $_ =~ /\./ } @{$attrs->{as}};
3457 # run through the resulting joinstructure (starting from our current slot)
3458 # and unset collapse if proven unnesessary
3459 if ($attrs->{collapse} && ref $attrs->{from} eq 'ARRAY') {
3461 if (@{$attrs->{from}} > 1) {
3463 # find where our table-spec starts and consider only things after us
3464 my @fromlist = @{$attrs->{from}};
3466 my $t = shift @fromlist;
3467 $t = $t->[0] if ref $t eq 'ARRAY'; #me vs join from-spec mismatch
3468 last if ($t->{-alias} && $t->{-alias} eq $alias);
3472 $attrs->{collapse} = ! $_->[0]{-is_single}
3477 # no joins - no collapse
3478 $attrs->{collapse} = 0;
3482 if (! $attrs->{order_by} and $attrs->{collapse}) {
3483 # default order for collapsing unless the user asked for something
3484 $attrs->{order_by} = [ map { "$alias.$_" } $source->primary_columns ];
3485 $attrs->{_ordered_for_collapse} = 1;
3488 # if both page and offset are specified, produce a combined offset
3489 # even though it doesn't make much sense, this is what pre 081xx has
3491 if (my $page = delete $attrs->{page}) {
3493 ($attrs->{rows} * ($page - 1))
3495 ($attrs->{offset} || 0)
3499 return $self->{_attrs} = $attrs;
3503 my ($self, $attr) = @_;
3505 if (ref $attr eq 'HASH') {
3506 return $self->_rollout_hash($attr);
3507 } elsif (ref $attr eq 'ARRAY') {
3508 return $self->_rollout_array($attr);
3514 sub _rollout_array {
3515 my ($self, $attr) = @_;
3518 foreach my $element (@{$attr}) {
3519 if (ref $element eq 'HASH') {
3520 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3521 } elsif (ref $element eq 'ARRAY') {
3522 # XXX - should probably recurse here
3523 push( @rolled_array, @{$self->_rollout_array($element)} );
3525 push( @rolled_array, $element );
3528 return \@rolled_array;
3532 my ($self, $attr) = @_;
3535 foreach my $key (keys %{$attr}) {
3536 push( @rolled_array, { $key => $attr->{$key} } );
3538 return \@rolled_array;
3541 sub _calculate_score {
3542 my ($self, $a, $b) = @_;
3544 if (defined $a xor defined $b) {
3547 elsif (not defined $a) {
3551 if (ref $b eq 'HASH') {
3552 my ($b_key) = keys %{$b};
3553 if (ref $a eq 'HASH') {
3554 my ($a_key) = keys %{$a};
3555 if ($a_key eq $b_key) {
3556 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3561 return ($a eq $b_key) ? 1 : 0;
3564 if (ref $a eq 'HASH') {
3565 my ($a_key) = keys %{$a};
3566 return ($b eq $a_key) ? 1 : 0;
3568 return ($b eq $a) ? 1 : 0;
3573 sub _merge_joinpref_attr {
3574 my ($self, $orig, $import) = @_;
3576 return $import unless defined($orig);
3577 return $orig unless defined($import);
3579 $orig = $self->_rollout_attr($orig);
3580 $import = $self->_rollout_attr($import);
3583 foreach my $import_element ( @{$import} ) {
3584 # find best candidate from $orig to merge $b_element into
3585 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3586 foreach my $orig_element ( @{$orig} ) {
3587 my $score = $self->_calculate_score( $orig_element, $import_element );
3588 if ($score > $best_candidate->{score}) {
3589 $best_candidate->{position} = $position;
3590 $best_candidate->{score} = $score;
3594 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3595 $import_key = '' if not defined $import_key;
3597 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3598 push( @{$orig}, $import_element );
3600 my $orig_best = $orig->[$best_candidate->{position}];
3601 # merge orig_best and b_element together and replace original with merged
3602 if (ref $orig_best ne 'HASH') {
3603 $orig->[$best_candidate->{position}] = $import_element;
3604 } elsif (ref $import_element eq 'HASH') {
3605 my ($key) = keys %{$orig_best};
3606 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3609 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3620 require Hash::Merge;
3621 my $hm = Hash::Merge->new;
3623 $hm->specify_behavior({
3626 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
3628 if ($defl xor $defr) {
3629 return [ $defl ? $_[0] : $_[1] ];
3634 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
3638 return [$_[0], $_[1]];
3642 return $_[1] if !defined $_[0];
3643 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3644 return [$_[0], @{$_[1]}]
3647 return [] if !defined $_[0] and !keys %{$_[1]};
3648 return [ $_[1] ] if !defined $_[0];
3649 return [ $_[0] ] if !keys %{$_[1]};
3650 return [$_[0], $_[1]]
3655 return $_[0] if !defined $_[1];
3656 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3657 return [@{$_[0]}, $_[1]]
3660 my @ret = @{$_[0]} or return $_[1];
3661 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
3662 my %idx = map { $_ => 1 } @ret;
3663 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
3667 return [ $_[1] ] if ! @{$_[0]};
3668 return $_[0] if !keys %{$_[1]};
3669 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3670 return [ @{$_[0]}, $_[1] ];
3675 return [] if !keys %{$_[0]} and !defined $_[1];
3676 return [ $_[0] ] if !defined $_[1];
3677 return [ $_[1] ] if !keys %{$_[0]};
3678 return [$_[0], $_[1]]
3681 return [] if !keys %{$_[0]} and !@{$_[1]};
3682 return [ $_[0] ] if !@{$_[1]};
3683 return $_[1] if !keys %{$_[0]};
3684 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3685 return [ $_[0], @{$_[1]} ];
3688 return [] if !keys %{$_[0]} and !keys %{$_[1]};
3689 return [ $_[0] ] if !keys %{$_[1]};
3690 return [ $_[1] ] if !keys %{$_[0]};
3691 return [ $_[0] ] if $_[0] eq $_[1];
3692 return [ $_[0], $_[1] ];
3695 } => 'DBIC_RS_ATTR_MERGER');
3699 return $hm->merge ($_[1], $_[2]);
3703 sub STORABLE_freeze {
3704 my ($self, $cloning) = @_;
3705 my $to_serialize = { %$self };
3707 # A cursor in progress can't be serialized (and would make little sense anyway)
3708 delete $to_serialize->{cursor};
3710 # the parser can be regenerated
3711 delete $to_serialize->{_row_parser};
3713 # nor is it sensical to store a not-yet-fired-count pager
3714 if ($to_serialize->{pager} and ref $to_serialize->{pager}{total_entries} eq 'CODE') {
3715 delete $to_serialize->{pager};
3718 Storable::nfreeze($to_serialize);
3721 # need this hook for symmetry
3723 my ($self, $cloning, $serialized) = @_;
3725 %$self = %{ Storable::thaw($serialized) };
3731 =head2 throw_exception
3733 See L<DBIx::Class::Schema/throw_exception> for details.
3737 sub throw_exception {
3740 if (ref $self and my $rsrc = $self->result_source) {
3741 $rsrc->throw_exception(@_)
3744 DBIx::Class::Exception->throw(@_);
3748 # XXX: FIXME: Attributes docs need clearing up
3752 Attributes are used to refine a ResultSet in various ways when
3753 searching for data. They can be passed to any method which takes an
3754 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3757 These are in no particular order:
3763 =item Value: ( $order_by | \@order_by | \%order_by )
3767 Which column(s) to order the results by.
3769 [The full list of suitable values is documented in
3770 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3773 If a single column name, or an arrayref of names is supplied, the
3774 argument is passed through directly to SQL. The hashref syntax allows
3775 for connection-agnostic specification of ordering direction:
3777 For descending order:
3779 order_by => { -desc => [qw/col1 col2 col3/] }
3781 For explicit ascending order:
3783 order_by => { -asc => 'col' }
3785 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3786 supported, although you are strongly encouraged to use the hashref
3787 syntax as outlined above.
3793 =item Value: \@columns
3797 Shortcut to request a particular set of columns to be retrieved. Each
3798 column spec may be a string (a table column name), or a hash (in which
3799 case the key is the C<as> value, and the value is used as the C<select>
3800 expression). Adds C<me.> onto the start of any column without a C<.> in
3801 it and sets C<select> from that, then auto-populates C<as> from
3802 C<select> as normal. (You may also use the C<cols> attribute, as in
3803 earlier versions of DBIC.)
3805 Essentially C<columns> does the same as L</select> and L</as>.
3807 columns => [ 'foo', { bar => 'baz' } ]
3811 select => [qw/foo baz/],
3818 =item Value: \@columns
3822 Indicates additional columns to be selected from storage. Works the same
3823 as L</columns> but adds columns to the selection. (You may also use the
3824 C<include_columns> attribute, as in earlier versions of DBIC). For
3827 $schema->resultset('CD')->search(undef, {
3828 '+columns' => ['artist.name'],
3832 would return all CDs and include a 'name' column to the information
3833 passed to object inflation. Note that the 'artist' is the name of the
3834 column (or relationship) accessor, and 'name' is the name of the column
3835 accessor in the related table.
3837 B<NOTE:> You need to explicitly quote '+columns' when defining the attribute.
3838 Not doing so causes Perl to incorrectly interpret +columns as a bareword with a
3839 unary plus operator before it.
3841 =head2 include_columns
3845 =item Value: \@columns
3849 Deprecated. Acts as a synonym for L</+columns> for backward compatibility.
3855 =item Value: \@select_columns
3859 Indicates which columns should be selected from the storage. You can use
3860 column names, or in the case of RDBMS back ends, function or stored procedure
3863 $rs = $schema->resultset('Employee')->search(undef, {
3866 { count => 'employeeid' },
3867 { max => { length => 'name' }, -as => 'longest_name' }
3872 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
3874 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
3875 use L</select>, to instruct DBIx::Class how to store the result of the column.
3876 Also note that the L</as> attribute has nothing to do with the SQL-side 'AS'
3877 identifier aliasing. You can however alias a function, so you can use it in
3878 e.g. an C<ORDER BY> clause. This is done via the C<-as> B<select function
3879 attribute> supplied as shown in the example above.
3881 B<NOTE:> You need to explicitly quote '+select'/'+as' when defining the attributes.
3882 Not doing so causes Perl to incorrectly interpret them as a bareword with a
3883 unary plus operator before it.
3889 Indicates additional columns to be selected from storage. Works the same as
3890 L</select> but adds columns to the default selection, instead of specifying
3899 Indicates additional column names for those added via L</+select>. See L</as>.
3907 =item Value: \@inflation_names
3911 Indicates column names for object inflation. That is L</as> indicates the
3912 slot name in which the column value will be stored within the
3913 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
3914 identifier by the C<get_column> method (or via the object accessor B<if one
3915 with the same name already exists>) as shown below. The L</as> attribute has
3916 B<nothing to do> with the SQL-side C<AS>. See L</select> for details.
3918 $rs = $schema->resultset('Employee')->search(undef, {
3921 { count => 'employeeid' },
3922 { max => { length => 'name' }, -as => 'longest_name' }
3931 If the object against which the search is performed already has an accessor
3932 matching a column name specified in C<as>, the value can be retrieved using
3933 the accessor as normal:
3935 my $name = $employee->name();
3937 If on the other hand an accessor does not exist in the object, you need to
3938 use C<get_column> instead:
3940 my $employee_count = $employee->get_column('employee_count');
3942 You can create your own accessors if required - see
3943 L<DBIx::Class::Manual::Cookbook> for details.
3949 =item Value: ($rel_name | \@rel_names | \%rel_names)
3953 Contains a list of relationships that should be joined for this query. For
3956 # Get CDs by Nine Inch Nails
3957 my $rs = $schema->resultset('CD')->search(
3958 { 'artist.name' => 'Nine Inch Nails' },
3959 { join => 'artist' }
3962 Can also contain a hash reference to refer to the other relation's relations.
3965 package MyApp::Schema::Track;
3966 use base qw/DBIx::Class/;
3967 __PACKAGE__->table('track');
3968 __PACKAGE__->add_columns(qw/trackid cd position title/);
3969 __PACKAGE__->set_primary_key('trackid');
3970 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
3973 # In your application
3974 my $rs = $schema->resultset('Artist')->search(
3975 { 'track.title' => 'Teardrop' },
3977 join => { cd => 'track' },
3978 order_by => 'artist.name',
3982 You need to use the relationship (not the table) name in conditions,
3983 because they are aliased as such. The current table is aliased as "me", so
3984 you need to use me.column_name in order to avoid ambiguity. For example:
3986 # Get CDs from 1984 with a 'Foo' track
3987 my $rs = $schema->resultset('CD')->search(
3990 'tracks.name' => 'Foo'
3992 { join => 'tracks' }
3995 If the same join is supplied twice, it will be aliased to <rel>_2 (and
3996 similarly for a third time). For e.g.
3998 my $rs = $schema->resultset('Artist')->search({
3999 'cds.title' => 'Down to Earth',
4000 'cds_2.title' => 'Popular',
4002 join => [ qw/cds cds/ ],
4005 will return a set of all artists that have both a cd with title 'Down
4006 to Earth' and a cd with title 'Popular'.
4008 If you want to fetch related objects from other tables as well, see C<prefetch>
4011 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
4017 =item Value: ($rel_name | \@rel_names | \%rel_names)
4021 Contains one or more relationships that should be fetched along with
4022 the main query (when they are accessed afterwards the data will
4023 already be available, without extra queries to the database). This is
4024 useful for when you know you will need the related objects, because it
4025 saves at least one query:
4027 my $rs = $schema->resultset('Tag')->search(
4036 The initial search results in SQL like the following:
4038 SELECT tag.*, cd.*, artist.* FROM tag
4039 JOIN cd ON tag.cd = cd.cdid
4040 JOIN artist ON cd.artist = artist.artistid
4042 L<DBIx::Class> has no need to go back to the database when we access the
4043 C<cd> or C<artist> relationships, which saves us two SQL statements in this
4046 Simple prefetches will be joined automatically, so there is no need
4047 for a C<join> attribute in the above search.
4049 L</prefetch> can be used with the any of the relationship types and
4050 multiple prefetches can be specified together. Below is a more complex
4051 example that prefetches a CD's artist, its liner notes (if present),
4052 the cover image, the tracks on that cd, and the guests on those
4056 My::Schema::CD->belongs_to( artist => 'My::Schema::Artist' );
4057 My::Schema::CD->might_have( liner_note => 'My::Schema::LinerNotes' );
4058 My::Schema::CD->has_one( cover_image => 'My::Schema::Artwork' );
4059 My::Schema::CD->has_many( tracks => 'My::Schema::Track' );
4061 My::Schema::Artist->belongs_to( record_label => 'My::Schema::RecordLabel' );
4063 My::Schema::Track->has_many( guests => 'My::Schema::Guest' );
4066 my $rs = $schema->resultset('CD')->search(
4070 { artist => 'record_label'}, # belongs_to => belongs_to
4071 'liner_note', # might_have
4072 'cover_image', # has_one
4073 { tracks => 'guests' }, # has_many => has_many
4078 This will produce SQL like the following:
4080 SELECT cd.*, artist.*, record_label.*, liner_note.*, cover_image.*,
4084 ON artist.artistid = me.artistid
4085 JOIN record_label record_label
4086 ON record_label.labelid = artist.labelid
4087 LEFT JOIN track tracks
4088 ON tracks.cdid = me.cdid
4089 LEFT JOIN guest guests
4090 ON guests.trackid = track.trackid
4091 LEFT JOIN liner_notes liner_note
4092 ON liner_note.cdid = me.cdid
4093 JOIN cd_artwork cover_image
4094 ON cover_image.cdid = me.cdid
4097 Now the C<artist>, C<record_label>, C<liner_note>, C<cover_image>,
4098 C<tracks>, and C<guests> of the CD will all be available through the
4099 relationship accessors without the need for additional queries to the
4102 However, there is one caveat to be observed: it can be dangerous to
4103 prefetch more than one L<has_many|DBIx::Class::Relationship/has_many>
4104 relationship on a given level. e.g.:
4106 my $rs = $schema->resultset('CD')->search(
4110 'tracks', # has_many
4111 { cd_to_producer => 'producer' }, # has_many => belongs_to (i.e. m2m)
4116 In fact, C<DBIx::Class> will emit the following warning:
4118 Prefetching multiple has_many rels tracks and cd_to_producer at top
4119 level will explode the number of row objects retrievable via ->next
4120 or ->all. Use at your own risk.
4122 The collapser currently can't identify duplicate tuples for multiple
4123 L<has_many|DBIx::Class::Relationship/has_many> relationships and as a
4124 result the second L<has_many|DBIx::Class::Relationship/has_many>
4125 relation could contain redundant objects.
4127 =head3 Using L</prefetch> with L</join>
4129 L</prefetch> implies a L</join> with the equivalent argument, and is
4130 properly merged with any existing L</join> specification. So the
4133 my $rs = $schema->resultset('CD')->search(
4134 {'record_label.name' => 'Music Product Ltd.'},
4136 join => {artist => 'record_label'},
4137 prefetch => 'artist',
4141 ... will work, searching on the record label's name, but only
4142 prefetching the C<artist>.
4144 =head3 Using L</prefetch> with L</select> / L</+select> / L</as> / L</+as>
4146 L</prefetch> implies a L</+select>/L</+as> with the fields of the
4147 prefetched relations. So given:
4149 my $rs = $schema->resultset('CD')->search(
4152 select => ['cd.title'],
4154 prefetch => 'artist',
4158 The L</select> becomes: C<'cd.title', 'artist.*'> and the L</as>
4159 becomes: C<'cd_title', 'artist.*'>.
4163 Prefetch does a lot of deep magic. As such, it may not behave exactly
4164 as you might expect.
4170 Prefetch uses the L</cache> to populate the prefetched relationships. This
4171 may or may not be what you want.
4175 If you specify a condition on a prefetched relationship, ONLY those
4176 rows that match the prefetched condition will be fetched into that relationship.
4177 This means that adding prefetch to a search() B<may alter> what is returned by
4178 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
4180 my $artist_rs = $schema->resultset('Artist')->search({
4186 my $count = $artist_rs->first->cds->count;
4188 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
4190 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
4192 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
4194 that cmp_ok() may or may not pass depending on the datasets involved. This
4195 behavior may or may not survive the 0.09 transition.
4207 Makes the resultset paged and specifies the page to retrieve. Effectively
4208 identical to creating a non-pages resultset and then calling ->page($page)
4211 If L</rows> attribute is not specified it defaults to 10 rows per page.
4213 When you have a paged resultset, L</count> will only return the number
4214 of rows in the page. To get the total, use the L</pager> and call
4215 C<total_entries> on it.
4225 Specifies the maximum number of rows for direct retrieval or the number of
4226 rows per page if the page attribute or method is used.
4232 =item Value: $offset
4236 Specifies the (zero-based) row number for the first row to be returned, or the
4237 of the first row of the first page if paging is used.
4239 =head2 software_limit
4243 =item Value: (0 | 1)
4247 When combined with L</rows> and/or L</offset> the generated SQL will not
4248 include any limit dialect stanzas. Instead the entire result will be selected
4249 as if no limits were specified, and DBIC will perform the limit locally, by
4250 artificially advancing and finishing the resulting L</cursor>.
4252 This is the recommended way of performing resultset limiting when no sane RDBMS
4253 implementation is available (e.g.
4254 L<Sybase ASE|DBIx::Class::Storage::DBI::Sybase::ASE> using the
4255 L<Generic Sub Query|DBIx::Class::SQLMaker::LimitDialects/GenericSubQ> hack)
4261 =item Value: \@columns
4265 A arrayref of columns to group by. Can include columns of joined tables.
4267 group_by => [qw/ column1 column2 ... /]
4273 =item Value: $condition
4277 HAVING is a select statement attribute that is applied between GROUP BY and
4278 ORDER BY. It is applied to the after the grouping calculations have been
4281 having => { 'count_employee' => { '>=', 100 } }
4283 or with an in-place function in which case literal SQL is required:
4285 having => \[ 'count(employee) >= ?', [ count => 100 ] ]
4291 =item Value: (0 | 1)
4295 Set to 1 to group by all columns. If the resultset already has a group_by
4296 attribute, this setting is ignored and an appropriate warning is issued.
4302 Adds to the WHERE clause.
4304 # only return rows WHERE deleted IS NULL for all searches
4305 __PACKAGE__->resultset_attributes({ where => { deleted => undef } }); )
4307 Can be overridden by passing C<< { where => undef } >> as an attribute
4310 For more complicated where clauses see L<SQL::Abstract/WHERE CLAUSES>.
4316 Set to 1 to cache search results. This prevents extra SQL queries if you
4317 revisit rows in your ResultSet:
4319 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
4321 while( my $artist = $resultset->next ) {
4325 $rs->first; # without cache, this would issue a query
4327 By default, searches are not cached.
4329 For more examples of using these attributes, see
4330 L<DBIx::Class::Manual::Cookbook>.
4336 =item Value: ( 'update' | 'shared' )
4340 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT