1 package DBIx::Class::ResultSet;
5 use base qw/DBIx::Class/;
6 use Carp::Clan qw/^DBIx::Class/;
7 use DBIx::Class::Exception;
10 use DBIx::Class::ResultSetColumn;
11 use DBIx::Class::ResultSourceHandle;
13 use Scalar::Util qw/blessed weaken/;
22 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class _source_handle/);
26 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
30 my $users_rs = $schema->resultset('User');
31 while( $user = $users_rs->next) {
32 print $user->username;
35 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
36 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
40 A ResultSet is an object which stores a set of conditions representing
41 a query. It is the backbone of DBIx::Class (i.e. the really
42 important/useful bit).
44 No SQL is executed on the database when a ResultSet is created, it
45 just stores all the conditions needed to create the query.
47 A basic ResultSet representing the data of an entire table is returned
48 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
49 L<Source|DBIx::Class::Manual::Glossary/Source> name.
51 my $users_rs = $schema->resultset('User');
53 A new ResultSet is returned from calling L</search> on an existing
54 ResultSet. The new one will contain all the conditions of the
55 original, plus any new conditions added in the C<search> call.
57 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
58 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
61 The query that the ResultSet represents is B<only> executed against
62 the database when these methods are called:
63 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
65 If a resultset is used in a numeric context it returns the L</count>.
66 However, if it is used in a boolean context it is B<always> true. So if
67 you want to check if a resultset has any results, you must use C<if $rs
72 =head2 Chaining resultsets
74 Let's say you've got a query that needs to be run to return some data
75 to the user. But, you have an authorization system in place that
76 prevents certain users from seeing certain information. So, you want
77 to construct the basic query in one method, but add constraints to it in
82 my $request = $self->get_request; # Get a request object somehow.
83 my $schema = $self->get_schema; # Get the DBIC schema object somehow.
85 my $cd_rs = $schema->resultset('CD')->search({
86 title => $request->param('title'),
87 year => $request->param('year'),
90 $self->apply_security_policy( $cd_rs );
95 sub apply_security_policy {
104 =head3 Resolving conditions and attributes
106 When a resultset is chained from another resultset, conditions and
107 attributes with the same keys need resolving.
109 L</join>, L</prefetch>, L</+select>, L</+as> attributes are merged
110 into the existing ones from the original resultset.
112 The L</where> and L</having> attributes, and any search conditions, are
113 merged with an SQL C<AND> to the existing condition from the original
116 All other attributes are overridden by any new ones supplied in the
119 =head2 Multiple queries
121 Since a resultset just defines a query, you can do all sorts of
122 things with it with the same object.
124 # Don't hit the DB yet.
125 my $cd_rs = $schema->resultset('CD')->search({
126 title => 'something',
130 # Each of these hits the DB individually.
131 my $count = $cd_rs->count;
132 my $most_recent = $cd_rs->get_column('date_released')->max();
133 my @records = $cd_rs->all;
135 And it's not just limited to SELECT statements.
141 $cd_rs->create({ artist => 'Fred' });
143 Which is the same as:
145 $schema->resultset('CD')->create({
146 title => 'something',
151 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
159 =item Arguments: $source, \%$attrs
161 =item Return Value: $rs
165 The resultset constructor. Takes a source object (usually a
166 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
167 L</ATTRIBUTES> below). Does not perform any queries -- these are
168 executed as needed by the other methods.
170 Generally you won't need to construct a resultset manually. You'll
171 automatically get one from e.g. a L</search> called in scalar context:
173 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
175 IMPORTANT: If called on an object, proxies to new_result instead so
177 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
179 will return a CD object, not a ResultSet.
185 return $class->new_result(@_) if ref $class;
187 my ($source, $attrs) = @_;
188 $source = $source->handle
189 unless $source->isa('DBIx::Class::ResultSourceHandle');
190 $attrs = { %{$attrs||{}} };
192 if ($attrs->{page}) {
193 $attrs->{rows} ||= 10;
196 $attrs->{alias} ||= 'me';
198 # Creation of {} and bless separated to mitigate RH perl bug
199 # see https://bugzilla.redhat.com/show_bug.cgi?id=196836
201 _source_handle => $source,
202 cond => $attrs->{where},
210 $attrs->{result_class} || $source->resolve->result_class
220 =item Arguments: $cond, \%attrs?
222 =item Return Value: $resultset (scalar context), @row_objs (list context)
226 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
227 my $new_rs = $cd_rs->search({ year => 2005 });
229 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
230 # year = 2005 OR year = 2004
232 If you need to pass in additional attributes but no additional condition,
233 call it as C<search(undef, \%attrs)>.
235 # "SELECT name, artistid FROM $artist_table"
236 my @all_artists = $schema->resultset('Artist')->search(undef, {
237 columns => [qw/name artistid/],
240 For a list of attributes that can be passed to C<search>, see
241 L</ATTRIBUTES>. For more examples of using this function, see
242 L<Searching|DBIx::Class::Manual::Cookbook/Searching>. For a complete
243 documentation for the first argument, see L<SQL::Abstract>.
245 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
251 my $rs = $self->search_rs( @_ );
253 my $want = wantarray;
257 elsif (defined $want) {
261 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense');
269 =item Arguments: $cond, \%attrs?
271 =item Return Value: $resultset
275 This method does the same exact thing as search() except it will
276 always return a resultset, even in list context.
283 # Special-case handling for (undef, undef).
284 if ( @_ == 2 && !defined $_[1] && !defined $_[0] ) {
289 $call_attrs = pop(@_) if (
290 @_ > 1 and ( ! defined $_[-1] or ref $_[-1] eq 'HASH' )
293 # see if we can keep the cache (no $rs changes)
295 my %safe = (alias => 1, cache => 1);
296 if ( ! List::Util::first { !$safe{$_} } keys %$call_attrs and (
299 ref $_[0] eq 'HASH' && ! keys %{$_[0]}
301 ref $_[0] eq 'ARRAY' && ! @{$_[0]}
303 $cache = $self->get_cache;
306 my $old_attrs = { %{$self->{attrs}} };
307 my $old_having = delete $old_attrs->{having};
308 my $old_where = delete $old_attrs->{where};
310 # reset the selector list
311 if (List::Util::first { exists $call_attrs->{$_} } qw{columns select as}) {
312 delete @{$old_attrs}{qw{select as columns +select +as +columns include_columns}};
315 my $new_attrs = { %{$old_attrs}, %{$call_attrs} };
317 # merge new attrs into inherited
318 foreach my $key (qw/join prefetch +select +as +columns include_columns bind/) {
319 next unless exists $call_attrs->{$key};
320 $new_attrs->{$key} = $self->_merge_attr($old_attrs->{$key}, $call_attrs->{$key});
323 # rip apart the rest of @_, parse a condition
326 if (ref $_[0] eq 'HASH') {
327 (keys %{$_[0]}) ? $_[0] : undef
333 $self->throw_exception('Odd number of arguments to search')
341 carp 'search( %condition ) is deprecated, use search( \%condition ) instead'
342 if (@_ > 1 and ! $self->result_source->result_class->isa('DBIx::Class::CDBICompat') );
344 for ($old_where, $call_cond) {
346 $new_attrs->{where} = $self->_stack_cond (
347 $_, $new_attrs->{where}
352 if (defined $old_having) {
353 $new_attrs->{having} = $self->_stack_cond (
354 $old_having, $new_attrs->{having}
358 my $rs = (ref $self)->new($self->result_source, $new_attrs);
360 $rs->set_cache($cache) if ($cache);
366 my ($self, $left, $right) = @_;
367 if (defined $left xor defined $right) {
368 return defined $left ? $left : $right;
370 elsif (defined $left) {
371 return { -and => [ map
372 { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
380 =head2 search_literal
384 =item Arguments: $sql_fragment, @bind_values
386 =item Return Value: $resultset (scalar context), @row_objs (list context)
390 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
391 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
393 Pass a literal chunk of SQL to be added to the conditional part of the
396 CAVEAT: C<search_literal> is provided for Class::DBI compatibility and should
397 only be used in that context. C<search_literal> is a convenience method.
398 It is equivalent to calling $schema->search(\[]), but if you want to ensure
399 columns are bound correctly, use C<search>.
401 Example of how to use C<search> instead of C<search_literal>
403 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
404 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
407 See L<DBIx::Class::Manual::Cookbook/Searching> and
408 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
409 require C<search_literal>.
414 my ($self, $sql, @bind) = @_;
416 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
419 return $self->search(\[ $sql, map [ __DUMMY__ => $_ ], @bind ], ($attr || () ));
426 =item Arguments: \%columns_values | @pk_values, \%attrs?
428 =item Return Value: $row_object | undef
432 Finds and returns a single row based on supplied criteria. Takes either a
433 hashref with the same format as L</create> (including inference of foreign
434 keys from related objects), or a list of primary key values in the same
435 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
436 declaration on the L</result_source>.
438 In either case an attempt is made to combine conditions already existing on
439 the resultset with the condition passed to this method.
441 To aid with preparing the correct query for the storage you may supply the
442 C<key> attribute, which is the name of a
443 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
444 unique constraint corresponding to the
445 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
446 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
447 to construct a query that satisfies the named unique constraint fully (
448 non-NULL values for each column member of the constraint) an exception is
451 If no C<key> is specified, the search is carried over all unique constraints
452 which are fully defined by the available condition.
454 If no such constraint is found, C<find> currently defaults to a simple
455 C<< search->(\%column_values) >> which may or may not do what you expect.
456 Note that this fallback behavior may be deprecated in further versions. If
457 you need to search with arbitrary conditions - use L</search>. If the query
458 resulting from this fallback produces more than one row, a warning to the
459 effect is issued, though only the first row is constructed and returned as
462 In addition to C<key>, L</find> recognizes and applies standard
463 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
465 Note that if you have extra concerns about the correctness of the resulting
466 query you need to specify the C<key> attribute and supply the entire condition
467 as an argument to find (since it is not always possible to perform the
468 combination of the resultset condition with the supplied one, especially if
469 the resultset condition contains literal sql).
471 For example, to find a row by its primary key:
473 my $cd = $schema->resultset('CD')->find(5);
475 You can also find a row by a specific unique constraint:
477 my $cd = $schema->resultset('CD')->find(
479 artist => 'Massive Attack',
480 title => 'Mezzanine',
482 { key => 'cd_artist_title' }
485 See also L</find_or_create> and L</update_or_create>.
491 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
493 my $rsrc = $self->result_source;
495 # Parse out the condition from input
497 if (ref $_[0] eq 'HASH') {
498 $call_cond = { %{$_[0]} };
501 my $constraint = exists $attrs->{key} ? $attrs->{key} : 'primary';
502 my @c_cols = $rsrc->unique_constraint_columns($constraint);
504 $self->throw_exception(
505 "No constraint columns, maybe a malformed '$constraint' constraint?"
508 $self->throw_exception (
509 'find() expects either a column/value hashref, or a list of values '
510 . "corresponding to the columns of the specified unique constraint '$constraint'"
511 ) unless @c_cols == @_;
514 @{$call_cond}{@c_cols} = @_;
518 for my $key (keys %$call_cond) {
520 my $keyref = ref($call_cond->{$key})
522 my $relinfo = $rsrc->relationship_info($key)
524 my $val = delete $call_cond->{$key};
526 next if $keyref eq 'ARRAY'; # has_many for multi_create
528 my $rel_q = $rsrc->_resolve_condition(
529 $relinfo->{cond}, $val, $key
531 die "Can't handle complex relationship conditions in find" if ref($rel_q) ne 'HASH';
532 @related{keys %$rel_q} = values %$rel_q;
536 # relationship conditions take precedence (?)
537 @{$call_cond}{keys %related} = values %related;
539 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
541 if (exists $attrs->{key}) {
542 $final_cond = $self->_qualify_cond_columns (
544 $self->_build_unique_cond (
552 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
553 # This means that we got here after a merger of relationship conditions
554 # in ::Relationship::Base::search_related (the row method), and furthermore
555 # the relationship is of the 'single' type. This means that the condition
556 # provided by the relationship (already attached to $self) is sufficient,
557 # as there can be only one row in the database that would satisfy the
561 # no key was specified - fall down to heuristics mode:
562 # run through all unique queries registered on the resultset, and
563 # 'OR' all qualifying queries together
564 my (@unique_queries, %seen_column_combinations);
565 for my $c_name ($rsrc->unique_constraint_names) {
566 next if $seen_column_combinations{
567 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
570 push @unique_queries, try {
571 $self->_build_unique_cond ($c_name, $call_cond)
575 $final_cond = @unique_queries
576 ? [ map { $self->_qualify_cond_columns($_, $alias) } @unique_queries ]
577 : $self->_non_unique_find_fallback ($call_cond, $attrs)
581 # Run the query, passing the result_class since it should propagate for find
582 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
583 if (keys %{$rs->_resolved_attrs->{collapse}}) {
585 carp "Query returned more than one row" if $rs->next;
593 # This is a stop-gap method as agreed during the discussion on find() cleanup:
594 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
596 # It is invoked when find() is called in legacy-mode with insufficiently-unique
597 # condition. It is provided for overrides until a saner way forward is devised
599 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
600 # the road. Please adjust your tests accordingly to catch this situation early
601 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
603 # The method will not be removed without an adequately complete replacement
604 # for strict-mode enforcement
605 sub _non_unique_find_fallback {
606 my ($self, $cond, $attrs) = @_;
608 return $self->_qualify_cond_columns(
610 exists $attrs->{alias}
612 : $self->{attrs}{alias}
617 sub _qualify_cond_columns {
618 my ($self, $cond, $alias) = @_;
620 my %aliased = %$cond;
621 for (keys %aliased) {
622 $aliased{"$alias.$_"} = delete $aliased{$_}
629 sub _build_unique_cond {
630 my ($self, $constraint_name, $extra_cond) = @_;
632 my @c_cols = $self->result_source->unique_constraint_columns($constraint_name);
634 # combination may fail if $self->{cond} is non-trivial
635 my ($final_cond) = try {
636 $self->_merge_with_rscond ($extra_cond)
641 # trim out everything not in $columns
642 $final_cond = { map { $_ => $final_cond->{$_} } @c_cols };
644 if (my @missing = grep { ! defined $final_cond->{$_} } (@c_cols) ) {
645 $self->throw_exception( sprintf ( "Unable to satisfy requested constraint '%s', no values for column(s): %s",
647 join (', ', map { "'$_'" } @missing),
654 =head2 search_related
658 =item Arguments: $rel, $cond, \%attrs?
660 =item Return Value: $new_resultset
664 $new_rs = $cd_rs->search_related('artist', {
668 Searches the specified relationship, optionally specifying a condition and
669 attributes for matching records. See L</ATTRIBUTES> for more information.
674 return shift->related_resultset(shift)->search(@_);
677 =head2 search_related_rs
679 This method works exactly the same as search_related, except that
680 it guarantees a resultset, even in list context.
684 sub search_related_rs {
685 return shift->related_resultset(shift)->search_rs(@_);
692 =item Arguments: none
694 =item Return Value: $cursor
698 Returns a storage-driven cursor to the given resultset. See
699 L<DBIx::Class::Cursor> for more information.
706 my $attrs = $self->_resolved_attrs_copy;
708 return $self->{cursor}
709 ||= $self->result_source->storage->select($attrs->{from}, $attrs->{select},
710 $attrs->{where},$attrs);
717 =item Arguments: $cond?
719 =item Return Value: $row_object | undef
723 my $cd = $schema->resultset('CD')->single({ year => 2001 });
725 Inflates the first result without creating a cursor if the resultset has
726 any records in it; if not returns C<undef>. Used by L</find> as a lean version
729 While this method can take an optional search condition (just like L</search>)
730 being a fast-code-path it does not recognize search attributes. If you need to
731 add extra joins or similar, call L</search> and then chain-call L</single> on the
732 L<DBIx::Class::ResultSet> returned.
738 As of 0.08100, this method enforces the assumption that the preceding
739 query returns only one row. If more than one row is returned, you will receive
742 Query returned more than one row
744 In this case, you should be using L</next> or L</find> instead, or if you really
745 know what you are doing, use the L</rows> attribute to explicitly limit the size
748 This method will also throw an exception if it is called on a resultset prefetching
749 has_many, as such a prefetch implies fetching multiple rows from the database in
750 order to assemble the resulting object.
757 my ($self, $where) = @_;
759 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
762 my $attrs = $self->_resolved_attrs_copy;
764 if (keys %{$attrs->{collapse}}) {
765 $self->throw_exception(
766 'single() can not be used on resultsets prefetching has_many. Use find( \%cond ) or next() instead'
771 if (defined $attrs->{where}) {
774 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
775 $where, delete $attrs->{where} ]
778 $attrs->{where} = $where;
782 my @data = $self->result_source->storage->select_single(
783 $attrs->{from}, $attrs->{select},
784 $attrs->{where}, $attrs
787 return (@data ? ($self->_construct_object(@data))[0] : undef);
793 # Recursively collapse the query, accumulating values for each column.
795 sub _collapse_query {
796 my ($self, $query, $collapsed) = @_;
800 if (ref $query eq 'ARRAY') {
801 foreach my $subquery (@$query) {
802 next unless ref $subquery; # -or
803 $collapsed = $self->_collapse_query($subquery, $collapsed);
806 elsif (ref $query eq 'HASH') {
807 if (keys %$query and (keys %$query)[0] eq '-and') {
808 foreach my $subquery (@{$query->{-and}}) {
809 $collapsed = $self->_collapse_query($subquery, $collapsed);
813 foreach my $col (keys %$query) {
814 my $value = $query->{$col};
815 $collapsed->{$col}{$value}++;
827 =item Arguments: $cond?
829 =item Return Value: $resultsetcolumn
833 my $max_length = $rs->get_column('length')->max;
835 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
840 my ($self, $column) = @_;
841 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
849 =item Arguments: $cond, \%attrs?
851 =item Return Value: $resultset (scalar context), @row_objs (list context)
855 # WHERE title LIKE '%blue%'
856 $cd_rs = $rs->search_like({ title => '%blue%'});
858 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
859 that this is simply a convenience method retained for ex Class::DBI users.
860 You most likely want to use L</search> with specific operators.
862 For more information, see L<DBIx::Class::Manual::Cookbook>.
864 This method is deprecated and will be removed in 0.09. Use L</search()>
865 instead. An example conversion is:
867 ->search_like({ foo => 'bar' });
871 ->search({ foo => { like => 'bar' } });
878 'search_like() is deprecated and will be removed in DBIC version 0.09.'
879 .' Instead use ->search({ x => { -like => "y%" } })'
880 .' (note the outer pair of {}s - they are important!)'
882 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
883 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
884 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
885 return $class->search($query, { %$attrs });
892 =item Arguments: $first, $last
894 =item Return Value: $resultset (scalar context), @row_objs (list context)
898 Returns a resultset or object list representing a subset of elements from the
899 resultset slice is called on. Indexes are from 0, i.e., to get the first
902 my ($one, $two, $three) = $rs->slice(0, 2);
907 my ($self, $min, $max) = @_;
908 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
909 $attrs->{offset} = $self->{attrs}{offset} || 0;
910 $attrs->{offset} += $min;
911 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
912 return $self->search(undef, $attrs);
913 #my $slice = (ref $self)->new($self->result_source, $attrs);
914 #return (wantarray ? $slice->all : $slice);
921 =item Arguments: none
923 =item Return Value: $result | undef
927 Returns the next element in the resultset (C<undef> is there is none).
929 Can be used to efficiently iterate over records in the resultset:
931 my $rs = $schema->resultset('CD')->search;
932 while (my $cd = $rs->next) {
936 Note that you need to store the resultset object, and call C<next> on it.
937 Calling C<< resultset('Table')->next >> repeatedly will always return the
938 first record from the resultset.
944 if (my $cache = $self->get_cache) {
945 $self->{all_cache_position} ||= 0;
946 return $cache->[$self->{all_cache_position}++];
948 if ($self->{attrs}{cache}) {
949 delete $self->{pager};
950 $self->{all_cache_position} = 1;
951 return ($self->all)[0];
953 if ($self->{stashed_objects}) {
954 my $obj = shift(@{$self->{stashed_objects}});
955 delete $self->{stashed_objects} unless @{$self->{stashed_objects}};
959 exists $self->{stashed_row}
960 ? @{delete $self->{stashed_row}}
961 : $self->cursor->next
963 return undef unless (@row);
964 my ($row, @more) = $self->_construct_object(@row);
965 $self->{stashed_objects} = \@more if @more;
969 sub _construct_object {
970 my ($self, @row) = @_;
972 my $info = $self->_collapse_result($self->{_attrs}{as}, \@row)
974 my @new = $self->result_class->inflate_result($self->result_source, @$info);
975 @new = $self->{_attrs}{record_filter}->(@new)
976 if exists $self->{_attrs}{record_filter};
980 sub _collapse_result {
981 my ($self, $as_proto, $row) = @_;
985 # 'foo' => [ undef, 'foo' ]
986 # 'foo.bar' => [ 'foo', 'bar' ]
987 # 'foo.bar.baz' => [ 'foo.bar', 'baz' ]
989 my @construct_as = map { [ (/^(?:(.*)\.)?([^.]+)$/) ] } @$as_proto;
991 my %collapse = %{$self->{_attrs}{collapse}||{}};
995 # if we're doing collapsing (has_many prefetch) we need to grab records
996 # until the PK changes, so fill @pri_index. if not, we leave it empty so
997 # we know we don't have to bother.
999 # the reason for not using the collapse stuff directly is because if you
1000 # had for e.g. two artists in a row with no cds, the collapse info for
1001 # both would be NULL (undef) so you'd lose the second artist
1003 # store just the index so we can check the array positions from the row
1004 # without having to contruct the full hash
1006 if (keys %collapse) {
1007 my %pri = map { ($_ => 1) } $self->result_source->_pri_cols;
1008 foreach my $i (0 .. $#construct_as) {
1009 next if defined($construct_as[$i][0]); # only self table
1010 if (delete $pri{$construct_as[$i][1]}) {
1011 push(@pri_index, $i);
1013 last unless keys %pri; # short circuit (Johnny Five Is Alive!)
1017 # no need to do an if, it'll be empty if @pri_index is empty anyway
1019 my %pri_vals = map { ($_ => $copy[$_]) } @pri_index;
1023 do { # no need to check anything at the front, we always want the first row
1027 foreach my $this_as (@construct_as) {
1028 $const{$this_as->[0]||''}{$this_as->[1]} = shift(@copy);
1031 push(@const_rows, \%const);
1033 } until ( # no pri_index => no collapse => drop straight out
1036 do { # get another row, stash it, drop out if different PK
1038 @copy = $self->cursor->next;
1039 $self->{stashed_row} = \@copy;
1041 # last thing in do block, counts as true if anything doesn't match
1043 # check xor defined first for NULL vs. NOT NULL then if one is
1044 # defined the other must be so check string equality
1047 (defined $pri_vals{$_} ^ defined $copy[$_])
1048 || (defined $pri_vals{$_} && ($pri_vals{$_} ne $copy[$_]))
1053 my $alias = $self->{attrs}{alias};
1060 foreach my $const (@const_rows) {
1061 scalar @const_keys or do {
1062 @const_keys = sort { length($a) <=> length($b) } keys %$const;
1064 foreach my $key (@const_keys) {
1067 my @parts = split(/\./, $key);
1069 my $data = $const->{$key};
1070 foreach my $p (@parts) {
1071 $target = $target->[1]->{$p} ||= [];
1073 if ($cur eq ".${key}" && (my @ckey = @{$collapse{$cur}||[]})) {
1074 # collapsing at this point and on final part
1075 my $pos = $collapse_pos{$cur};
1076 CK: foreach my $ck (@ckey) {
1077 if (!defined $pos->{$ck} || $pos->{$ck} ne $data->{$ck}) {
1078 $collapse_pos{$cur} = $data;
1079 delete @collapse_pos{ # clear all positioning for sub-entries
1080 grep { m/^\Q${cur}.\E/ } keys %collapse_pos
1087 if (exists $collapse{$cur}) {
1088 $target = $target->[-1];
1091 $target->[0] = $data;
1093 $info->[0] = $const->{$key};
1101 =head2 result_source
1105 =item Arguments: $result_source?
1107 =item Return Value: $result_source
1111 An accessor for the primary ResultSource object from which this ResultSet
1118 =item Arguments: $result_class?
1120 =item Return Value: $result_class
1124 An accessor for the class to use when creating row objects. Defaults to
1125 C<< result_source->result_class >> - which in most cases is the name of the
1126 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1128 Note that changing the result_class will also remove any components
1129 that were originally loaded in the source class via
1130 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1131 in the original source class will not run.
1136 my ($self, $result_class) = @_;
1137 if ($result_class) {
1138 unless (ref $result_class) { # don't fire this for an object
1139 $self->ensure_class_loaded($result_class);
1141 $self->_result_class($result_class);
1142 # THIS LINE WOULD BE A BUG - this accessor specifically exists to
1143 # permit the user to set result class on one result set only; it only
1144 # chains if provided to search()
1145 #$self->{attrs}{result_class} = $result_class if ref $self;
1147 $self->_result_class;
1154 =item Arguments: $cond, \%attrs??
1156 =item Return Value: $count
1160 Performs an SQL C<COUNT> with the same query as the resultset was built
1161 with to find the number of elements. Passing arguments is equivalent to
1162 C<< $rs->search ($cond, \%attrs)->count >>
1168 return $self->search(@_)->count if @_ and defined $_[0];
1169 return scalar @{ $self->get_cache } if $self->get_cache;
1171 my $attrs = $self->_resolved_attrs_copy;
1173 # this is a little optimization - it is faster to do the limit
1174 # adjustments in software, instead of a subquery
1175 my $rows = delete $attrs->{rows};
1176 my $offset = delete $attrs->{offset};
1179 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1180 $crs = $self->_count_subq_rs ($attrs);
1183 $crs = $self->_count_rs ($attrs);
1185 my $count = $crs->next;
1187 $count -= $offset if $offset;
1188 $count = $rows if $rows and $rows < $count;
1189 $count = 0 if ($count < 0);
1198 =item Arguments: $cond, \%attrs??
1200 =item Return Value: $count_rs
1204 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1205 This can be very handy for subqueries:
1207 ->search( { amount => $some_rs->count_rs->as_query } )
1209 As with regular resultsets the SQL query will be executed only after
1210 the resultset is accessed via L</next> or L</all>. That would return
1211 the same single value obtainable via L</count>.
1217 return $self->search(@_)->count_rs if @_;
1219 # this may look like a lack of abstraction (count() does about the same)
1220 # but in fact an _rs *must* use a subquery for the limits, as the
1221 # software based limiting can not be ported if this $rs is to be used
1222 # in a subquery itself (i.e. ->as_query)
1223 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1224 return $self->_count_subq_rs;
1227 return $self->_count_rs;
1232 # returns a ResultSetColumn object tied to the count query
1235 my ($self, $attrs) = @_;
1237 my $rsrc = $self->result_source;
1238 $attrs ||= $self->_resolved_attrs;
1240 my $tmp_attrs = { %$attrs };
1241 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1242 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1244 # overwrite the selector (supplied by the storage)
1245 $tmp_attrs->{select} = $rsrc->storage->_count_select ($rsrc, $attrs);
1246 $tmp_attrs->{as} = 'count';
1248 my $tmp_rs = $rsrc->resultset_class->new($rsrc, $tmp_attrs)->get_column ('count');
1254 # same as above but uses a subquery
1256 sub _count_subq_rs {
1257 my ($self, $attrs) = @_;
1259 my $rsrc = $self->result_source;
1260 $attrs ||= $self->_resolved_attrs;
1262 my $sub_attrs = { %$attrs };
1263 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1264 delete @{$sub_attrs}{qw/collapse select _prefetch_select as order_by for/};
1266 # if we multi-prefetch we group_by primary keys only as this is what we would
1267 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1268 if ( keys %{$attrs->{collapse}} ) {
1269 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } ($rsrc->_pri_cols) ]
1272 # Calculate subquery selector
1273 if (my $g = $sub_attrs->{group_by}) {
1275 my $sql_maker = $rsrc->storage->sql_maker;
1277 # necessary as the group_by may refer to aliased functions
1279 for my $sel (@{$attrs->{select}}) {
1280 $sel_index->{$sel->{-as}} = $sel
1281 if (ref $sel eq 'HASH' and $sel->{-as});
1284 for my $g_part (@$g) {
1285 my $colpiece = $sel_index->{$g_part} || $g_part;
1287 # disqualify join-based group_by's. Arcane but possible query
1288 # also horrible horrible hack to alias a column (not a func.)
1289 # (probably need to introduce SQLA syntax)
1290 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1293 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1295 push @{$sub_attrs->{select}}, $colpiece;
1299 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1300 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1303 return $rsrc->resultset_class
1304 ->new ($rsrc, $sub_attrs)
1306 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1307 ->get_column ('count');
1314 =head2 count_literal
1318 =item Arguments: $sql_fragment, @bind_values
1320 =item Return Value: $count
1324 Counts the results in a literal query. Equivalent to calling L</search_literal>
1325 with the passed arguments, then L</count>.
1329 sub count_literal { shift->search_literal(@_)->count; }
1335 =item Arguments: none
1337 =item Return Value: @objects
1341 Returns all elements in the resultset. Called implicitly if the resultset
1342 is returned in list context.
1349 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1352 return @{ $self->get_cache } if $self->get_cache;
1356 if (keys %{$self->_resolved_attrs->{collapse}}) {
1357 # Using $self->cursor->all is really just an optimisation.
1358 # If we're collapsing has_many prefetches it probably makes
1359 # very little difference, and this is cleaner than hacking
1360 # _construct_object to survive the approach
1361 $self->cursor->reset;
1362 my @row = $self->cursor->next;
1364 push(@obj, $self->_construct_object(@row));
1365 @row = (exists $self->{stashed_row}
1366 ? @{delete $self->{stashed_row}}
1367 : $self->cursor->next);
1370 @obj = map { $self->_construct_object(@$_) } $self->cursor->all;
1373 $self->set_cache(\@obj) if $self->{attrs}{cache};
1382 =item Arguments: none
1384 =item Return Value: $self
1388 Resets the resultset's cursor, so you can iterate through the elements again.
1389 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1396 delete $self->{_attrs} if exists $self->{_attrs};
1397 $self->{all_cache_position} = 0;
1398 $self->cursor->reset;
1406 =item Arguments: none
1408 =item Return Value: $object | undef
1412 Resets the resultset and returns an object for the first result (or C<undef>
1413 if the resultset is empty).
1418 return $_[0]->reset->next;
1424 # Determines whether and what type of subquery is required for the $rs operation.
1425 # If grouping is necessary either supplies its own, or verifies the current one
1426 # After all is done delegates to the proper storage method.
1428 sub _rs_update_delete {
1429 my ($self, $op, $values) = @_;
1431 my $rsrc = $self->result_source;
1433 # if a condition exists we need to strip all table qualifiers
1434 # if this is not possible we'll force a subquery below
1435 my $cond = $rsrc->schema->storage->_strip_cond_qualifiers ($self->{cond});
1437 my $needs_group_by_subq = $self->_has_resolved_attr (qw/collapse group_by -join/);
1438 my $needs_subq = $needs_group_by_subq || (not defined $cond) || $self->_has_resolved_attr(qw/rows offset/);
1440 if ($needs_group_by_subq or $needs_subq) {
1442 # make a new $rs selecting only the PKs (that's all we really need)
1443 my $attrs = $self->_resolved_attrs_copy;
1446 delete $attrs->{$_} for qw/collapse _collapse_order_by select _prefetch_select as/;
1447 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } ($self->result_source->_pri_cols) ];
1449 if ($needs_group_by_subq) {
1450 # make sure no group_by was supplied, or if there is one - make sure it matches
1451 # the columns compiled above perfectly. Anything else can not be sanely executed
1452 # on most databases so croak right then and there
1454 if (my $g = $attrs->{group_by}) {
1455 my @current_group_by = map
1456 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1461 join ("\x00", sort @current_group_by)
1463 join ("\x00", sort @{$attrs->{columns}} )
1465 $self->throw_exception (
1466 "You have just attempted a $op operation on a resultset which does group_by"
1467 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1468 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1469 . ' kind of queries. Please retry the operation with a modified group_by or'
1470 . ' without using one at all.'
1475 $attrs->{group_by} = $attrs->{columns};
1479 my $subrs = (ref $self)->new($rsrc, $attrs);
1480 return $self->result_source->storage->_subq_update_delete($subrs, $op, $values);
1483 return $rsrc->storage->$op(
1485 $op eq 'update' ? $values : (),
1495 =item Arguments: \%values
1497 =item Return Value: $storage_rv
1501 Sets the specified columns in the resultset to the supplied values in a
1502 single query. Note that this will not run any accessor/set_column/update
1503 triggers, nor will it update any row object instances derived from this
1504 resultset (this includes the contents of the L<resultset cache|/set_cache>
1505 if any). See L</update_all> if you need to execute any on-update
1506 triggers or cascades defined either by you or a
1507 L<result component|DBIx::Class::Manual::Component/WHAT_IS_A_COMPONENT>.
1509 The return value is a pass through of what the underlying
1510 storage backend returned, and may vary. See L<DBI/execute> for the most
1516 my ($self, $values) = @_;
1517 $self->throw_exception('Values for update must be a hash')
1518 unless ref $values eq 'HASH';
1520 return $self->_rs_update_delete ('update', $values);
1527 =item Arguments: \%values
1529 =item Return Value: 1
1533 Fetches all objects and updates them one at a time via
1534 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
1535 triggers, while L</update> will not.
1540 my ($self, $values) = @_;
1541 $self->throw_exception('Values for update_all must be a hash')
1542 unless ref $values eq 'HASH';
1544 my $guard = $self->result_source->schema->txn_scope_guard;
1545 $_->update($values) for $self->all;
1554 =item Arguments: none
1556 =item Return Value: $storage_rv
1560 Deletes the rows matching this resultset in a single query. Note that this
1561 will not run any delete triggers, nor will it alter the
1562 L<in_storage|DBIx::Class::Row/in_storage> status of any row object instances
1563 derived from this resultset (this includes the contents of the
1564 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
1565 execute any on-delete triggers or cascades defined either by you or a
1566 L<result component|DBIx::Class::Manual::Component/WHAT_IS_A_COMPONENT>.
1568 The return value is a pass through of what the underlying storage backend
1569 returned, and may vary. See L<DBI/execute> for the most common case.
1575 $self->throw_exception('delete does not accept any arguments')
1578 return $self->_rs_update_delete ('delete');
1585 =item Arguments: none
1587 =item Return Value: 1
1591 Fetches all objects and deletes them one at a time via
1592 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
1593 triggers, while L</delete> will not.
1599 $self->throw_exception('delete_all does not accept any arguments')
1602 my $guard = $self->result_source->schema->txn_scope_guard;
1603 $_->delete for $self->all;
1612 =item Arguments: \@data;
1616 Accepts either an arrayref of hashrefs or alternatively an arrayref of arrayrefs.
1617 For the arrayref of hashrefs style each hashref should be a structure suitable
1618 forsubmitting to a $resultset->create(...) method.
1620 In void context, C<insert_bulk> in L<DBIx::Class::Storage::DBI> is used
1621 to insert the data, as this is a faster method.
1623 Otherwise, each set of data is inserted into the database using
1624 L<DBIx::Class::ResultSet/create>, and the resulting objects are
1625 accumulated into an array. The array itself, or an array reference
1626 is returned depending on scalar or list context.
1628 Example: Assuming an Artist Class that has many CDs Classes relating:
1630 my $Artist_rs = $schema->resultset("Artist");
1632 ## Void Context Example
1633 $Artist_rs->populate([
1634 { artistid => 4, name => 'Manufactured Crap', cds => [
1635 { title => 'My First CD', year => 2006 },
1636 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
1639 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
1640 { title => 'My parents sold me to a record company', year => 2005 },
1641 { title => 'Why Am I So Ugly?', year => 2006 },
1642 { title => 'I Got Surgery and am now Popular', year => 2007 }
1647 ## Array Context Example
1648 my ($ArtistOne, $ArtistTwo, $ArtistThree) = $Artist_rs->populate([
1649 { name => "Artist One"},
1650 { name => "Artist Two"},
1651 { name => "Artist Three", cds=> [
1652 { title => "First CD", year => 2007},
1653 { title => "Second CD", year => 2008},
1657 print $ArtistOne->name; ## response is 'Artist One'
1658 print $ArtistThree->cds->count ## reponse is '2'
1660 For the arrayref of arrayrefs style, the first element should be a list of the
1661 fieldsnames to which the remaining elements are rows being inserted. For
1664 $Arstist_rs->populate([
1665 [qw/artistid name/],
1666 [100, 'A Formally Unknown Singer'],
1667 [101, 'A singer that jumped the shark two albums ago'],
1668 [102, 'An actually cool singer'],
1671 Please note an important effect on your data when choosing between void and
1672 wantarray context. Since void context goes straight to C<insert_bulk> in
1673 L<DBIx::Class::Storage::DBI> this will skip any component that is overriding
1674 C<insert>. So if you are using something like L<DBIx-Class-UUIDColumns> to
1675 create primary keys for you, you will find that your PKs are empty. In this
1676 case you will have to use the wantarray context in order to create those
1684 # cruft placed in standalone method
1685 my $data = $self->_normalize_populate_args(@_);
1687 if(defined wantarray) {
1689 foreach my $item (@$data) {
1690 push(@created, $self->create($item));
1692 return wantarray ? @created : \@created;
1694 my $first = $data->[0];
1696 # if a column is a registered relationship, and is a non-blessed hash/array, consider
1697 # it relationship data
1698 my (@rels, @columns);
1699 for (keys %$first) {
1700 my $ref = ref $first->{$_};
1701 $self->result_source->has_relationship($_) && ($ref eq 'ARRAY' or $ref eq 'HASH')
1707 my @pks = $self->result_source->primary_columns;
1709 ## do the belongs_to relationships
1710 foreach my $index (0..$#$data) {
1712 # delegate to create() for any dataset without primary keys with specified relationships
1713 if (grep { !defined $data->[$index]->{$_} } @pks ) {
1715 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
1716 my @ret = $self->populate($data);
1722 foreach my $rel (@rels) {
1723 next unless ref $data->[$index]->{$rel} eq "HASH";
1724 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
1725 my ($reverse) = keys %{$self->result_source->reverse_relationship_info($rel)};
1726 my $related = $result->result_source->_resolve_condition(
1727 $result->result_source->relationship_info($reverse)->{cond},
1732 delete $data->[$index]->{$rel};
1733 $data->[$index] = {%{$data->[$index]}, %$related};
1735 push @columns, keys %$related if $index == 0;
1739 ## inherit the data locked in the conditions of the resultset
1740 my ($rs_data) = $self->_merge_with_rscond({});
1741 delete @{$rs_data}{@columns};
1742 my @inherit_cols = keys %$rs_data;
1743 my @inherit_data = values %$rs_data;
1745 ## do bulk insert on current row
1746 $self->result_source->storage->insert_bulk(
1747 $self->result_source,
1748 [@columns, @inherit_cols],
1749 [ map { [ @$_{@columns}, @inherit_data ] } @$data ],
1752 ## do the has_many relationships
1753 foreach my $item (@$data) {
1755 foreach my $rel (@rels) {
1756 next unless $item->{$rel} && ref $item->{$rel} eq "ARRAY";
1758 my $parent = $self->find({map { $_ => $item->{$_} } @pks})
1759 || $self->throw_exception('Cannot find the relating object.');
1761 my $child = $parent->$rel;
1763 my $related = $child->result_source->_resolve_condition(
1764 $parent->result_source->relationship_info($rel)->{cond},
1769 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
1770 my @populate = map { {%$_, %$related} } @rows_to_add;
1772 $child->populate( \@populate );
1779 # populate() argumnets went over several incarnations
1780 # What we ultimately support is AoH
1781 sub _normalize_populate_args {
1782 my ($self, $arg) = @_;
1784 if (ref $arg eq 'ARRAY') {
1785 if (ref $arg->[0] eq 'HASH') {
1788 elsif (ref $arg->[0] eq 'ARRAY') {
1790 my @colnames = @{$arg->[0]};
1791 foreach my $values (@{$arg}[1 .. $#$arg]) {
1792 push @ret, { map { $colnames[$_] => $values->[$_] } (0 .. $#colnames) };
1798 $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
1805 =item Arguments: none
1807 =item Return Value: $pager
1811 Return Value a L<Data::Page> object for the current resultset. Only makes
1812 sense for queries with a C<page> attribute.
1814 To get the full count of entries for a paged resultset, call
1815 C<total_entries> on the L<Data::Page> object.
1819 # make a wizard good for both a scalar and a hashref
1820 my $mk_lazy_count_wizard = sub {
1821 require Variable::Magic;
1823 my $stash = { total_rs => shift };
1824 my $slot = shift; # only used by the hashref magic
1826 my $magic = Variable::Magic::wizard (
1827 data => sub { $stash },
1833 # set value lazily, and dispell for good
1834 ${$_[0]} = $_[1]{total_rs}->count;
1835 Variable::Magic::dispell (${$_[0]}, $_[1]{magic_selfref});
1839 # an explicit set implies dispell as well
1840 # the unless() is to work around "fun and giggles" below
1841 Variable::Magic::dispell (${$_[0]}, $_[1]{magic_selfref})
1842 unless (caller(2))[3] eq 'DBIx::Class::ResultSet::pager';
1849 if ($_[2] eq $slot and !$_[1]{inactive}) {
1850 my $cnt = $_[1]{total_rs}->count;
1851 $_[0]->{$slot} = $cnt;
1853 # attempting to dispell in a fetch handle (works in store), seems
1854 # to invariable segfault on 5.10, 5.12, 5.13 :(
1855 # so use an inactivator instead
1856 #Variable::Magic::dispell (%{$_[0]}, $_[1]{magic_selfref});
1862 if (! $_[1]{inactive} and $_[2] eq $slot) {
1863 #Variable::Magic::dispell (%{$_[0]}, $_[1]{magic_selfref});
1865 unless (caller(2))[3] eq 'DBIx::Class::ResultSet::pager';
1872 $stash->{magic_selfref} = $magic;
1873 weaken ($stash->{magic_selfref}); # this fails on 5.8.1
1878 # the tie class for 5.8.1
1880 package # hide from pause
1881 DBIx::Class::__DBIC_LAZY_RS_COUNT__;
1882 use base qw/Tie::Hash/;
1884 sub FIRSTKEY { my $dummy = scalar keys %{$_[0]{data}}; each %{$_[0]{data}} }
1885 sub NEXTKEY { each %{$_[0]{data}} }
1886 sub EXISTS { exists $_[0]{data}{$_[1]} }
1887 sub DELETE { delete $_[0]{data}{$_[1]} }
1888 sub CLEAR { %{$_[0]{data}} = () }
1889 sub SCALAR { scalar %{$_[0]{data}} }
1892 $_[1]{data} = {%{$_[1]{selfref}}};
1893 %{$_[1]{selfref}} = ();
1894 Scalar::Util::weaken ($_[1]{selfref});
1895 return bless ($_[1], $_[0]);
1899 if ($_[1] eq $_[0]{slot}) {
1900 my $cnt = $_[0]{data}{$_[1]} = $_[0]{total_rs}->count;
1901 untie %{$_[0]{selfref}};
1902 %{$_[0]{selfref}} = %{$_[0]{data}};
1911 $_[0]{data}{$_[1]} = $_[2];
1912 if ($_[1] eq $_[0]{slot}) {
1913 untie %{$_[0]{selfref}};
1914 %{$_[0]{selfref}} = %{$_[0]{data}};
1923 return $self->{pager} if $self->{pager};
1925 if ($self->get_cache) {
1926 $self->throw_exception ('Pagers on cached resultsets are not supported');
1929 my $attrs = $self->{attrs};
1930 $self->throw_exception("Can't create pager for non-paged rs")
1931 unless $self->{attrs}{page};
1932 $attrs->{rows} ||= 10;
1934 # throw away the paging flags and re-run the count (possibly
1935 # with a subselect) to get the real total count
1936 my $count_attrs = { %$attrs };
1937 delete $count_attrs->{$_} for qw/rows offset page pager/;
1938 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
1941 ### the following may seem awkward and dirty, but it's a thought-experiment
1942 ### necessary for future development of DBIx::DS. Do *NOT* change this code
1943 ### before talking to ribasushi/mst
1945 my $pager = Data::Page->new(
1946 0, #start with an empty set
1948 $self->{attrs}{page},
1951 my $data_slot = 'total_entries';
1953 # Since we are interested in a cached value (once it's set - it's set), every
1954 # technique will detach from the magic-host once the time comes to fire the
1955 # ->count (or in the segfaulting case of >= 5.10 it will deactivate itself)
1957 if ($] < 5.008003) {
1958 # 5.8.1 throws 'Modification of a read-only value attempted' when one tries
1959 # to weakref the magic container :(
1961 tie (%$pager, 'DBIx::Class::__DBIC_LAZY_RS_COUNT__',
1962 { slot => $data_slot, total_rs => $total_rs, selfref => $pager }
1965 elsif ($] < 5.010) {
1966 # We can use magic on the hash value slot. It's interesting that the magic is
1967 # attached to the hash-slot, and does *not* stop working once I do the dummy
1968 # assignments after the cast()
1969 # tested on 5.8.3 and 5.8.9
1970 my $magic = $mk_lazy_count_wizard->($total_rs);
1971 Variable::Magic::cast ( $pager->{$data_slot}, $magic );
1973 # this is for fun and giggles
1974 $pager->{$data_slot} = -1;
1975 $pager->{$data_slot} = 0;
1977 # this does not work for scalars, but works with
1979 #my %vals = %$pager;
1984 # And the uvar magic
1985 # works on 5.10.1, 5.12.1 and 5.13.4 in its current form,
1986 # however see the wizard maker for more notes
1987 my $magic = $mk_lazy_count_wizard->($total_rs, $data_slot);
1988 Variable::Magic::cast ( %$pager, $magic );
1991 $pager->{$data_slot} = -1;
1992 $pager->{$data_slot} = 0;
2000 return $self->{pager} = $pager;
2007 =item Arguments: $page_number
2009 =item Return Value: $rs
2013 Returns a resultset for the $page_number page of the resultset on which page
2014 is called, where each page contains a number of rows equal to the 'rows'
2015 attribute set on the resultset (10 by default).
2020 my ($self, $page) = @_;
2021 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2028 =item Arguments: \%vals
2030 =item Return Value: $rowobject
2034 Creates a new row object in the resultset's result class and returns
2035 it. The row is not inserted into the database at this point, call
2036 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2037 will tell you whether the row object has been inserted or not.
2039 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2044 my ($self, $values) = @_;
2045 $self->throw_exception( "new_result needs a hash" )
2046 unless (ref $values eq 'HASH');
2048 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2052 @$cols_from_relations
2053 ? (-cols_from_relations => $cols_from_relations)
2055 -source_handle => $self->_source_handle,
2056 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2059 return $self->result_class->new(\%new);
2062 # _merge_with_rscond
2064 # Takes a simple hash of K/V data and returns its copy merged with the
2065 # condition already present on the resultset. Additionally returns an
2066 # arrayref of value/condition names, which were inferred from related
2067 # objects (this is needed for in-memory related objects)
2068 sub _merge_with_rscond {
2069 my ($self, $data) = @_;
2071 my (%new_data, @cols_from_relations);
2073 my $alias = $self->{attrs}{alias};
2075 if (! defined $self->{cond}) {
2076 # just massage $data below
2078 elsif ($self->{cond} eq $DBIx::Class::ResultSource::UNRESOLVABLE_CONDITION) {
2079 %new_data = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
2080 @cols_from_relations = keys %new_data;
2082 elsif (ref $self->{cond} ne 'HASH') {
2083 $self->throw_exception(
2084 "Can't abstract implicit construct, resultset condition not a hash"
2088 # precendence must be given to passed values over values inherited from
2089 # the cond, so the order here is important.
2090 my $collapsed_cond = $self->_collapse_cond($self->{cond});
2091 my %implied = %{$self->_remove_alias($collapsed_cond, $alias)};
2093 while ( my($col, $value) = each %implied ) {
2094 my $vref = ref $value;
2095 if ($vref eq 'HASH' && keys(%$value) && (keys %$value)[0] eq '=') {
2096 $new_data{$col} = $value->{'='};
2098 elsif( !$vref or $vref eq 'SCALAR' or blessed($value) ) {
2099 $new_data{$col} = $value;
2106 %{ $self->_remove_alias($data, $alias) },
2109 return (\%new_data, \@cols_from_relations);
2112 # _has_resolved_attr
2114 # determines if the resultset defines at least one
2115 # of the attributes supplied
2117 # used to determine if a subquery is neccessary
2119 # supports some virtual attributes:
2121 # This will scan for any joins being present on the resultset.
2122 # It is not a mere key-search but a deep inspection of {from}
2125 sub _has_resolved_attr {
2126 my ($self, @attr_names) = @_;
2128 my $attrs = $self->_resolved_attrs;
2132 for my $n (@attr_names) {
2133 if (grep { $n eq $_ } (qw/-join/) ) {
2134 $extra_checks{$n}++;
2138 my $attr = $attrs->{$n};
2140 next if not defined $attr;
2142 if (ref $attr eq 'HASH') {
2143 return 1 if keys %$attr;
2145 elsif (ref $attr eq 'ARRAY') {
2153 # a resolved join is expressed as a multi-level from
2155 $extra_checks{-join}
2157 ref $attrs->{from} eq 'ARRAY'
2159 @{$attrs->{from}} > 1
2167 # Recursively collapse the condition.
2169 sub _collapse_cond {
2170 my ($self, $cond, $collapsed) = @_;
2174 if (ref $cond eq 'ARRAY') {
2175 foreach my $subcond (@$cond) {
2176 next unless ref $subcond; # -or
2177 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2180 elsif (ref $cond eq 'HASH') {
2181 if (keys %$cond and (keys %$cond)[0] eq '-and') {
2182 foreach my $subcond (@{$cond->{-and}}) {
2183 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2187 foreach my $col (keys %$cond) {
2188 my $value = $cond->{$col};
2189 $collapsed->{$col} = $value;
2199 # Remove the specified alias from the specified query hash. A copy is made so
2200 # the original query is not modified.
2203 my ($self, $query, $alias) = @_;
2205 my %orig = %{ $query || {} };
2208 foreach my $key (keys %orig) {
2210 $unaliased{$key} = $orig{$key};
2213 $unaliased{$1} = $orig{$key}
2214 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2224 =item Arguments: none
2226 =item Return Value: \[ $sql, @bind ]
2230 Returns the SQL query and bind vars associated with the invocant.
2232 This is generally used as the RHS for a subquery.
2239 my $attrs = $self->_resolved_attrs_copy;
2244 # my ($sql, \@bind, \%dbi_bind_attrs) = _select_args_to_query (...)
2245 # $sql also has no wrapping parenthesis in list ctx
2247 my $sqlbind = $self->result_source->storage
2248 ->_select_args_to_query ($attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs);
2257 =item Arguments: \%vals, \%attrs?
2259 =item Return Value: $rowobject
2263 my $artist = $schema->resultset('Artist')->find_or_new(
2264 { artist => 'fred' }, { key => 'artists' });
2266 $cd->cd_to_producer->find_or_new({ producer => $producer },
2267 { key => 'primary });
2269 Find an existing record from this resultset using L</find>. if none exists,
2270 instantiate a new result object and return it. The object will not be saved
2271 into your storage until you call L<DBIx::Class::Row/insert> on it.
2273 You most likely want this method when looking for existing rows using a unique
2274 constraint that is not the primary key, or looking for related rows.
2276 If you want objects to be saved immediately, use L</find_or_create> instead.
2278 B<Note>: Make sure to read the documentation of L</find> and understand the
2279 significance of the C<key> attribute, as its lack may skew your search, and
2280 subsequently result in spurious new objects.
2282 B<Note>: Take care when using C<find_or_new> with a table having
2283 columns with default values that you intend to be automatically
2284 supplied by the database (e.g. an auto_increment primary key column).
2285 In normal usage, the value of such columns should NOT be included at
2286 all in the call to C<find_or_new>, even when set to C<undef>.
2292 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2293 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2294 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2297 return $self->new_result($hash);
2304 =item Arguments: \%vals
2306 =item Return Value: a L<DBIx::Class::Row> $object
2310 Attempt to create a single new row or a row with multiple related rows
2311 in the table represented by the resultset (and related tables). This
2312 will not check for duplicate rows before inserting, use
2313 L</find_or_create> to do that.
2315 To create one row for this resultset, pass a hashref of key/value
2316 pairs representing the columns of the table and the values you wish to
2317 store. If the appropriate relationships are set up, foreign key fields
2318 can also be passed an object representing the foreign row, and the
2319 value will be set to its primary key.
2321 To create related objects, pass a hashref of related-object column values
2322 B<keyed on the relationship name>. If the relationship is of type C<multi>
2323 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2324 The process will correctly identify columns holding foreign keys, and will
2325 transparently populate them from the keys of the corresponding relation.
2326 This can be applied recursively, and will work correctly for a structure
2327 with an arbitrary depth and width, as long as the relationships actually
2328 exists and the correct column data has been supplied.
2331 Instead of hashrefs of plain related data (key/value pairs), you may
2332 also pass new or inserted objects. New objects (not inserted yet, see
2333 L</new>), will be inserted into their appropriate tables.
2335 Effectively a shortcut for C<< ->new_result(\%vals)->insert >>.
2337 Example of creating a new row.
2339 $person_rs->create({
2340 name=>"Some Person",
2341 email=>"somebody@someplace.com"
2344 Example of creating a new row and also creating rows in a related C<has_many>
2345 or C<has_one> resultset. Note Arrayref.
2348 { artistid => 4, name => 'Manufactured Crap', cds => [
2349 { title => 'My First CD', year => 2006 },
2350 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2355 Example of creating a new row and also creating a row in a related
2356 C<belongs_to> resultset. Note Hashref.
2359 title=>"Music for Silly Walks",
2362 name=>"Silly Musician",
2370 When subclassing ResultSet never attempt to override this method. Since
2371 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2372 lot of the internals simply never call it, so your override will be
2373 bypassed more often than not. Override either L<new|DBIx::Class::Row/new>
2374 or L<insert|DBIx::Class::Row/insert> depending on how early in the
2375 L</create> process you need to intervene.
2382 my ($self, $attrs) = @_;
2383 $self->throw_exception( "create needs a hashref" )
2384 unless ref $attrs eq 'HASH';
2385 return $self->new_result($attrs)->insert;
2388 =head2 find_or_create
2392 =item Arguments: \%vals, \%attrs?
2394 =item Return Value: $rowobject
2398 $cd->cd_to_producer->find_or_create({ producer => $producer },
2399 { key => 'primary' });
2401 Tries to find a record based on its primary key or unique constraints; if none
2402 is found, creates one and returns that instead.
2404 my $cd = $schema->resultset('CD')->find_or_create({
2406 artist => 'Massive Attack',
2407 title => 'Mezzanine',
2411 Also takes an optional C<key> attribute, to search by a specific key or unique
2412 constraint. For example:
2414 my $cd = $schema->resultset('CD')->find_or_create(
2416 artist => 'Massive Attack',
2417 title => 'Mezzanine',
2419 { key => 'cd_artist_title' }
2422 B<Note>: Make sure to read the documentation of L</find> and understand the
2423 significance of the C<key> attribute, as its lack may skew your search, and
2424 subsequently result in spurious row creation.
2426 B<Note>: Because find_or_create() reads from the database and then
2427 possibly inserts based on the result, this method is subject to a race
2428 condition. Another process could create a record in the table after
2429 the find has completed and before the create has started. To avoid
2430 this problem, use find_or_create() inside a transaction.
2432 B<Note>: Take care when using C<find_or_create> with a table having
2433 columns with default values that you intend to be automatically
2434 supplied by the database (e.g. an auto_increment primary key column).
2435 In normal usage, the value of such columns should NOT be included at
2436 all in the call to C<find_or_create>, even when set to C<undef>.
2438 See also L</find> and L</update_or_create>. For information on how to declare
2439 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2443 sub find_or_create {
2445 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2446 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2447 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2450 return $self->create($hash);
2453 =head2 update_or_create
2457 =item Arguments: \%col_values, { key => $unique_constraint }?
2459 =item Return Value: $row_object
2463 $resultset->update_or_create({ col => $val, ... });
2465 Like L</find_or_create>, but if a row is found it is immediately updated via
2466 C<< $found_row->update (\%col_values) >>.
2469 Takes an optional C<key> attribute to search on a specific unique constraint.
2472 # In your application
2473 my $cd = $schema->resultset('CD')->update_or_create(
2475 artist => 'Massive Attack',
2476 title => 'Mezzanine',
2479 { key => 'cd_artist_title' }
2482 $cd->cd_to_producer->update_or_create({
2483 producer => $producer,
2489 B<Note>: Make sure to read the documentation of L</find> and understand the
2490 significance of the C<key> attribute, as its lack may skew your search, and
2491 subsequently result in spurious row creation.
2493 B<Note>: Take care when using C<update_or_create> with a table having
2494 columns with default values that you intend to be automatically
2495 supplied by the database (e.g. an auto_increment primary key column).
2496 In normal usage, the value of such columns should NOT be included at
2497 all in the call to C<update_or_create>, even when set to C<undef>.
2499 See also L</find> and L</find_or_create>. For information on how to declare
2500 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2504 sub update_or_create {
2506 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2507 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2509 my $row = $self->find($cond, $attrs);
2511 $row->update($cond);
2515 return $self->create($cond);
2518 =head2 update_or_new
2522 =item Arguments: \%col_values, { key => $unique_constraint }?
2524 =item Return Value: $rowobject
2528 $resultset->update_or_new({ col => $val, ... });
2530 Like L</find_or_new> but if a row is found it is immediately updated via
2531 C<< $found_row->update (\%col_values) >>.
2535 # In your application
2536 my $cd = $schema->resultset('CD')->update_or_new(
2538 artist => 'Massive Attack',
2539 title => 'Mezzanine',
2542 { key => 'cd_artist_title' }
2545 if ($cd->in_storage) {
2546 # the cd was updated
2549 # the cd is not yet in the database, let's insert it
2553 B<Note>: Make sure to read the documentation of L</find> and understand the
2554 significance of the C<key> attribute, as its lack may skew your search, and
2555 subsequently result in spurious new objects.
2557 B<Note>: Take care when using C<update_or_new> with a table having
2558 columns with default values that you intend to be automatically
2559 supplied by the database (e.g. an auto_increment primary key column).
2560 In normal usage, the value of such columns should NOT be included at
2561 all in the call to C<update_or_new>, even when set to C<undef>.
2563 See also L</find>, L</find_or_create> and L</find_or_new>.
2569 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
2570 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2572 my $row = $self->find( $cond, $attrs );
2573 if ( defined $row ) {
2574 $row->update($cond);
2578 return $self->new_result($cond);
2585 =item Arguments: none
2587 =item Return Value: \@cache_objects | undef
2591 Gets the contents of the cache for the resultset, if the cache is set.
2593 The cache is populated either by using the L</prefetch> attribute to
2594 L</search> or by calling L</set_cache>.
2606 =item Arguments: \@cache_objects
2608 =item Return Value: \@cache_objects
2612 Sets the contents of the cache for the resultset. Expects an arrayref
2613 of objects of the same class as those produced by the resultset. Note that
2614 if the cache is set the resultset will return the cached objects rather
2615 than re-querying the database even if the cache attr is not set.
2617 The contents of the cache can also be populated by using the
2618 L</prefetch> attribute to L</search>.
2623 my ( $self, $data ) = @_;
2624 $self->throw_exception("set_cache requires an arrayref")
2625 if defined($data) && (ref $data ne 'ARRAY');
2626 $self->{all_cache} = $data;
2633 =item Arguments: none
2635 =item Return Value: undef
2639 Clears the cache for the resultset.
2644 shift->set_cache(undef);
2651 =item Arguments: none
2653 =item Return Value: true, if the resultset has been paginated
2661 return !!$self->{attrs}{page};
2668 =item Arguments: none
2670 =item Return Value: true, if the resultset has been ordered with C<order_by>.
2678 return scalar $self->result_source->storage->_extract_order_columns($self->{attrs}{order_by});
2681 =head2 related_resultset
2685 =item Arguments: $relationship_name
2687 =item Return Value: $resultset
2691 Returns a related resultset for the supplied relationship name.
2693 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
2697 sub related_resultset {
2698 my ($self, $rel) = @_;
2700 $self->{related_resultsets} ||= {};
2701 return $self->{related_resultsets}{$rel} ||= do {
2702 my $rsrc = $self->result_source;
2703 my $rel_info = $rsrc->relationship_info($rel);
2705 $self->throw_exception(
2706 "search_related: result source '" . $rsrc->source_name .
2707 "' has no such relationship $rel")
2710 my $attrs = $self->_chain_relationship($rel);
2712 my $join_count = $attrs->{seen_join}{$rel};
2714 my $alias = $self->result_source->storage
2715 ->relname_to_table_alias($rel, $join_count);
2717 # since this is search_related, and we already slid the select window inwards
2718 # (the select/as attrs were deleted in the beginning), we need to flip all
2719 # left joins to inner, so we get the expected results
2720 # read the comment on top of the actual function to see what this does
2721 $attrs->{from} = $rsrc->schema->storage->_inner_join_to_node ($attrs->{from}, $alias);
2724 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
2725 delete @{$attrs}{qw(result_class alias)};
2729 if (my $cache = $self->get_cache) {
2730 if ($cache->[0] && $cache->[0]->related_resultset($rel)->get_cache) {
2731 $new_cache = [ map { @{$_->related_resultset($rel)->get_cache} }
2736 my $rel_source = $rsrc->related_source($rel);
2740 # The reason we do this now instead of passing the alias to the
2741 # search_rs below is that if you wrap/overload resultset on the
2742 # source you need to know what alias it's -going- to have for things
2743 # to work sanely (e.g. RestrictWithObject wants to be able to add
2744 # extra query restrictions, and these may need to be $alias.)
2746 my $rel_attrs = $rel_source->resultset_attributes;
2747 local $rel_attrs->{alias} = $alias;
2749 $rel_source->resultset
2753 where => $attrs->{where},
2756 $new->set_cache($new_cache) if $new_cache;
2761 =head2 current_source_alias
2765 =item Arguments: none
2767 =item Return Value: $source_alias
2771 Returns the current table alias for the result source this resultset is built
2772 on, that will be used in the SQL query. Usually it is C<me>.
2774 Currently the source alias that refers to the result set returned by a
2775 L</search>/L</find> family method depends on how you got to the resultset: it's
2776 C<me> by default, but eg. L</search_related> aliases it to the related result
2777 source name (and keeps C<me> referring to the original result set). The long
2778 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
2779 (and make this method unnecessary).
2781 Thus it's currently necessary to use this method in predefined queries (see
2782 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
2783 source alias of the current result set:
2785 # in a result set class
2787 my ($self, $user) = @_;
2789 my $me = $self->current_source_alias;
2791 return $self->search(
2792 "$me.modified" => $user->id,
2798 sub current_source_alias {
2801 return ($self->{attrs} || {})->{alias} || 'me';
2804 =head2 as_subselect_rs
2808 =item Arguments: none
2810 =item Return Value: $resultset
2814 Act as a barrier to SQL symbols. The resultset provided will be made into a
2815 "virtual view" by including it as a subquery within the from clause. From this
2816 point on, any joined tables are inaccessible to ->search on the resultset (as if
2817 it were simply where-filtered without joins). For example:
2819 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
2821 # 'x' now pollutes the query namespace
2823 # So the following works as expected
2824 my $ok_rs = $rs->search({'x.other' => 1});
2826 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
2827 # def) we look for one row with contradictory terms and join in another table
2828 # (aliased 'x_2') which we never use
2829 my $broken_rs = $rs->search({'x.name' => 'def'});
2831 my $rs2 = $rs->as_subselect_rs;
2833 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
2834 my $not_joined_rs = $rs2->search({'x.other' => 1});
2836 # works as expected: finds a 'table' row related to two x rows (abc and def)
2837 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
2839 Another example of when one might use this would be to select a subset of
2840 columns in a group by clause:
2842 my $rs = $schema->resultset('Bar')->search(undef, {
2843 group_by => [qw{ id foo_id baz_id }],
2844 })->as_subselect_rs->search(undef, {
2845 columns => [qw{ id foo_id }]
2848 In the above example normally columns would have to be equal to the group by,
2849 but because we isolated the group by into a subselect the above works.
2853 sub as_subselect_rs {
2856 my $attrs = $self->_resolved_attrs;
2858 my $fresh_rs = (ref $self)->new (
2859 $self->result_source
2862 # these pieces will be locked in the subquery
2863 delete $fresh_rs->{cond};
2864 delete @{$fresh_rs->{attrs}}{qw/where bind/};
2866 return $fresh_rs->search( {}, {
2868 $attrs->{alias} => $self->as_query,
2869 -alias => $attrs->{alias},
2870 -source_handle => $self->result_source->handle,
2872 alias => $attrs->{alias},
2876 # This code is called by search_related, and makes sure there
2877 # is clear separation between the joins before, during, and
2878 # after the relationship. This information is needed later
2879 # in order to properly resolve prefetch aliases (any alias
2880 # with a relation_chain_depth less than the depth of the
2881 # current prefetch is not considered)
2883 # The increments happen twice per join. An even number means a
2884 # relationship specified via a search_related, whereas an odd
2885 # number indicates a join/prefetch added via attributes
2887 # Also this code will wrap the current resultset (the one we
2888 # chain to) in a subselect IFF it contains limiting attributes
2889 sub _chain_relationship {
2890 my ($self, $rel) = @_;
2891 my $source = $self->result_source;
2892 my $attrs = { %{$self->{attrs}||{}} };
2894 # we need to take the prefetch the attrs into account before we
2895 # ->_resolve_join as otherwise they get lost - captainL
2896 my $join = $self->_merge_attr( $attrs->{join}, $attrs->{prefetch} );
2898 delete @{$attrs}{qw/join prefetch collapse group_by distinct select as columns +select +as +columns/};
2900 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
2903 my @force_subq_attrs = qw/offset rows group_by having/;
2906 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
2908 $self->_has_resolved_attr (@force_subq_attrs)
2910 # Nuke the prefetch (if any) before the new $rs attrs
2911 # are resolved (prefetch is useless - we are wrapping
2912 # a subquery anyway).
2913 my $rs_copy = $self->search;
2914 $rs_copy->{attrs}{join} = $self->_merge_attr (
2915 $rs_copy->{attrs}{join},
2916 delete $rs_copy->{attrs}{prefetch},
2920 -source_handle => $source->handle,
2921 -alias => $attrs->{alias},
2922 $attrs->{alias} => $rs_copy->as_query,
2924 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
2925 $seen->{-relation_chain_depth} = 0;
2927 elsif ($attrs->{from}) { #shallow copy suffices
2928 $from = [ @{$attrs->{from}} ];
2932 -source_handle => $source->handle,
2933 -alias => $attrs->{alias},
2934 $attrs->{alias} => $source->from,
2938 my $jpath = ($seen->{-relation_chain_depth})
2939 ? $from->[-1][0]{-join_path}
2942 my @requested_joins = $source->_resolve_join(
2949 push @$from, @requested_joins;
2951 $seen->{-relation_chain_depth}++;
2953 # if $self already had a join/prefetch specified on it, the requested
2954 # $rel might very well be already included. What we do in this case
2955 # is effectively a no-op (except that we bump up the chain_depth on
2956 # the join in question so we could tell it *is* the search_related)
2959 # we consider the last one thus reverse
2960 for my $j (reverse @requested_joins) {
2961 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
2962 if ($rel eq $last_j) {
2963 $j->[0]{-relation_chain_depth}++;
2969 unless ($already_joined) {
2970 push @$from, $source->_resolve_join(
2978 $seen->{-relation_chain_depth}++;
2980 return {%$attrs, from => $from, seen_join => $seen};
2983 # too many times we have to do $attrs = { %{$self->_resolved_attrs} }
2984 sub _resolved_attrs_copy {
2986 return { %{$self->_resolved_attrs (@_)} };
2989 sub _resolved_attrs {
2991 return $self->{_attrs} if $self->{_attrs};
2993 my $attrs = { %{ $self->{attrs} || {} } };
2994 my $source = $self->result_source;
2995 my $alias = $attrs->{alias};
2997 $attrs->{columns} ||= delete $attrs->{cols} if exists $attrs->{cols};
3000 # build columns (as long as select isn't set) into a set of as/select hashes
3001 unless ( $attrs->{select} ) {
3004 if ( ref $attrs->{columns} eq 'ARRAY' ) {
3005 @cols = @{ delete $attrs->{columns}}
3006 } elsif ( defined $attrs->{columns} ) {
3007 @cols = delete $attrs->{columns}
3009 @cols = $source->columns
3013 if ( ref $_ eq 'HASH' ) {
3016 my $key = /^\Q${alias}.\E(.+)$/
3022 push @colbits, { $key => $value };
3027 # add the additional columns on
3028 foreach (qw{include_columns +columns}) {
3029 if ( $attrs->{$_} ) {
3030 my @list = ( ref($attrs->{$_}) eq 'ARRAY' )
3031 ? @{ delete $attrs->{$_} }
3032 : delete $attrs->{$_};
3034 if ( ref($_) eq 'HASH' ) {
3037 my $key = ( split /\./, $_ )[-1];
3038 my $value = ( /\./ ? $_ : "$alias.$_" );
3039 push @colbits, { $key => $value };
3045 # start with initial select items
3046 if ( $attrs->{select} ) {
3048 ( ref $attrs->{select} eq 'ARRAY' )
3049 ? [ @{ $attrs->{select} } ]
3050 : [ $attrs->{select} ];
3052 if ( $attrs->{as} ) {
3055 ref $attrs->{as} eq 'ARRAY'
3056 ? [ @{ $attrs->{as} } ]
3060 $attrs->{as} = [ map {
3061 m/^\Q${alias}.\E(.+)$/
3064 } @{ $attrs->{select} }
3070 # otherwise we intialise select & as to empty
3071 $attrs->{select} = [];
3075 # now add colbits to select/as
3076 push @{ $attrs->{select} }, map values %{$_}, @colbits;
3077 push @{ $attrs->{as} }, map keys %{$_}, @colbits;
3079 if ( my $adds = delete $attrs->{'+select'} ) {
3080 $adds = [$adds] unless ref $adds eq 'ARRAY';
3081 push @{ $attrs->{select} },
3082 map { /\./ || ref $_ ? $_ : "$alias.$_" } @$adds;
3084 if ( my $adds = delete $attrs->{'+as'} ) {
3085 $adds = [$adds] unless ref $adds eq 'ARRAY';
3086 push @{ $attrs->{as} }, @$adds;
3089 $attrs->{from} ||= [{
3090 -source_handle => $source->handle,
3091 -alias => $self->{attrs}{alias},
3092 $self->{attrs}{alias} => $source->from,
3095 if ( $attrs->{join} || $attrs->{prefetch} ) {
3097 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3098 if ref $attrs->{from} ne 'ARRAY';
3100 my $join = delete $attrs->{join} || {};
3102 if ( defined $attrs->{prefetch} ) {
3103 $join = $self->_merge_attr( $join, $attrs->{prefetch} );
3106 $attrs->{from} = # have to copy here to avoid corrupting the original
3108 @{ $attrs->{from} },
3109 $source->_resolve_join(
3112 { %{ $attrs->{seen_join} || {} } },
3113 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3114 ? $attrs->{from}[-1][0]{-join_path}
3121 if ( defined $attrs->{order_by} ) {
3122 $attrs->{order_by} = (
3123 ref( $attrs->{order_by} ) eq 'ARRAY'
3124 ? [ @{ $attrs->{order_by} } ]
3125 : [ $attrs->{order_by} || () ]
3129 if ($attrs->{group_by} and ref $attrs->{group_by} ne 'ARRAY') {
3130 $attrs->{group_by} = [ $attrs->{group_by} ];
3133 # generate the distinct induced group_by early, as prefetch will be carried via a
3134 # subquery (since a group_by is present)
3135 if (delete $attrs->{distinct}) {
3136 if ($attrs->{group_by}) {
3137 carp ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3140 $attrs->{group_by} = $source->storage->_group_over_selection (
3141 @{$attrs}{qw/from select order_by/}
3146 $attrs->{collapse} ||= {};
3147 if ( my $prefetch = delete $attrs->{prefetch} ) {
3148 $prefetch = $self->_merge_attr( {}, $prefetch );
3150 my $prefetch_ordering = [];
3152 # this is a separate structure (we don't look in {from} directly)
3153 # as the resolver needs to shift things off the lists to work
3154 # properly (identical-prefetches on different branches)
3156 if (ref $attrs->{from} eq 'ARRAY') {
3158 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3160 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3161 next unless $j->[0]{-alias};
3162 next unless $j->[0]{-join_path};
3163 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3165 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3168 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3169 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3174 $source->_resolve_prefetch( $prefetch, $alias, $join_map, $prefetch_ordering, $attrs->{collapse} );
3176 # we need to somehow mark which columns came from prefetch
3177 $attrs->{_prefetch_select} = [ map { $_->[0] } @prefetch ];
3179 push @{ $attrs->{select} }, @{$attrs->{_prefetch_select}};
3180 push @{ $attrs->{as} }, (map { $_->[1] } @prefetch);
3182 push( @{$attrs->{order_by}}, @$prefetch_ordering );
3183 $attrs->{_collapse_order_by} = \@$prefetch_ordering;
3186 # if both page and offset are specified, produce a combined offset
3187 # even though it doesn't make much sense, this is what pre 081xx has
3189 if (my $page = delete $attrs->{page}) {
3191 ($attrs->{rows} * ($page - 1))
3193 ($attrs->{offset} || 0)
3197 return $self->{_attrs} = $attrs;
3201 my ($self, $attr) = @_;
3203 if (ref $attr eq 'HASH') {
3204 return $self->_rollout_hash($attr);
3205 } elsif (ref $attr eq 'ARRAY') {
3206 return $self->_rollout_array($attr);
3212 sub _rollout_array {
3213 my ($self, $attr) = @_;
3216 foreach my $element (@{$attr}) {
3217 if (ref $element eq 'HASH') {
3218 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3219 } elsif (ref $element eq 'ARRAY') {
3220 # XXX - should probably recurse here
3221 push( @rolled_array, @{$self->_rollout_array($element)} );
3223 push( @rolled_array, $element );
3226 return \@rolled_array;
3230 my ($self, $attr) = @_;
3233 foreach my $key (keys %{$attr}) {
3234 push( @rolled_array, { $key => $attr->{$key} } );
3236 return \@rolled_array;
3239 sub _calculate_score {
3240 my ($self, $a, $b) = @_;
3242 if (defined $a xor defined $b) {
3245 elsif (not defined $a) {
3249 if (ref $b eq 'HASH') {
3250 my ($b_key) = keys %{$b};
3251 if (ref $a eq 'HASH') {
3252 my ($a_key) = keys %{$a};
3253 if ($a_key eq $b_key) {
3254 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3259 return ($a eq $b_key) ? 1 : 0;
3262 if (ref $a eq 'HASH') {
3263 my ($a_key) = keys %{$a};
3264 return ($b eq $a_key) ? 1 : 0;
3266 return ($b eq $a) ? 1 : 0;
3272 my ($self, $orig, $import) = @_;
3274 return $import unless defined($orig);
3275 return $orig unless defined($import);
3277 $orig = $self->_rollout_attr($orig);
3278 $import = $self->_rollout_attr($import);
3281 foreach my $import_element ( @{$import} ) {
3282 # find best candidate from $orig to merge $b_element into
3283 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3284 foreach my $orig_element ( @{$orig} ) {
3285 my $score = $self->_calculate_score( $orig_element, $import_element );
3286 if ($score > $best_candidate->{score}) {
3287 $best_candidate->{position} = $position;
3288 $best_candidate->{score} = $score;
3292 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3294 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3295 push( @{$orig}, $import_element );
3297 my $orig_best = $orig->[$best_candidate->{position}];
3298 # merge orig_best and b_element together and replace original with merged
3299 if (ref $orig_best ne 'HASH') {
3300 $orig->[$best_candidate->{position}] = $import_element;
3301 } elsif (ref $import_element eq 'HASH') {
3302 my ($key) = keys %{$orig_best};
3303 $orig->[$best_candidate->{position}] = { $key => $self->_merge_attr($orig_best->{$key}, $import_element->{$key}) };
3306 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3316 $self->_source_handle($_[0]->handle);
3318 $self->_source_handle->resolve;
3322 =head2 throw_exception
3324 See L<DBIx::Class::Schema/throw_exception> for details.
3328 sub throw_exception {
3331 if (ref $self && $self->_source_handle->schema) {
3332 $self->_source_handle->schema->throw_exception(@_)
3335 DBIx::Class::Exception->throw(@_);
3339 # XXX: FIXME: Attributes docs need clearing up
3343 Attributes are used to refine a ResultSet in various ways when
3344 searching for data. They can be passed to any method which takes an
3345 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3348 These are in no particular order:
3354 =item Value: ( $order_by | \@order_by | \%order_by )
3358 Which column(s) to order the results by.
3360 [The full list of suitable values is documented in
3361 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3364 If a single column name, or an arrayref of names is supplied, the
3365 argument is passed through directly to SQL. The hashref syntax allows
3366 for connection-agnostic specification of ordering direction:
3368 For descending order:
3370 order_by => { -desc => [qw/col1 col2 col3/] }
3372 For explicit ascending order:
3374 order_by => { -asc => 'col' }
3376 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3377 supported, although you are strongly encouraged to use the hashref
3378 syntax as outlined above.
3384 =item Value: \@columns
3388 Shortcut to request a particular set of columns to be retrieved. Each
3389 column spec may be a string (a table column name), or a hash (in which
3390 case the key is the C<as> value, and the value is used as the C<select>
3391 expression). Adds C<me.> onto the start of any column without a C<.> in
3392 it and sets C<select> from that, then auto-populates C<as> from
3393 C<select> as normal. (You may also use the C<cols> attribute, as in
3394 earlier versions of DBIC.)
3396 Essentially C<columns> does the same as L</select> and L</as>.
3398 columns => [ 'foo', { bar => 'baz' } ]
3402 select => [qw/foo baz/],
3409 =item Value: \@columns
3413 Indicates additional columns to be selected from storage. Works the same
3414 as L</columns> but adds columns to the selection. (You may also use the
3415 C<include_columns> attribute, as in earlier versions of DBIC). For
3418 $schema->resultset('CD')->search(undef, {
3419 '+columns' => ['artist.name'],
3423 would return all CDs and include a 'name' column to the information
3424 passed to object inflation. Note that the 'artist' is the name of the
3425 column (or relationship) accessor, and 'name' is the name of the column
3426 accessor in the related table.
3428 =head2 include_columns
3432 =item Value: \@columns
3436 Deprecated. Acts as a synonym for L</+columns> for backward compatibility.
3442 =item Value: \@select_columns
3446 Indicates which columns should be selected from the storage. You can use
3447 column names, or in the case of RDBMS back ends, function or stored procedure
3450 $rs = $schema->resultset('Employee')->search(undef, {
3453 { count => 'employeeid' },
3454 { max => { length => 'name' }, -as => 'longest_name' }
3459 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
3461 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
3462 use L</select>, to instruct DBIx::Class how to store the result of the column.
3463 Also note that the L</as> attribute has nothing to do with the SQL-side 'AS'
3464 identifier aliasing. You can however alias a function, so you can use it in
3465 e.g. an C<ORDER BY> clause. This is done via the C<-as> B<select function
3466 attribute> supplied as shown in the example above.
3472 Indicates additional columns to be selected from storage. Works the same as
3473 L</select> but adds columns to the default selection, instead of specifying
3482 Indicates additional column names for those added via L</+select>. See L</as>.
3490 =item Value: \@inflation_names
3494 Indicates column names for object inflation. That is L</as> indicates the
3495 slot name in which the column value will be stored within the
3496 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
3497 identifier by the C<get_column> method (or via the object accessor B<if one
3498 with the same name already exists>) as shown below. The L</as> attribute has
3499 B<nothing to do> with the SQL-side C<AS>. See L</select> for details.
3501 $rs = $schema->resultset('Employee')->search(undef, {
3504 { count => 'employeeid' },
3505 { max => { length => 'name' }, -as => 'longest_name' }
3514 If the object against which the search is performed already has an accessor
3515 matching a column name specified in C<as>, the value can be retrieved using
3516 the accessor as normal:
3518 my $name = $employee->name();
3520 If on the other hand an accessor does not exist in the object, you need to
3521 use C<get_column> instead:
3523 my $employee_count = $employee->get_column('employee_count');
3525 You can create your own accessors if required - see
3526 L<DBIx::Class::Manual::Cookbook> for details.
3532 =item Value: ($rel_name | \@rel_names | \%rel_names)
3536 Contains a list of relationships that should be joined for this query. For
3539 # Get CDs by Nine Inch Nails
3540 my $rs = $schema->resultset('CD')->search(
3541 { 'artist.name' => 'Nine Inch Nails' },
3542 { join => 'artist' }
3545 Can also contain a hash reference to refer to the other relation's relations.
3548 package MyApp::Schema::Track;
3549 use base qw/DBIx::Class/;
3550 __PACKAGE__->table('track');
3551 __PACKAGE__->add_columns(qw/trackid cd position title/);
3552 __PACKAGE__->set_primary_key('trackid');
3553 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
3556 # In your application
3557 my $rs = $schema->resultset('Artist')->search(
3558 { 'track.title' => 'Teardrop' },
3560 join => { cd => 'track' },
3561 order_by => 'artist.name',
3565 You need to use the relationship (not the table) name in conditions,
3566 because they are aliased as such. The current table is aliased as "me", so
3567 you need to use me.column_name in order to avoid ambiguity. For example:
3569 # Get CDs from 1984 with a 'Foo' track
3570 my $rs = $schema->resultset('CD')->search(
3573 'tracks.name' => 'Foo'
3575 { join => 'tracks' }
3578 If the same join is supplied twice, it will be aliased to <rel>_2 (and
3579 similarly for a third time). For e.g.
3581 my $rs = $schema->resultset('Artist')->search({
3582 'cds.title' => 'Down to Earth',
3583 'cds_2.title' => 'Popular',
3585 join => [ qw/cds cds/ ],
3588 will return a set of all artists that have both a cd with title 'Down
3589 to Earth' and a cd with title 'Popular'.
3591 If you want to fetch related objects from other tables as well, see C<prefetch>
3594 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
3600 =item Value: ($rel_name | \@rel_names | \%rel_names)
3604 Contains one or more relationships that should be fetched along with
3605 the main query (when they are accessed afterwards the data will
3606 already be available, without extra queries to the database). This is
3607 useful for when you know you will need the related objects, because it
3608 saves at least one query:
3610 my $rs = $schema->resultset('Tag')->search(
3619 The initial search results in SQL like the following:
3621 SELECT tag.*, cd.*, artist.* FROM tag
3622 JOIN cd ON tag.cd = cd.cdid
3623 JOIN artist ON cd.artist = artist.artistid
3625 L<DBIx::Class> has no need to go back to the database when we access the
3626 C<cd> or C<artist> relationships, which saves us two SQL statements in this
3629 Simple prefetches will be joined automatically, so there is no need
3630 for a C<join> attribute in the above search.
3632 C<prefetch> can be used with the following relationship types: C<belongs_to>,
3633 C<has_one> (or if you're using C<add_relationship>, any relationship declared
3634 with an accessor type of 'single' or 'filter'). A more complex example that
3635 prefetches an artists cds, the tracks on those cds, and the tags associated
3636 with that artist is given below (assuming many-to-many from artists to tags):
3638 my $rs = $schema->resultset('Artist')->search(
3642 { cds => 'tracks' },
3643 { artist_tags => 'tags' }
3649 B<NOTE:> If you specify a C<prefetch> attribute, the C<join> and C<select>
3650 attributes will be ignored.
3652 B<CAVEATs>: Prefetch does a lot of deep magic. As such, it may not behave
3653 exactly as you might expect.
3659 Prefetch uses the L</cache> to populate the prefetched relationships. This
3660 may or may not be what you want.
3664 If you specify a condition on a prefetched relationship, ONLY those
3665 rows that match the prefetched condition will be fetched into that relationship.
3666 This means that adding prefetch to a search() B<may alter> what is returned by
3667 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
3669 my $artist_rs = $schema->resultset('Artist')->search({
3675 my $count = $artist_rs->first->cds->count;
3677 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
3679 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
3681 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
3683 that cmp_ok() may or may not pass depending on the datasets involved. This
3684 behavior may or may not survive the 0.09 transition.
3696 Makes the resultset paged and specifies the page to retrieve. Effectively
3697 identical to creating a non-pages resultset and then calling ->page($page)
3700 If L<rows> attribute is not specified it defaults to 10 rows per page.
3702 When you have a paged resultset, L</count> will only return the number
3703 of rows in the page. To get the total, use the L</pager> and call
3704 C<total_entries> on it.
3714 Specifies the maximum number of rows for direct retrieval or the number of
3715 rows per page if the page attribute or method is used.
3721 =item Value: $offset
3725 Specifies the (zero-based) row number for the first row to be returned, or the
3726 of the first row of the first page if paging is used.
3732 =item Value: \@columns
3736 A arrayref of columns to group by. Can include columns of joined tables.
3738 group_by => [qw/ column1 column2 ... /]
3744 =item Value: $condition
3748 HAVING is a select statement attribute that is applied between GROUP BY and
3749 ORDER BY. It is applied to the after the grouping calculations have been
3752 having => { 'count(employee)' => { '>=', 100 } }
3758 =item Value: (0 | 1)
3762 Set to 1 to group by all columns. If the resultset already has a group_by
3763 attribute, this setting is ignored and an appropriate warning is issued.
3769 Adds to the WHERE clause.
3771 # only return rows WHERE deleted IS NULL for all searches
3772 __PACKAGE__->resultset_attributes({ where => { deleted => undef } }); )
3774 Can be overridden by passing C<< { where => undef } >> as an attribute
3781 Set to 1 to cache search results. This prevents extra SQL queries if you
3782 revisit rows in your ResultSet:
3784 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
3786 while( my $artist = $resultset->next ) {
3790 $rs->first; # without cache, this would issue a query
3792 By default, searches are not cached.
3794 For more examples of using these attributes, see
3795 L<DBIx::Class::Manual::Cookbook>.
3801 =item Value: ( 'update' | 'shared' )
3805 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT