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;
14 use Scalar::Util qw/blessed weaken/;
16 use Storable qw/nfreeze thaw/;
20 # De-duplication in _merge_attr() is disabled, but left in for reference
21 *__HM_DEDUP = sub () { 0 };
29 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class _source_handle/);
33 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
37 my $users_rs = $schema->resultset('User');
38 while( $user = $users_rs->next) {
39 print $user->username;
42 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
43 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
47 A ResultSet is an object which stores a set of conditions representing
48 a query. It is the backbone of DBIx::Class (i.e. the really
49 important/useful bit).
51 No SQL is executed on the database when a ResultSet is created, it
52 just stores all the conditions needed to create the query.
54 A basic ResultSet representing the data of an entire table is returned
55 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
56 L<Source|DBIx::Class::Manual::Glossary/Source> name.
58 my $users_rs = $schema->resultset('User');
60 A new ResultSet is returned from calling L</search> on an existing
61 ResultSet. The new one will contain all the conditions of the
62 original, plus any new conditions added in the C<search> call.
64 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
65 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
68 The query that the ResultSet represents is B<only> executed against
69 the database when these methods are called:
70 L</find>, L</next>, L</all>, L</first>, L</single>, L</count>.
72 If a resultset is used in a numeric context it returns the L</count>.
73 However, if it is used in a boolean context it is B<always> true. So if
74 you want to check if a resultset has any results, you must use C<if $rs
79 =head2 Chaining resultsets
81 Let's say you've got a query that needs to be run to return some data
82 to the user. But, you have an authorization system in place that
83 prevents certain users from seeing certain information. So, you want
84 to construct the basic query in one method, but add constraints to it in
89 my $request = $self->get_request; # Get a request object somehow.
90 my $schema = $self->get_schema; # Get the DBIC schema object somehow.
92 my $cd_rs = $schema->resultset('CD')->search({
93 title => $request->param('title'),
94 year => $request->param('year'),
97 $self->apply_security_policy( $cd_rs );
102 sub apply_security_policy {
111 =head3 Resolving conditions and attributes
113 When a resultset is chained from another resultset, conditions and
114 attributes with the same keys need resolving.
116 L</join>, L</prefetch>, L</+select>, L</+as> attributes are merged
117 into the existing ones from the original resultset.
119 The L</where> and L</having> attributes, and any search conditions, are
120 merged with an SQL C<AND> to the existing condition from the original
123 All other attributes are overridden by any new ones supplied in the
126 =head2 Multiple queries
128 Since a resultset just defines a query, you can do all sorts of
129 things with it with the same object.
131 # Don't hit the DB yet.
132 my $cd_rs = $schema->resultset('CD')->search({
133 title => 'something',
137 # Each of these hits the DB individually.
138 my $count = $cd_rs->count;
139 my $most_recent = $cd_rs->get_column('date_released')->max();
140 my @records = $cd_rs->all;
142 And it's not just limited to SELECT statements.
148 $cd_rs->create({ artist => 'Fred' });
150 Which is the same as:
152 $schema->resultset('CD')->create({
153 title => 'something',
158 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
166 =item Arguments: $source, \%$attrs
168 =item Return Value: $rs
172 The resultset constructor. Takes a source object (usually a
173 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
174 L</ATTRIBUTES> below). Does not perform any queries -- these are
175 executed as needed by the other methods.
177 Generally you won't need to construct a resultset manually. You'll
178 automatically get one from e.g. a L</search> called in scalar context:
180 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
182 IMPORTANT: If called on an object, proxies to new_result instead so
184 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
186 will return a CD object, not a ResultSet.
192 return $class->new_result(@_) if ref $class;
194 my ($source, $attrs) = @_;
195 $source = $source->handle
196 unless $source->isa('DBIx::Class::ResultSourceHandle');
197 $attrs = { %{$attrs||{}} };
199 if ($attrs->{page}) {
200 $attrs->{rows} ||= 10;
203 $attrs->{alias} ||= 'me';
205 # Creation of {} and bless separated to mitigate RH perl bug
206 # see https://bugzilla.redhat.com/show_bug.cgi?id=196836
208 _source_handle => $source,
209 cond => $attrs->{where},
217 $attrs->{result_class} || $source->resolve->result_class
227 =item Arguments: $cond, \%attrs?
229 =item Return Value: $resultset (scalar context), @row_objs (list context)
233 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
234 my $new_rs = $cd_rs->search({ year => 2005 });
236 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
237 # year = 2005 OR year = 2004
239 If you need to pass in additional attributes but no additional condition,
240 call it as C<search(undef, \%attrs)>.
242 # "SELECT name, artistid FROM $artist_table"
243 my @all_artists = $schema->resultset('Artist')->search(undef, {
244 columns => [qw/name artistid/],
247 For a list of attributes that can be passed to C<search>, see
248 L</ATTRIBUTES>. For more examples of using this function, see
249 L<Searching|DBIx::Class::Manual::Cookbook/Searching>. For a complete
250 documentation for the first argument, see L<SQL::Abstract>.
252 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
256 Note that L</search> does not process/deflate any of the values passed in the
257 L<SQL::Abstract>-compatible search condition structure. This is unlike other
258 condition-bound methods L</new>, L</create> and L</find>. The user must ensure
259 manually that any value passed to this method will stringify to something the
260 RDBMS knows how to deal with. A notable example is the handling of L<DateTime>
261 objects, for more info see:
262 L<DBIx::Class::Manual::Cookbook/Formatting_DateTime_objects_in_queries>.
268 my $rs = $self->search_rs( @_ );
270 my $want = wantarray;
274 elsif (defined $want) {
278 $self->throw_exception ('->search is *not* a mutator, calling it in void context makes no sense');
286 =item Arguments: $cond, \%attrs?
288 =item Return Value: $resultset
292 This method does the same exact thing as search() except it will
293 always return a resultset, even in list context.
300 # Special-case handling for (undef, undef).
301 if ( @_ == 2 && !defined $_[1] && !defined $_[0] ) {
306 $call_attrs = pop(@_) if (
307 @_ > 1 and ( ! defined $_[-1] or ref $_[-1] eq 'HASH' )
310 # see if we can keep the cache (no $rs changes)
312 my %safe = (alias => 1, cache => 1);
313 if ( ! List::Util::first { !$safe{$_} } keys %$call_attrs and (
316 ref $_[0] eq 'HASH' && ! keys %{$_[0]}
318 ref $_[0] eq 'ARRAY' && ! @{$_[0]}
320 $cache = $self->get_cache;
323 my $old_attrs = { %{$self->{attrs}} };
324 my $old_having = delete $old_attrs->{having};
325 my $old_where = delete $old_attrs->{where};
327 # reset the selector list
328 if (List::Util::first { exists $call_attrs->{$_} } qw{columns select as}) {
329 delete @{$old_attrs}{qw{select as columns +select +as +columns include_columns}};
332 my $new_attrs = { %{$old_attrs}, %{$call_attrs} };
334 # merge new attrs into inherited
335 foreach my $key (qw/join prefetch/) {
336 next unless exists $call_attrs->{$key};
337 $new_attrs->{$key} = $self->_merge_joinpref_attr($old_attrs->{$key}, $call_attrs->{$key});
339 foreach my $key (qw/+select +as +columns include_columns bind/) {
340 next unless exists $call_attrs->{$key};
341 $new_attrs->{$key} = $self->_merge_attr($old_attrs->{$key}, $call_attrs->{$key});
344 # rip apart the rest of @_, parse a condition
347 if (ref $_[0] eq 'HASH') {
348 (keys %{$_[0]}) ? $_[0] : undef
354 $self->throw_exception('Odd number of arguments to search')
362 carp 'search( %condition ) is deprecated, use search( \%condition ) instead'
363 if (@_ > 1 and ! $self->result_source->result_class->isa('DBIx::Class::CDBICompat') );
365 for ($old_where, $call_cond) {
367 $new_attrs->{where} = $self->_stack_cond (
368 $_, $new_attrs->{where}
373 if (defined $old_having) {
374 $new_attrs->{having} = $self->_stack_cond (
375 $old_having, $new_attrs->{having}
379 my $rs = (ref $self)->new($self->result_source, $new_attrs);
381 $rs->set_cache($cache) if ($cache);
387 my ($self, $left, $right) = @_;
388 if (defined $left xor defined $right) {
389 return defined $left ? $left : $right;
391 elsif (defined $left) {
392 return { -and => [ map
393 { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
401 =head2 search_literal
405 =item Arguments: $sql_fragment, @bind_values
407 =item Return Value: $resultset (scalar context), @row_objs (list context)
411 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
412 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
414 Pass a literal chunk of SQL to be added to the conditional part of the
417 CAVEAT: C<search_literal> is provided for Class::DBI compatibility and should
418 only be used in that context. C<search_literal> is a convenience method.
419 It is equivalent to calling $schema->search(\[]), but if you want to ensure
420 columns are bound correctly, use C<search>.
422 Example of how to use C<search> instead of C<search_literal>
424 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
425 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
428 See L<DBIx::Class::Manual::Cookbook/Searching> and
429 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
430 require C<search_literal>.
435 my ($self, $sql, @bind) = @_;
437 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
440 return $self->search(\[ $sql, map [ __DUMMY__ => $_ ], @bind ], ($attr || () ));
447 =item Arguments: \%columns_values | @pk_values, \%attrs?
449 =item Return Value: $row_object | undef
453 Finds and returns a single row based on supplied criteria. Takes either a
454 hashref with the same format as L</create> (including inference of foreign
455 keys from related objects), or a list of primary key values in the same
456 order as the L<primary columns|DBIx::Class::ResultSource/primary_columns>
457 declaration on the L</result_source>.
459 In either case an attempt is made to combine conditions already existing on
460 the resultset with the condition passed to this method.
462 To aid with preparing the correct query for the storage you may supply the
463 C<key> attribute, which is the name of a
464 L<unique constraint|DBIx::Class::ResultSource/add_unique_constraint> (the
465 unique constraint corresponding to the
466 L<primary columns|DBIx::Class::ResultSource/primary_columns> is always named
467 C<primary>). If the C<key> attribute has been supplied, and DBIC is unable
468 to construct a query that satisfies the named unique constraint fully (
469 non-NULL values for each column member of the constraint) an exception is
472 If no C<key> is specified, the search is carried over all unique constraints
473 which are fully defined by the available condition.
475 If no such constraint is found, C<find> currently defaults to a simple
476 C<< search->(\%column_values) >> which may or may not do what you expect.
477 Note that this fallback behavior may be deprecated in further versions. If
478 you need to search with arbitrary conditions - use L</search>. If the query
479 resulting from this fallback produces more than one row, a warning to the
480 effect is issued, though only the first row is constructed and returned as
483 In addition to C<key>, L</find> recognizes and applies standard
484 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
486 Note that if you have extra concerns about the correctness of the resulting
487 query you need to specify the C<key> attribute and supply the entire condition
488 as an argument to find (since it is not always possible to perform the
489 combination of the resultset condition with the supplied one, especially if
490 the resultset condition contains literal sql).
492 For example, to find a row by its primary key:
494 my $cd = $schema->resultset('CD')->find(5);
496 You can also find a row by a specific unique constraint:
498 my $cd = $schema->resultset('CD')->find(
500 artist => 'Massive Attack',
501 title => 'Mezzanine',
503 { key => 'cd_artist_title' }
506 See also L</find_or_create> and L</update_or_create>.
512 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
514 my $rsrc = $self->result_source;
516 # Parse out the condition from input
518 if (ref $_[0] eq 'HASH') {
519 $call_cond = { %{$_[0]} };
522 my $constraint = exists $attrs->{key} ? $attrs->{key} : 'primary';
523 my @c_cols = $rsrc->unique_constraint_columns($constraint);
525 $self->throw_exception(
526 "No constraint columns, maybe a malformed '$constraint' constraint?"
529 $self->throw_exception (
530 'find() expects either a column/value hashref, or a list of values '
531 . "corresponding to the columns of the specified unique constraint '$constraint'"
532 ) unless @c_cols == @_;
535 @{$call_cond}{@c_cols} = @_;
539 for my $key (keys %$call_cond) {
541 my $keyref = ref($call_cond->{$key})
543 my $relinfo = $rsrc->relationship_info($key)
545 my $val = delete $call_cond->{$key};
547 next if $keyref eq 'ARRAY'; # has_many for multi_create
549 my $rel_q = $rsrc->_resolve_condition(
550 $relinfo->{cond}, $val, $key
552 die "Can't handle complex relationship conditions in find" if ref($rel_q) ne 'HASH';
553 @related{keys %$rel_q} = values %$rel_q;
557 # relationship conditions take precedence (?)
558 @{$call_cond}{keys %related} = values %related;
560 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
562 if (exists $attrs->{key}) {
563 $final_cond = $self->_qualify_cond_columns (
565 $self->_build_unique_cond (
573 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
574 # This means that we got here after a merger of relationship conditions
575 # in ::Relationship::Base::search_related (the row method), and furthermore
576 # the relationship is of the 'single' type. This means that the condition
577 # provided by the relationship (already attached to $self) is sufficient,
578 # as there can be only one row in the database that would satisfy the
582 # no key was specified - fall down to heuristics mode:
583 # run through all unique queries registered on the resultset, and
584 # 'OR' all qualifying queries together
585 my (@unique_queries, %seen_column_combinations);
586 for my $c_name ($rsrc->unique_constraint_names) {
587 next if $seen_column_combinations{
588 join "\x00", sort $rsrc->unique_constraint_columns($c_name)
591 push @unique_queries, try {
592 $self->_build_unique_cond ($c_name, $call_cond)
596 $final_cond = @unique_queries
597 ? [ map { $self->_qualify_cond_columns($_, $alias) } @unique_queries ]
598 : $self->_non_unique_find_fallback ($call_cond, $attrs)
602 # Run the query, passing the result_class since it should propagate for find
603 my $rs = $self->search ($final_cond, {result_class => $self->result_class, %$attrs});
604 if (keys %{$rs->_resolved_attrs->{collapse}}) {
606 carp "Query returned more than one row" if $rs->next;
614 # This is a stop-gap method as agreed during the discussion on find() cleanup:
615 # http://lists.scsys.co.uk/pipermail/dbix-class/2010-October/009535.html
617 # It is invoked when find() is called in legacy-mode with insufficiently-unique
618 # condition. It is provided for overrides until a saner way forward is devised
620 # *NOTE* This is not a public method, and it's *GUARANTEED* to disappear down
621 # the road. Please adjust your tests accordingly to catch this situation early
622 # DBIx::Class::ResultSet->can('_non_unique_find_fallback') is reasonable
624 # The method will not be removed without an adequately complete replacement
625 # for strict-mode enforcement
626 sub _non_unique_find_fallback {
627 my ($self, $cond, $attrs) = @_;
629 return $self->_qualify_cond_columns(
631 exists $attrs->{alias}
633 : $self->{attrs}{alias}
638 sub _qualify_cond_columns {
639 my ($self, $cond, $alias) = @_;
641 my %aliased = %$cond;
642 for (keys %aliased) {
643 $aliased{"$alias.$_"} = delete $aliased{$_}
650 sub _build_unique_cond {
651 my ($self, $constraint_name, $extra_cond) = @_;
653 my @c_cols = $self->result_source->unique_constraint_columns($constraint_name);
655 # combination may fail if $self->{cond} is non-trivial
656 my ($final_cond) = try {
657 $self->_merge_with_rscond ($extra_cond)
662 # trim out everything not in $columns
663 $final_cond = { map { $_ => $final_cond->{$_} } @c_cols };
665 if (my @missing = grep { ! defined $final_cond->{$_} } (@c_cols) ) {
666 $self->throw_exception( sprintf ( "Unable to satisfy requested constraint '%s', no values for column(s): %s",
668 join (', ', map { "'$_'" } @missing),
675 =head2 search_related
679 =item Arguments: $rel, $cond, \%attrs?
681 =item Return Value: $new_resultset
685 $new_rs = $cd_rs->search_related('artist', {
689 Searches the specified relationship, optionally specifying a condition and
690 attributes for matching records. See L</ATTRIBUTES> for more information.
695 return shift->related_resultset(shift)->search(@_);
698 =head2 search_related_rs
700 This method works exactly the same as search_related, except that
701 it guarantees a resultset, even in list context.
705 sub search_related_rs {
706 return shift->related_resultset(shift)->search_rs(@_);
713 =item Arguments: none
715 =item Return Value: $cursor
719 Returns a storage-driven cursor to the given resultset. See
720 L<DBIx::Class::Cursor> for more information.
727 my $attrs = $self->_resolved_attrs_copy;
729 return $self->{cursor}
730 ||= $self->result_source->storage->select($attrs->{from}, $attrs->{select},
731 $attrs->{where},$attrs);
738 =item Arguments: $cond?
740 =item Return Value: $row_object | undef
744 my $cd = $schema->resultset('CD')->single({ year => 2001 });
746 Inflates the first result without creating a cursor if the resultset has
747 any records in it; if not returns C<undef>. Used by L</find> as a lean version
750 While this method can take an optional search condition (just like L</search>)
751 being a fast-code-path it does not recognize search attributes. If you need to
752 add extra joins or similar, call L</search> and then chain-call L</single> on the
753 L<DBIx::Class::ResultSet> returned.
759 As of 0.08100, this method enforces the assumption that the preceding
760 query returns only one row. If more than one row is returned, you will receive
763 Query returned more than one row
765 In this case, you should be using L</next> or L</find> instead, or if you really
766 know what you are doing, use the L</rows> attribute to explicitly limit the size
769 This method will also throw an exception if it is called on a resultset prefetching
770 has_many, as such a prefetch implies fetching multiple rows from the database in
771 order to assemble the resulting object.
778 my ($self, $where) = @_;
780 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
783 my $attrs = $self->_resolved_attrs_copy;
785 if (keys %{$attrs->{collapse}}) {
786 $self->throw_exception(
787 'single() can not be used on resultsets prefetching has_many. Use find( \%cond ) or next() instead'
792 if (defined $attrs->{where}) {
795 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
796 $where, delete $attrs->{where} ]
799 $attrs->{where} = $where;
803 my @data = $self->result_source->storage->select_single(
804 $attrs->{from}, $attrs->{select},
805 $attrs->{where}, $attrs
808 return (@data ? ($self->_construct_object(@data))[0] : undef);
814 # Recursively collapse the query, accumulating values for each column.
816 sub _collapse_query {
817 my ($self, $query, $collapsed) = @_;
821 if (ref $query eq 'ARRAY') {
822 foreach my $subquery (@$query) {
823 next unless ref $subquery; # -or
824 $collapsed = $self->_collapse_query($subquery, $collapsed);
827 elsif (ref $query eq 'HASH') {
828 if (keys %$query and (keys %$query)[0] eq '-and') {
829 foreach my $subquery (@{$query->{-and}}) {
830 $collapsed = $self->_collapse_query($subquery, $collapsed);
834 foreach my $col (keys %$query) {
835 my $value = $query->{$col};
836 $collapsed->{$col}{$value}++;
848 =item Arguments: $cond?
850 =item Return Value: $resultsetcolumn
854 my $max_length = $rs->get_column('length')->max;
856 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
861 my ($self, $column) = @_;
862 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
870 =item Arguments: $cond, \%attrs?
872 =item Return Value: $resultset (scalar context), @row_objs (list context)
876 # WHERE title LIKE '%blue%'
877 $cd_rs = $rs->search_like({ title => '%blue%'});
879 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
880 that this is simply a convenience method retained for ex Class::DBI users.
881 You most likely want to use L</search> with specific operators.
883 For more information, see L<DBIx::Class::Manual::Cookbook>.
885 This method is deprecated and will be removed in 0.09. Use L</search()>
886 instead. An example conversion is:
888 ->search_like({ foo => 'bar' });
892 ->search({ foo => { like => 'bar' } });
899 'search_like() is deprecated and will be removed in DBIC version 0.09.'
900 .' Instead use ->search({ x => { -like => "y%" } })'
901 .' (note the outer pair of {}s - they are important!)'
903 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
904 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
905 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
906 return $class->search($query, { %$attrs });
913 =item Arguments: $first, $last
915 =item Return Value: $resultset (scalar context), @row_objs (list context)
919 Returns a resultset or object list representing a subset of elements from the
920 resultset slice is called on. Indexes are from 0, i.e., to get the first
923 my ($one, $two, $three) = $rs->slice(0, 2);
928 my ($self, $min, $max) = @_;
929 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
930 $attrs->{offset} = $self->{attrs}{offset} || 0;
931 $attrs->{offset} += $min;
932 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
933 return $self->search(undef, $attrs);
934 #my $slice = (ref $self)->new($self->result_source, $attrs);
935 #return (wantarray ? $slice->all : $slice);
942 =item Arguments: none
944 =item Return Value: $result | undef
948 Returns the next element in the resultset (C<undef> is there is none).
950 Can be used to efficiently iterate over records in the resultset:
952 my $rs = $schema->resultset('CD')->search;
953 while (my $cd = $rs->next) {
957 Note that you need to store the resultset object, and call C<next> on it.
958 Calling C<< resultset('Table')->next >> repeatedly will always return the
959 first record from the resultset.
965 if (my $cache = $self->get_cache) {
966 $self->{all_cache_position} ||= 0;
967 return $cache->[$self->{all_cache_position}++];
969 if ($self->{attrs}{cache}) {
970 delete $self->{pager};
971 $self->{all_cache_position} = 1;
972 return ($self->all)[0];
974 if ($self->{stashed_objects}) {
975 my $obj = shift(@{$self->{stashed_objects}});
976 delete $self->{stashed_objects} unless @{$self->{stashed_objects}};
980 exists $self->{stashed_row}
981 ? @{delete $self->{stashed_row}}
982 : $self->cursor->next
984 return undef unless (@row);
985 my ($row, @more) = $self->_construct_object(@row);
986 $self->{stashed_objects} = \@more if @more;
990 sub _construct_object {
991 my ($self, @row) = @_;
993 my $info = $self->_collapse_result($self->{_attrs}{as}, \@row)
995 my @new = $self->result_class->inflate_result($self->result_source, @$info);
996 @new = $self->{_attrs}{record_filter}->(@new)
997 if exists $self->{_attrs}{record_filter};
1001 sub _collapse_result {
1002 my ($self, $as_proto, $row) = @_;
1006 # 'foo' => [ undef, 'foo' ]
1007 # 'foo.bar' => [ 'foo', 'bar' ]
1008 # 'foo.bar.baz' => [ 'foo.bar', 'baz' ]
1010 my @construct_as = map { [ (/^(?:(.*)\.)?([^.]+)$/) ] } @$as_proto;
1012 my %collapse = %{$self->{_attrs}{collapse}||{}};
1016 # if we're doing collapsing (has_many prefetch) we need to grab records
1017 # until the PK changes, so fill @pri_index. if not, we leave it empty so
1018 # we know we don't have to bother.
1020 # the reason for not using the collapse stuff directly is because if you
1021 # had for e.g. two artists in a row with no cds, the collapse info for
1022 # both would be NULL (undef) so you'd lose the second artist
1024 # store just the index so we can check the array positions from the row
1025 # without having to contruct the full hash
1027 if (keys %collapse) {
1028 my %pri = map { ($_ => 1) } $self->result_source->_pri_cols;
1029 foreach my $i (0 .. $#construct_as) {
1030 next if defined($construct_as[$i][0]); # only self table
1031 if (delete $pri{$construct_as[$i][1]}) {
1032 push(@pri_index, $i);
1034 last unless keys %pri; # short circuit (Johnny Five Is Alive!)
1038 # no need to do an if, it'll be empty if @pri_index is empty anyway
1040 my %pri_vals = map { ($_ => $copy[$_]) } @pri_index;
1044 do { # no need to check anything at the front, we always want the first row
1048 foreach my $this_as (@construct_as) {
1049 $const{$this_as->[0]||''}{$this_as->[1]} = shift(@copy);
1052 push(@const_rows, \%const);
1054 } until ( # no pri_index => no collapse => drop straight out
1057 do { # get another row, stash it, drop out if different PK
1059 @copy = $self->cursor->next;
1060 $self->{stashed_row} = \@copy;
1062 # last thing in do block, counts as true if anything doesn't match
1064 # check xor defined first for NULL vs. NOT NULL then if one is
1065 # defined the other must be so check string equality
1068 (defined $pri_vals{$_} ^ defined $copy[$_])
1069 || (defined $pri_vals{$_} && ($pri_vals{$_} ne $copy[$_]))
1074 my $alias = $self->{attrs}{alias};
1081 foreach my $const (@const_rows) {
1082 scalar @const_keys or do {
1083 @const_keys = sort { length($a) <=> length($b) } keys %$const;
1085 foreach my $key (@const_keys) {
1088 my @parts = split(/\./, $key);
1090 my $data = $const->{$key};
1091 foreach my $p (@parts) {
1092 $target = $target->[1]->{$p} ||= [];
1094 if ($cur eq ".${key}" && (my @ckey = @{$collapse{$cur}||[]})) {
1095 # collapsing at this point and on final part
1096 my $pos = $collapse_pos{$cur};
1097 CK: foreach my $ck (@ckey) {
1098 if (!defined $pos->{$ck} || $pos->{$ck} ne $data->{$ck}) {
1099 $collapse_pos{$cur} = $data;
1100 delete @collapse_pos{ # clear all positioning for sub-entries
1101 grep { m/^\Q${cur}.\E/ } keys %collapse_pos
1108 if (exists $collapse{$cur}) {
1109 $target = $target->[-1];
1112 $target->[0] = $data;
1114 $info->[0] = $const->{$key};
1122 =head2 result_source
1126 =item Arguments: $result_source?
1128 =item Return Value: $result_source
1132 An accessor for the primary ResultSource object from which this ResultSet
1139 =item Arguments: $result_class?
1141 =item Return Value: $result_class
1145 An accessor for the class to use when creating row objects. Defaults to
1146 C<< result_source->result_class >> - which in most cases is the name of the
1147 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1149 Note that changing the result_class will also remove any components
1150 that were originally loaded in the source class via
1151 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1152 in the original source class will not run.
1157 my ($self, $result_class) = @_;
1158 if ($result_class) {
1159 unless (ref $result_class) { # don't fire this for an object
1160 $self->ensure_class_loaded($result_class);
1162 $self->_result_class($result_class);
1163 # THIS LINE WOULD BE A BUG - this accessor specifically exists to
1164 # permit the user to set result class on one result set only; it only
1165 # chains if provided to search()
1166 #$self->{attrs}{result_class} = $result_class if ref $self;
1168 $self->_result_class;
1175 =item Arguments: $cond, \%attrs??
1177 =item Return Value: $count
1181 Performs an SQL C<COUNT> with the same query as the resultset was built
1182 with to find the number of elements. Passing arguments is equivalent to
1183 C<< $rs->search ($cond, \%attrs)->count >>
1189 return $self->search(@_)->count if @_ and defined $_[0];
1190 return scalar @{ $self->get_cache } if $self->get_cache;
1192 my $attrs = $self->_resolved_attrs_copy;
1194 # this is a little optimization - it is faster to do the limit
1195 # adjustments in software, instead of a subquery
1196 my $rows = delete $attrs->{rows};
1197 my $offset = delete $attrs->{offset};
1200 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1201 $crs = $self->_count_subq_rs ($attrs);
1204 $crs = $self->_count_rs ($attrs);
1206 my $count = $crs->next;
1208 $count -= $offset if $offset;
1209 $count = $rows if $rows and $rows < $count;
1210 $count = 0 if ($count < 0);
1219 =item Arguments: $cond, \%attrs??
1221 =item Return Value: $count_rs
1225 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1226 This can be very handy for subqueries:
1228 ->search( { amount => $some_rs->count_rs->as_query } )
1230 As with regular resultsets the SQL query will be executed only after
1231 the resultset is accessed via L</next> or L</all>. That would return
1232 the same single value obtainable via L</count>.
1238 return $self->search(@_)->count_rs if @_;
1240 # this may look like a lack of abstraction (count() does about the same)
1241 # but in fact an _rs *must* use a subquery for the limits, as the
1242 # software based limiting can not be ported if this $rs is to be used
1243 # in a subquery itself (i.e. ->as_query)
1244 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1245 return $self->_count_subq_rs;
1248 return $self->_count_rs;
1253 # returns a ResultSetColumn object tied to the count query
1256 my ($self, $attrs) = @_;
1258 my $rsrc = $self->result_source;
1259 $attrs ||= $self->_resolved_attrs;
1261 my $tmp_attrs = { %$attrs };
1262 # take off any limits, record_filter is cdbi, and no point of ordering nor locking a count
1263 delete @{$tmp_attrs}{qw/rows offset order_by record_filter for/};
1265 # overwrite the selector (supplied by the storage)
1266 $tmp_attrs->{select} = $rsrc->storage->_count_select ($rsrc, $attrs);
1267 $tmp_attrs->{as} = 'count';
1269 my $tmp_rs = $rsrc->resultset_class->new($rsrc, $tmp_attrs)->get_column ('count');
1275 # same as above but uses a subquery
1277 sub _count_subq_rs {
1278 my ($self, $attrs) = @_;
1280 my $rsrc = $self->result_source;
1281 $attrs ||= $self->_resolved_attrs;
1283 my $sub_attrs = { %$attrs };
1284 # extra selectors do not go in the subquery and there is no point of ordering it, nor locking it
1285 delete @{$sub_attrs}{qw/collapse select _prefetch_select as order_by for/};
1287 # if we multi-prefetch we group_by primary keys only as this is what we would
1288 # get out of the rs via ->next/->all. We *DO WANT* to clobber old group_by regardless
1289 if ( keys %{$attrs->{collapse}} ) {
1290 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } ($rsrc->_pri_cols) ]
1293 # Calculate subquery selector
1294 if (my $g = $sub_attrs->{group_by}) {
1296 my $sql_maker = $rsrc->storage->sql_maker;
1298 # necessary as the group_by may refer to aliased functions
1300 for my $sel (@{$attrs->{select}}) {
1301 $sel_index->{$sel->{-as}} = $sel
1302 if (ref $sel eq 'HASH' and $sel->{-as});
1305 for my $g_part (@$g) {
1306 my $colpiece = $sel_index->{$g_part} || $g_part;
1308 # disqualify join-based group_by's. Arcane but possible query
1309 # also horrible horrible hack to alias a column (not a func.)
1310 # (probably need to introduce SQLA syntax)
1311 if ($colpiece =~ /\./ && $colpiece !~ /^$attrs->{alias}\./) {
1314 $colpiece = \ sprintf ('%s AS %s', map { $sql_maker->_quote ($_) } ($colpiece, $as) );
1316 push @{$sub_attrs->{select}}, $colpiece;
1320 my @pcols = map { "$attrs->{alias}.$_" } ($rsrc->primary_columns);
1321 $sub_attrs->{select} = @pcols ? \@pcols : [ 1 ];
1324 return $rsrc->resultset_class
1325 ->new ($rsrc, $sub_attrs)
1327 ->search ({}, { columns => { count => $rsrc->storage->_count_select ($rsrc, $attrs) } })
1328 ->get_column ('count');
1335 =head2 count_literal
1339 =item Arguments: $sql_fragment, @bind_values
1341 =item Return Value: $count
1345 Counts the results in a literal query. Equivalent to calling L</search_literal>
1346 with the passed arguments, then L</count>.
1350 sub count_literal { shift->search_literal(@_)->count; }
1356 =item Arguments: none
1358 =item Return Value: @objects
1362 Returns all elements in the resultset. Called implicitly if the resultset
1363 is returned in list context.
1370 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1373 return @{ $self->get_cache } if $self->get_cache;
1377 if (keys %{$self->_resolved_attrs->{collapse}}) {
1378 # Using $self->cursor->all is really just an optimisation.
1379 # If we're collapsing has_many prefetches it probably makes
1380 # very little difference, and this is cleaner than hacking
1381 # _construct_object to survive the approach
1382 $self->cursor->reset;
1383 my @row = $self->cursor->next;
1385 push(@obj, $self->_construct_object(@row));
1386 @row = (exists $self->{stashed_row}
1387 ? @{delete $self->{stashed_row}}
1388 : $self->cursor->next);
1391 @obj = map { $self->_construct_object(@$_) } $self->cursor->all;
1394 $self->set_cache(\@obj) if $self->{attrs}{cache};
1403 =item Arguments: none
1405 =item Return Value: $self
1409 Resets the resultset's cursor, so you can iterate through the elements again.
1410 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1417 delete $self->{_attrs} if exists $self->{_attrs};
1418 $self->{all_cache_position} = 0;
1419 $self->cursor->reset;
1427 =item Arguments: none
1429 =item Return Value: $object | undef
1433 Resets the resultset and returns an object for the first result (or C<undef>
1434 if the resultset is empty).
1439 return $_[0]->reset->next;
1445 # Determines whether and what type of subquery is required for the $rs operation.
1446 # If grouping is necessary either supplies its own, or verifies the current one
1447 # After all is done delegates to the proper storage method.
1449 sub _rs_update_delete {
1450 my ($self, $op, $values) = @_;
1452 my $rsrc = $self->result_source;
1454 # if a condition exists we need to strip all table qualifiers
1455 # if this is not possible we'll force a subquery below
1456 my $cond = $rsrc->schema->storage->_strip_cond_qualifiers ($self->{cond});
1458 my $needs_group_by_subq = $self->_has_resolved_attr (qw/collapse group_by -join/);
1459 my $needs_subq = $needs_group_by_subq || (not defined $cond) || $self->_has_resolved_attr(qw/rows offset/);
1461 if ($needs_group_by_subq or $needs_subq) {
1463 # make a new $rs selecting only the PKs (that's all we really need)
1464 my $attrs = $self->_resolved_attrs_copy;
1467 delete $attrs->{$_} for qw/collapse _collapse_order_by select _prefetch_select as/;
1468 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } ($self->result_source->_pri_cols) ];
1470 if ($needs_group_by_subq) {
1471 # make sure no group_by was supplied, or if there is one - make sure it matches
1472 # the columns compiled above perfectly. Anything else can not be sanely executed
1473 # on most databases so croak right then and there
1475 if (my $g = $attrs->{group_by}) {
1476 my @current_group_by = map
1477 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1482 join ("\x00", sort @current_group_by)
1484 join ("\x00", sort @{$attrs->{columns}} )
1486 $self->throw_exception (
1487 "You have just attempted a $op operation on a resultset which does group_by"
1488 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1489 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1490 . ' kind of queries. Please retry the operation with a modified group_by or'
1491 . ' without using one at all.'
1496 $attrs->{group_by} = $attrs->{columns};
1500 my $subrs = (ref $self)->new($rsrc, $attrs);
1501 return $self->result_source->storage->_subq_update_delete($subrs, $op, $values);
1504 return $rsrc->storage->$op(
1506 $op eq 'update' ? $values : (),
1516 =item Arguments: \%values
1518 =item Return Value: $storage_rv
1522 Sets the specified columns in the resultset to the supplied values in a
1523 single query. Note that this will not run any accessor/set_column/update
1524 triggers, nor will it update any row object instances derived from this
1525 resultset (this includes the contents of the L<resultset cache|/set_cache>
1526 if any). See L</update_all> if you need to execute any on-update
1527 triggers or cascades defined either by you or a
1528 L<result component|DBIx::Class::Manual::Component/WHAT_IS_A_COMPONENT>.
1530 The return value is a pass through of what the underlying
1531 storage backend returned, and may vary. See L<DBI/execute> for the most
1536 Note that L</update> does not process/deflate any of the values passed in.
1537 This is unlike the corresponding L<DBIx::Class::Row/update>. The user must
1538 ensure manually that any value passed to this method will stringify to
1539 something the RDBMS knows how to deal with. A notable example is the
1540 handling of L<DateTime> objects, for more info see:
1541 L<DBIx::Class::Manual::Cookbook/Formatting_DateTime_objects_in_queries>.
1546 my ($self, $values) = @_;
1547 $self->throw_exception('Values for update must be a hash')
1548 unless ref $values eq 'HASH';
1550 return $self->_rs_update_delete ('update', $values);
1557 =item Arguments: \%values
1559 =item Return Value: 1
1563 Fetches all objects and updates them one at a time via
1564 L<DBIx::Class::Row/update>. Note that C<update_all> will run DBIC defined
1565 triggers, while L</update> will not.
1570 my ($self, $values) = @_;
1571 $self->throw_exception('Values for update_all must be a hash')
1572 unless ref $values eq 'HASH';
1574 my $guard = $self->result_source->schema->txn_scope_guard;
1575 $_->update($values) for $self->all;
1584 =item Arguments: none
1586 =item Return Value: $storage_rv
1590 Deletes the rows matching this resultset in a single query. Note that this
1591 will not run any delete triggers, nor will it alter the
1592 L<in_storage|DBIx::Class::Row/in_storage> status of any row object instances
1593 derived from this resultset (this includes the contents of the
1594 L<resultset cache|/set_cache> if any). See L</delete_all> if you need to
1595 execute any on-delete triggers or cascades defined either by you or a
1596 L<result component|DBIx::Class::Manual::Component/WHAT_IS_A_COMPONENT>.
1598 The return value is a pass through of what the underlying storage backend
1599 returned, and may vary. See L<DBI/execute> for the most common case.
1605 $self->throw_exception('delete does not accept any arguments')
1608 return $self->_rs_update_delete ('delete');
1615 =item Arguments: none
1617 =item Return Value: 1
1621 Fetches all objects and deletes them one at a time via
1622 L<DBIx::Class::Row/delete>. Note that C<delete_all> will run DBIC defined
1623 triggers, while L</delete> will not.
1629 $self->throw_exception('delete_all does not accept any arguments')
1632 my $guard = $self->result_source->schema->txn_scope_guard;
1633 $_->delete for $self->all;
1642 =item Arguments: \@data;
1646 Accepts either an arrayref of hashrefs or alternatively an arrayref of arrayrefs.
1647 For the arrayref of hashrefs style each hashref should be a structure suitable
1648 forsubmitting to a $resultset->create(...) method.
1650 In void context, C<insert_bulk> in L<DBIx::Class::Storage::DBI> is used
1651 to insert the data, as this is a faster method.
1653 Otherwise, each set of data is inserted into the database using
1654 L<DBIx::Class::ResultSet/create>, and the resulting objects are
1655 accumulated into an array. The array itself, or an array reference
1656 is returned depending on scalar or list context.
1658 Example: Assuming an Artist Class that has many CDs Classes relating:
1660 my $Artist_rs = $schema->resultset("Artist");
1662 ## Void Context Example
1663 $Artist_rs->populate([
1664 { artistid => 4, name => 'Manufactured Crap', cds => [
1665 { title => 'My First CD', year => 2006 },
1666 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
1669 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
1670 { title => 'My parents sold me to a record company', year => 2005 },
1671 { title => 'Why Am I So Ugly?', year => 2006 },
1672 { title => 'I Got Surgery and am now Popular', year => 2007 }
1677 ## Array Context Example
1678 my ($ArtistOne, $ArtistTwo, $ArtistThree) = $Artist_rs->populate([
1679 { name => "Artist One"},
1680 { name => "Artist Two"},
1681 { name => "Artist Three", cds=> [
1682 { title => "First CD", year => 2007},
1683 { title => "Second CD", year => 2008},
1687 print $ArtistOne->name; ## response is 'Artist One'
1688 print $ArtistThree->cds->count ## reponse is '2'
1690 For the arrayref of arrayrefs style, the first element should be a list of the
1691 fieldsnames to which the remaining elements are rows being inserted. For
1694 $Arstist_rs->populate([
1695 [qw/artistid name/],
1696 [100, 'A Formally Unknown Singer'],
1697 [101, 'A singer that jumped the shark two albums ago'],
1698 [102, 'An actually cool singer'],
1701 Please note an important effect on your data when choosing between void and
1702 wantarray context. Since void context goes straight to C<insert_bulk> in
1703 L<DBIx::Class::Storage::DBI> this will skip any component that is overriding
1704 C<insert>. So if you are using something like L<DBIx-Class-UUIDColumns> to
1705 create primary keys for you, you will find that your PKs are empty. In this
1706 case you will have to use the wantarray context in order to create those
1714 # cruft placed in standalone method
1715 my $data = $self->_normalize_populate_args(@_);
1717 if(defined wantarray) {
1719 foreach my $item (@$data) {
1720 push(@created, $self->create($item));
1722 return wantarray ? @created : \@created;
1724 my $first = $data->[0];
1726 # if a column is a registered relationship, and is a non-blessed hash/array, consider
1727 # it relationship data
1728 my (@rels, @columns);
1729 for (keys %$first) {
1730 my $ref = ref $first->{$_};
1731 $self->result_source->has_relationship($_) && ($ref eq 'ARRAY' or $ref eq 'HASH')
1737 my @pks = $self->result_source->primary_columns;
1739 ## do the belongs_to relationships
1740 foreach my $index (0..$#$data) {
1742 # delegate to create() for any dataset without primary keys with specified relationships
1743 if (grep { !defined $data->[$index]->{$_} } @pks ) {
1745 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
1746 my @ret = $self->populate($data);
1752 foreach my $rel (@rels) {
1753 next unless ref $data->[$index]->{$rel} eq "HASH";
1754 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
1755 my ($reverse) = keys %{$self->result_source->reverse_relationship_info($rel)};
1756 my $related = $result->result_source->_resolve_condition(
1757 $result->result_source->relationship_info($reverse)->{cond},
1762 delete $data->[$index]->{$rel};
1763 $data->[$index] = {%{$data->[$index]}, %$related};
1765 push @columns, keys %$related if $index == 0;
1769 ## inherit the data locked in the conditions of the resultset
1770 my ($rs_data) = $self->_merge_with_rscond({});
1771 delete @{$rs_data}{@columns};
1772 my @inherit_cols = keys %$rs_data;
1773 my @inherit_data = values %$rs_data;
1775 ## do bulk insert on current row
1776 $self->result_source->storage->insert_bulk(
1777 $self->result_source,
1778 [@columns, @inherit_cols],
1779 [ map { [ @$_{@columns}, @inherit_data ] } @$data ],
1782 ## do the has_many relationships
1783 foreach my $item (@$data) {
1785 foreach my $rel (@rels) {
1786 next unless $item->{$rel} && ref $item->{$rel} eq "ARRAY";
1788 my $parent = $self->find({map { $_ => $item->{$_} } @pks})
1789 || $self->throw_exception('Cannot find the relating object.');
1791 my $child = $parent->$rel;
1793 my $related = $child->result_source->_resolve_condition(
1794 $parent->result_source->relationship_info($rel)->{cond},
1799 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
1800 my @populate = map { {%$_, %$related} } @rows_to_add;
1802 $child->populate( \@populate );
1809 # populate() argumnets went over several incarnations
1810 # What we ultimately support is AoH
1811 sub _normalize_populate_args {
1812 my ($self, $arg) = @_;
1814 if (ref $arg eq 'ARRAY') {
1815 if (ref $arg->[0] eq 'HASH') {
1818 elsif (ref $arg->[0] eq 'ARRAY') {
1820 my @colnames = @{$arg->[0]};
1821 foreach my $values (@{$arg}[1 .. $#$arg]) {
1822 push @ret, { map { $colnames[$_] => $values->[$_] } (0 .. $#colnames) };
1828 $self->throw_exception('Populate expects an arrayref of hashrefs or arrayref of arrayrefs');
1835 =item Arguments: none
1837 =item Return Value: $pager
1841 Return Value a L<Data::Page> object for the current resultset. Only makes
1842 sense for queries with a C<page> attribute.
1844 To get the full count of entries for a paged resultset, call
1845 C<total_entries> on the L<Data::Page> object.
1849 # make a wizard good for both a scalar and a hashref
1850 my $mk_lazy_count_wizard = sub {
1851 require Variable::Magic;
1853 my $stash = { total_rs => shift };
1854 my $slot = shift; # only used by the hashref magic
1856 my $magic = Variable::Magic::wizard (
1857 data => sub { $stash },
1863 # set value lazily, and dispell for good
1864 ${$_[0]} = $_[1]{total_rs}->count;
1865 Variable::Magic::dispell (${$_[0]}, $_[1]{magic_selfref});
1869 # an explicit set implies dispell as well
1870 # the unless() is to work around "fun and giggles" below
1871 Variable::Magic::dispell (${$_[0]}, $_[1]{magic_selfref})
1872 unless (caller(2))[3] eq 'DBIx::Class::ResultSet::pager';
1879 if ($_[2] eq $slot and !$_[1]{inactive}) {
1880 my $cnt = $_[1]{total_rs}->count;
1881 $_[0]->{$slot} = $cnt;
1883 # attempting to dispell in a fetch handle (works in store), seems
1884 # to invariable segfault on 5.10, 5.12, 5.13 :(
1885 # so use an inactivator instead
1886 #Variable::Magic::dispell (%{$_[0]}, $_[1]{magic_selfref});
1892 if (! $_[1]{inactive} and $_[2] eq $slot) {
1893 #Variable::Magic::dispell (%{$_[0]}, $_[1]{magic_selfref});
1895 unless (caller(2))[3] eq 'DBIx::Class::ResultSet::pager';
1902 $stash->{magic_selfref} = $magic;
1903 weaken ($stash->{magic_selfref}); # this fails on 5.8.1
1908 # the tie class for 5.8.1
1910 package # hide from pause
1911 DBIx::Class::__DBIC_LAZY_RS_COUNT__;
1912 use base qw/Tie::Hash/;
1914 sub FIRSTKEY { my $dummy = scalar keys %{$_[0]{data}}; each %{$_[0]{data}} }
1915 sub NEXTKEY { each %{$_[0]{data}} }
1916 sub EXISTS { exists $_[0]{data}{$_[1]} }
1917 sub DELETE { delete $_[0]{data}{$_[1]} }
1918 sub CLEAR { %{$_[0]{data}} = () }
1919 sub SCALAR { scalar %{$_[0]{data}} }
1922 $_[1]{data} = {%{$_[1]{selfref}}};
1923 %{$_[1]{selfref}} = ();
1924 Scalar::Util::weaken ($_[1]{selfref});
1925 return bless ($_[1], $_[0]);
1929 if ($_[1] eq $_[0]{slot}) {
1930 my $cnt = $_[0]{data}{$_[1]} = $_[0]{total_rs}->count;
1931 untie %{$_[0]{selfref}};
1932 %{$_[0]{selfref}} = %{$_[0]{data}};
1941 $_[0]{data}{$_[1]} = $_[2];
1942 if ($_[1] eq $_[0]{slot}) {
1943 untie %{$_[0]{selfref}};
1944 %{$_[0]{selfref}} = %{$_[0]{data}};
1953 return $self->{pager} if $self->{pager};
1955 if ($self->get_cache) {
1956 $self->throw_exception ('Pagers on cached resultsets are not supported');
1959 my $attrs = $self->{attrs};
1960 if (!defined $attrs->{page}) {
1961 $self->throw_exception("Can't create pager for non-paged rs");
1963 elsif ($attrs->{page} <= 0) {
1964 $self->throw_exception('Invalid page number (page-numbers are 1-based)');
1966 $attrs->{rows} ||= 10;
1968 # throw away the paging flags and re-run the count (possibly
1969 # with a subselect) to get the real total count
1970 my $count_attrs = { %$attrs };
1971 delete $count_attrs->{$_} for qw/rows offset page pager/;
1972 my $total_rs = (ref $self)->new($self->result_source, $count_attrs);
1975 ### the following may seem awkward and dirty, but it's a thought-experiment
1976 ### necessary for future development of DBIx::DS. Do *NOT* change this code
1977 ### before talking to ribasushi/mst
1979 my $pager = Data::Page->new(
1980 0, #start with an empty set
1982 $self->{attrs}{page},
1985 my $data_slot = 'total_entries';
1987 # Since we are interested in a cached value (once it's set - it's set), every
1988 # technique will detach from the magic-host once the time comes to fire the
1989 # ->count (or in the segfaulting case of >= 5.10 it will deactivate itself)
1991 if ($] < 5.008003) {
1992 # 5.8.1 throws 'Modification of a read-only value attempted' when one tries
1993 # to weakref the magic container :(
1995 tie (%$pager, 'DBIx::Class::__DBIC_LAZY_RS_COUNT__',
1996 { slot => $data_slot, total_rs => $total_rs, selfref => $pager }
1999 elsif ($] < 5.010) {
2000 # We can use magic on the hash value slot. It's interesting that the magic is
2001 # attached to the hash-slot, and does *not* stop working once I do the dummy
2002 # assignments after the cast()
2003 # tested on 5.8.3 and 5.8.9
2004 my $magic = $mk_lazy_count_wizard->($total_rs);
2005 Variable::Magic::cast ( $pager->{$data_slot}, $magic );
2007 # this is for fun and giggles
2008 $pager->{$data_slot} = -1;
2009 $pager->{$data_slot} = 0;
2011 # this does not work for scalars, but works with
2013 #my %vals = %$pager;
2018 # And the uvar magic
2019 # works on 5.10.1, 5.12.1 and 5.13.4 in its current form,
2020 # however see the wizard maker for more notes
2021 my $magic = $mk_lazy_count_wizard->($total_rs, $data_slot);
2022 Variable::Magic::cast ( %$pager, $magic );
2025 $pager->{$data_slot} = -1;
2026 $pager->{$data_slot} = 0;
2034 return $self->{pager} = $pager;
2041 =item Arguments: $page_number
2043 =item Return Value: $rs
2047 Returns a resultset for the $page_number page of the resultset on which page
2048 is called, where each page contains a number of rows equal to the 'rows'
2049 attribute set on the resultset (10 by default).
2054 my ($self, $page) = @_;
2055 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
2062 =item Arguments: \%vals
2064 =item Return Value: $rowobject
2068 Creates a new row object in the resultset's result class and returns
2069 it. The row is not inserted into the database at this point, call
2070 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
2071 will tell you whether the row object has been inserted or not.
2073 Passes the hashref of input on to L<DBIx::Class::Row/new>.
2078 my ($self, $values) = @_;
2079 $self->throw_exception( "new_result needs a hash" )
2080 unless (ref $values eq 'HASH');
2082 my ($merged_cond, $cols_from_relations) = $self->_merge_with_rscond($values);
2086 @$cols_from_relations
2087 ? (-cols_from_relations => $cols_from_relations)
2089 -source_handle => $self->_source_handle,
2090 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2093 return $self->result_class->new(\%new);
2096 # _merge_with_rscond
2098 # Takes a simple hash of K/V data and returns its copy merged with the
2099 # condition already present on the resultset. Additionally returns an
2100 # arrayref of value/condition names, which were inferred from related
2101 # objects (this is needed for in-memory related objects)
2102 sub _merge_with_rscond {
2103 my ($self, $data) = @_;
2105 my (%new_data, @cols_from_relations);
2107 my $alias = $self->{attrs}{alias};
2109 if (! defined $self->{cond}) {
2110 # just massage $data below
2112 elsif ($self->{cond} eq $DBIx::Class::ResultSource::UNRESOLVABLE_CONDITION) {
2113 %new_data = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
2114 @cols_from_relations = keys %new_data;
2116 elsif (ref $self->{cond} ne 'HASH') {
2117 $self->throw_exception(
2118 "Can't abstract implicit construct, resultset condition not a hash"
2122 # precendence must be given to passed values over values inherited from
2123 # the cond, so the order here is important.
2124 my $collapsed_cond = $self->_collapse_cond($self->{cond});
2125 my %implied = %{$self->_remove_alias($collapsed_cond, $alias)};
2127 while ( my($col, $value) = each %implied ) {
2128 my $vref = ref $value;
2129 if ($vref eq 'HASH' && keys(%$value) && (keys %$value)[0] eq '=') {
2130 $new_data{$col} = $value->{'='};
2132 elsif( !$vref or $vref eq 'SCALAR' or blessed($value) ) {
2133 $new_data{$col} = $value;
2140 %{ $self->_remove_alias($data, $alias) },
2143 return (\%new_data, \@cols_from_relations);
2146 # _has_resolved_attr
2148 # determines if the resultset defines at least one
2149 # of the attributes supplied
2151 # used to determine if a subquery is neccessary
2153 # supports some virtual attributes:
2155 # This will scan for any joins being present on the resultset.
2156 # It is not a mere key-search but a deep inspection of {from}
2159 sub _has_resolved_attr {
2160 my ($self, @attr_names) = @_;
2162 my $attrs = $self->_resolved_attrs;
2166 for my $n (@attr_names) {
2167 if (grep { $n eq $_ } (qw/-join/) ) {
2168 $extra_checks{$n}++;
2172 my $attr = $attrs->{$n};
2174 next if not defined $attr;
2176 if (ref $attr eq 'HASH') {
2177 return 1 if keys %$attr;
2179 elsif (ref $attr eq 'ARRAY') {
2187 # a resolved join is expressed as a multi-level from
2189 $extra_checks{-join}
2191 ref $attrs->{from} eq 'ARRAY'
2193 @{$attrs->{from}} > 1
2201 # Recursively collapse the condition.
2203 sub _collapse_cond {
2204 my ($self, $cond, $collapsed) = @_;
2208 if (ref $cond eq 'ARRAY') {
2209 foreach my $subcond (@$cond) {
2210 next unless ref $subcond; # -or
2211 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2214 elsif (ref $cond eq 'HASH') {
2215 if (keys %$cond and (keys %$cond)[0] eq '-and') {
2216 foreach my $subcond (@{$cond->{-and}}) {
2217 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2221 foreach my $col (keys %$cond) {
2222 my $value = $cond->{$col};
2223 $collapsed->{$col} = $value;
2233 # Remove the specified alias from the specified query hash. A copy is made so
2234 # the original query is not modified.
2237 my ($self, $query, $alias) = @_;
2239 my %orig = %{ $query || {} };
2242 foreach my $key (keys %orig) {
2244 $unaliased{$key} = $orig{$key};
2247 $unaliased{$1} = $orig{$key}
2248 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2258 =item Arguments: none
2260 =item Return Value: \[ $sql, @bind ]
2264 Returns the SQL query and bind vars associated with the invocant.
2266 This is generally used as the RHS for a subquery.
2273 my $attrs = $self->_resolved_attrs_copy;
2278 # my ($sql, \@bind, \%dbi_bind_attrs) = _select_args_to_query (...)
2279 # $sql also has no wrapping parenthesis in list ctx
2281 my $sqlbind = $self->result_source->storage
2282 ->_select_args_to_query ($attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs);
2291 =item Arguments: \%vals, \%attrs?
2293 =item Return Value: $rowobject
2297 my $artist = $schema->resultset('Artist')->find_or_new(
2298 { artist => 'fred' }, { key => 'artists' });
2300 $cd->cd_to_producer->find_or_new({ producer => $producer },
2301 { key => 'primary });
2303 Find an existing record from this resultset using L</find>. if none exists,
2304 instantiate a new result object and return it. The object will not be saved
2305 into your storage until you call L<DBIx::Class::Row/insert> on it.
2307 You most likely want this method when looking for existing rows using a unique
2308 constraint that is not the primary key, or looking for related rows.
2310 If you want objects to be saved immediately, use L</find_or_create> instead.
2312 B<Note>: Make sure to read the documentation of L</find> and understand the
2313 significance of the C<key> attribute, as its lack may skew your search, and
2314 subsequently result in spurious new objects.
2316 B<Note>: Take care when using C<find_or_new> with a table having
2317 columns with default values that you intend to be automatically
2318 supplied by the database (e.g. an auto_increment primary key column).
2319 In normal usage, the value of such columns should NOT be included at
2320 all in the call to C<find_or_new>, even when set to C<undef>.
2326 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2327 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2328 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2331 return $self->new_result($hash);
2338 =item Arguments: \%vals
2340 =item Return Value: a L<DBIx::Class::Row> $object
2344 Attempt to create a single new row or a row with multiple related rows
2345 in the table represented by the resultset (and related tables). This
2346 will not check for duplicate rows before inserting, use
2347 L</find_or_create> to do that.
2349 To create one row for this resultset, pass a hashref of key/value
2350 pairs representing the columns of the table and the values you wish to
2351 store. If the appropriate relationships are set up, foreign key fields
2352 can also be passed an object representing the foreign row, and the
2353 value will be set to its primary key.
2355 To create related objects, pass a hashref of related-object column values
2356 B<keyed on the relationship name>. If the relationship is of type C<multi>
2357 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2358 The process will correctly identify columns holding foreign keys, and will
2359 transparently populate them from the keys of the corresponding relation.
2360 This can be applied recursively, and will work correctly for a structure
2361 with an arbitrary depth and width, as long as the relationships actually
2362 exists and the correct column data has been supplied.
2365 Instead of hashrefs of plain related data (key/value pairs), you may
2366 also pass new or inserted objects. New objects (not inserted yet, see
2367 L</new>), will be inserted into their appropriate tables.
2369 Effectively a shortcut for C<< ->new_result(\%vals)->insert >>.
2371 Example of creating a new row.
2373 $person_rs->create({
2374 name=>"Some Person",
2375 email=>"somebody@someplace.com"
2378 Example of creating a new row and also creating rows in a related C<has_many>
2379 or C<has_one> resultset. Note Arrayref.
2382 { artistid => 4, name => 'Manufactured Crap', cds => [
2383 { title => 'My First CD', year => 2006 },
2384 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2389 Example of creating a new row and also creating a row in a related
2390 C<belongs_to> resultset. Note Hashref.
2393 title=>"Music for Silly Walks",
2396 name=>"Silly Musician",
2404 When subclassing ResultSet never attempt to override this method. Since
2405 it is a simple shortcut for C<< $self->new_result($attrs)->insert >>, a
2406 lot of the internals simply never call it, so your override will be
2407 bypassed more often than not. Override either L<new|DBIx::Class::Row/new>
2408 or L<insert|DBIx::Class::Row/insert> depending on how early in the
2409 L</create> process you need to intervene.
2416 my ($self, $attrs) = @_;
2417 $self->throw_exception( "create needs a hashref" )
2418 unless ref $attrs eq 'HASH';
2419 return $self->new_result($attrs)->insert;
2422 =head2 find_or_create
2426 =item Arguments: \%vals, \%attrs?
2428 =item Return Value: $rowobject
2432 $cd->cd_to_producer->find_or_create({ producer => $producer },
2433 { key => 'primary' });
2435 Tries to find a record based on its primary key or unique constraints; if none
2436 is found, creates one and returns that instead.
2438 my $cd = $schema->resultset('CD')->find_or_create({
2440 artist => 'Massive Attack',
2441 title => 'Mezzanine',
2445 Also takes an optional C<key> attribute, to search by a specific key or unique
2446 constraint. For example:
2448 my $cd = $schema->resultset('CD')->find_or_create(
2450 artist => 'Massive Attack',
2451 title => 'Mezzanine',
2453 { key => 'cd_artist_title' }
2456 B<Note>: Make sure to read the documentation of L</find> and understand the
2457 significance of the C<key> attribute, as its lack may skew your search, and
2458 subsequently result in spurious row creation.
2460 B<Note>: Because find_or_create() reads from the database and then
2461 possibly inserts based on the result, this method is subject to a race
2462 condition. Another process could create a record in the table after
2463 the find has completed and before the create has started. To avoid
2464 this problem, use find_or_create() inside a transaction.
2466 B<Note>: Take care when using C<find_or_create> with a table having
2467 columns with default values that you intend to be automatically
2468 supplied by the database (e.g. an auto_increment primary key column).
2469 In normal usage, the value of such columns should NOT be included at
2470 all in the call to C<find_or_create>, even when set to C<undef>.
2472 See also L</find> and L</update_or_create>. For information on how to declare
2473 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2477 sub find_or_create {
2479 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2480 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2481 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2484 return $self->create($hash);
2487 =head2 update_or_create
2491 =item Arguments: \%col_values, { key => $unique_constraint }?
2493 =item Return Value: $row_object
2497 $resultset->update_or_create({ col => $val, ... });
2499 Like L</find_or_create>, but if a row is found it is immediately updated via
2500 C<< $found_row->update (\%col_values) >>.
2503 Takes an optional C<key> attribute to search on a specific unique constraint.
2506 # In your application
2507 my $cd = $schema->resultset('CD')->update_or_create(
2509 artist => 'Massive Attack',
2510 title => 'Mezzanine',
2513 { key => 'cd_artist_title' }
2516 $cd->cd_to_producer->update_or_create({
2517 producer => $producer,
2523 B<Note>: Make sure to read the documentation of L</find> and understand the
2524 significance of the C<key> attribute, as its lack may skew your search, and
2525 subsequently result in spurious row creation.
2527 B<Note>: Take care when using C<update_or_create> with a table having
2528 columns with default values that you intend to be automatically
2529 supplied by the database (e.g. an auto_increment primary key column).
2530 In normal usage, the value of such columns should NOT be included at
2531 all in the call to C<update_or_create>, even when set to C<undef>.
2533 See also L</find> and L</find_or_create>. For information on how to declare
2534 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2538 sub update_or_create {
2540 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2541 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2543 my $row = $self->find($cond, $attrs);
2545 $row->update($cond);
2549 return $self->create($cond);
2552 =head2 update_or_new
2556 =item Arguments: \%col_values, { key => $unique_constraint }?
2558 =item Return Value: $rowobject
2562 $resultset->update_or_new({ col => $val, ... });
2564 Like L</find_or_new> but if a row is found it is immediately updated via
2565 C<< $found_row->update (\%col_values) >>.
2569 # In your application
2570 my $cd = $schema->resultset('CD')->update_or_new(
2572 artist => 'Massive Attack',
2573 title => 'Mezzanine',
2576 { key => 'cd_artist_title' }
2579 if ($cd->in_storage) {
2580 # the cd was updated
2583 # the cd is not yet in the database, let's insert it
2587 B<Note>: Make sure to read the documentation of L</find> and understand the
2588 significance of the C<key> attribute, as its lack may skew your search, and
2589 subsequently result in spurious new objects.
2591 B<Note>: Take care when using C<update_or_new> with a table having
2592 columns with default values that you intend to be automatically
2593 supplied by the database (e.g. an auto_increment primary key column).
2594 In normal usage, the value of such columns should NOT be included at
2595 all in the call to C<update_or_new>, even when set to C<undef>.
2597 See also L</find>, L</find_or_create> and L</find_or_new>.
2603 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
2604 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2606 my $row = $self->find( $cond, $attrs );
2607 if ( defined $row ) {
2608 $row->update($cond);
2612 return $self->new_result($cond);
2619 =item Arguments: none
2621 =item Return Value: \@cache_objects | undef
2625 Gets the contents of the cache for the resultset, if the cache is set.
2627 The cache is populated either by using the L</prefetch> attribute to
2628 L</search> or by calling L</set_cache>.
2640 =item Arguments: \@cache_objects
2642 =item Return Value: \@cache_objects
2646 Sets the contents of the cache for the resultset. Expects an arrayref
2647 of objects of the same class as those produced by the resultset. Note that
2648 if the cache is set the resultset will return the cached objects rather
2649 than re-querying the database even if the cache attr is not set.
2651 The contents of the cache can also be populated by using the
2652 L</prefetch> attribute to L</search>.
2657 my ( $self, $data ) = @_;
2658 $self->throw_exception("set_cache requires an arrayref")
2659 if defined($data) && (ref $data ne 'ARRAY');
2660 $self->{all_cache} = $data;
2667 =item Arguments: none
2669 =item Return Value: undef
2673 Clears the cache for the resultset.
2678 shift->set_cache(undef);
2685 =item Arguments: none
2687 =item Return Value: true, if the resultset has been paginated
2695 return !!$self->{attrs}{page};
2702 =item Arguments: none
2704 =item Return Value: true, if the resultset has been ordered with C<order_by>.
2712 return scalar $self->result_source->storage->_extract_order_columns($self->{attrs}{order_by});
2715 =head2 related_resultset
2719 =item Arguments: $relationship_name
2721 =item Return Value: $resultset
2725 Returns a related resultset for the supplied relationship name.
2727 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
2731 sub related_resultset {
2732 my ($self, $rel) = @_;
2734 $self->{related_resultsets} ||= {};
2735 return $self->{related_resultsets}{$rel} ||= do {
2736 my $rsrc = $self->result_source;
2737 my $rel_info = $rsrc->relationship_info($rel);
2739 $self->throw_exception(
2740 "search_related: result source '" . $rsrc->source_name .
2741 "' has no such relationship $rel")
2744 my $attrs = $self->_chain_relationship($rel);
2746 my $join_count = $attrs->{seen_join}{$rel};
2748 my $alias = $self->result_source->storage
2749 ->relname_to_table_alias($rel, $join_count);
2751 # since this is search_related, and we already slid the select window inwards
2752 # (the select/as attrs were deleted in the beginning), we need to flip all
2753 # left joins to inner, so we get the expected results
2754 # read the comment on top of the actual function to see what this does
2755 $attrs->{from} = $rsrc->schema->storage->_inner_join_to_node ($attrs->{from}, $alias);
2758 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
2759 delete @{$attrs}{qw(result_class alias)};
2763 if (my $cache = $self->get_cache) {
2764 if ($cache->[0] && $cache->[0]->related_resultset($rel)->get_cache) {
2765 $new_cache = [ map { @{$_->related_resultset($rel)->get_cache} }
2770 my $rel_source = $rsrc->related_source($rel);
2774 # The reason we do this now instead of passing the alias to the
2775 # search_rs below is that if you wrap/overload resultset on the
2776 # source you need to know what alias it's -going- to have for things
2777 # to work sanely (e.g. RestrictWithObject wants to be able to add
2778 # extra query restrictions, and these may need to be $alias.)
2780 my $rel_attrs = $rel_source->resultset_attributes;
2781 local $rel_attrs->{alias} = $alias;
2783 $rel_source->resultset
2787 where => $attrs->{where},
2790 $new->set_cache($new_cache) if $new_cache;
2795 =head2 current_source_alias
2799 =item Arguments: none
2801 =item Return Value: $source_alias
2805 Returns the current table alias for the result source this resultset is built
2806 on, that will be used in the SQL query. Usually it is C<me>.
2808 Currently the source alias that refers to the result set returned by a
2809 L</search>/L</find> family method depends on how you got to the resultset: it's
2810 C<me> by default, but eg. L</search_related> aliases it to the related result
2811 source name (and keeps C<me> referring to the original result set). The long
2812 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
2813 (and make this method unnecessary).
2815 Thus it's currently necessary to use this method in predefined queries (see
2816 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
2817 source alias of the current result set:
2819 # in a result set class
2821 my ($self, $user) = @_;
2823 my $me = $self->current_source_alias;
2825 return $self->search(
2826 "$me.modified" => $user->id,
2832 sub current_source_alias {
2835 return ($self->{attrs} || {})->{alias} || 'me';
2838 =head2 as_subselect_rs
2842 =item Arguments: none
2844 =item Return Value: $resultset
2848 Act as a barrier to SQL symbols. The resultset provided will be made into a
2849 "virtual view" by including it as a subquery within the from clause. From this
2850 point on, any joined tables are inaccessible to ->search on the resultset (as if
2851 it were simply where-filtered without joins). For example:
2853 my $rs = $schema->resultset('Bar')->search({'x.name' => 'abc'},{ join => 'x' });
2855 # 'x' now pollutes the query namespace
2857 # So the following works as expected
2858 my $ok_rs = $rs->search({'x.other' => 1});
2860 # But this doesn't: instead of finding a 'Bar' related to two x rows (abc and
2861 # def) we look for one row with contradictory terms and join in another table
2862 # (aliased 'x_2') which we never use
2863 my $broken_rs = $rs->search({'x.name' => 'def'});
2865 my $rs2 = $rs->as_subselect_rs;
2867 # doesn't work - 'x' is no longer accessible in $rs2, having been sealed away
2868 my $not_joined_rs = $rs2->search({'x.other' => 1});
2870 # works as expected: finds a 'table' row related to two x rows (abc and def)
2871 my $correctly_joined_rs = $rs2->search({'x.name' => 'def'});
2873 Another example of when one might use this would be to select a subset of
2874 columns in a group by clause:
2876 my $rs = $schema->resultset('Bar')->search(undef, {
2877 group_by => [qw{ id foo_id baz_id }],
2878 })->as_subselect_rs->search(undef, {
2879 columns => [qw{ id foo_id }]
2882 In the above example normally columns would have to be equal to the group by,
2883 but because we isolated the group by into a subselect the above works.
2887 sub as_subselect_rs {
2890 my $attrs = $self->_resolved_attrs;
2892 my $fresh_rs = (ref $self)->new (
2893 $self->result_source
2896 # these pieces will be locked in the subquery
2897 delete $fresh_rs->{cond};
2898 delete @{$fresh_rs->{attrs}}{qw/where bind/};
2900 return $fresh_rs->search( {}, {
2902 $attrs->{alias} => $self->as_query,
2903 -alias => $attrs->{alias},
2904 -source_handle => $self->result_source->handle,
2906 alias => $attrs->{alias},
2910 # This code is called by search_related, and makes sure there
2911 # is clear separation between the joins before, during, and
2912 # after the relationship. This information is needed later
2913 # in order to properly resolve prefetch aliases (any alias
2914 # with a relation_chain_depth less than the depth of the
2915 # current prefetch is not considered)
2917 # The increments happen twice per join. An even number means a
2918 # relationship specified via a search_related, whereas an odd
2919 # number indicates a join/prefetch added via attributes
2921 # Also this code will wrap the current resultset (the one we
2922 # chain to) in a subselect IFF it contains limiting attributes
2923 sub _chain_relationship {
2924 my ($self, $rel) = @_;
2925 my $source = $self->result_source;
2926 my $attrs = { %{$self->{attrs}||{}} };
2928 # we need to take the prefetch the attrs into account before we
2929 # ->_resolve_join as otherwise they get lost - captainL
2930 my $join = $self->_merge_joinpref_attr( $attrs->{join}, $attrs->{prefetch} );
2932 delete @{$attrs}{qw/join prefetch collapse group_by distinct select as columns +select +as +columns/};
2934 my $seen = { %{ (delete $attrs->{seen_join}) || {} } };
2937 my @force_subq_attrs = qw/offset rows group_by having/;
2940 ($attrs->{from} && ref $attrs->{from} ne 'ARRAY')
2942 $self->_has_resolved_attr (@force_subq_attrs)
2944 # Nuke the prefetch (if any) before the new $rs attrs
2945 # are resolved (prefetch is useless - we are wrapping
2946 # a subquery anyway).
2947 my $rs_copy = $self->search;
2948 $rs_copy->{attrs}{join} = $self->_merge_joinpref_attr (
2949 $rs_copy->{attrs}{join},
2950 delete $rs_copy->{attrs}{prefetch},
2954 -source_handle => $source->handle,
2955 -alias => $attrs->{alias},
2956 $attrs->{alias} => $rs_copy->as_query,
2958 delete @{$attrs}{@force_subq_attrs, qw/where bind/};
2959 $seen->{-relation_chain_depth} = 0;
2961 elsif ($attrs->{from}) { #shallow copy suffices
2962 $from = [ @{$attrs->{from}} ];
2966 -source_handle => $source->handle,
2967 -alias => $attrs->{alias},
2968 $attrs->{alias} => $source->from,
2972 my $jpath = ($seen->{-relation_chain_depth})
2973 ? $from->[-1][0]{-join_path}
2976 my @requested_joins = $source->_resolve_join(
2983 push @$from, @requested_joins;
2985 $seen->{-relation_chain_depth}++;
2987 # if $self already had a join/prefetch specified on it, the requested
2988 # $rel might very well be already included. What we do in this case
2989 # is effectively a no-op (except that we bump up the chain_depth on
2990 # the join in question so we could tell it *is* the search_related)
2993 # we consider the last one thus reverse
2994 for my $j (reverse @requested_joins) {
2995 my ($last_j) = keys %{$j->[0]{-join_path}[-1]};
2996 if ($rel eq $last_j) {
2997 $j->[0]{-relation_chain_depth}++;
3003 unless ($already_joined) {
3004 push @$from, $source->_resolve_join(
3012 $seen->{-relation_chain_depth}++;
3014 return {%$attrs, from => $from, seen_join => $seen};
3017 # too many times we have to do $attrs = { %{$self->_resolved_attrs} }
3018 sub _resolved_attrs_copy {
3020 return { %{$self->_resolved_attrs (@_)} };
3023 sub _resolved_attrs {
3025 return $self->{_attrs} if $self->{_attrs};
3027 my $attrs = { %{ $self->{attrs} || {} } };
3028 my $source = $self->result_source;
3029 my $alias = $attrs->{alias};
3032 # resolve selectors, this one is quite hairy
3034 my $selection_pieces;
3036 $attrs->{columns} ||= delete $attrs->{cols}
3037 if exists $attrs->{cols};
3039 # disassemble columns / +columns
3041 $selection_pieces->{columns}{select},
3042 $selection_pieces->{columns}{as},
3043 $selection_pieces->{'+columns'}{select},
3044 $selection_pieces->{'+columns'}{as},
3049 for my $colbit (@$_) {
3051 if (ref $colbit eq 'HASH') {
3052 for my $as (keys %$colbit) {
3053 push @sel, $colbit->{$as};
3066 (ref $attrs->{columns} eq 'ARRAY' ? delete $attrs->{columns} : [ delete $attrs->{columns} ]),
3067 # include_columns is a legacy add-on to +columns
3068 [ map { ref $_ eq 'ARRAY' ? @$_ : ($_ || () ) } delete @{$attrs}{qw/+columns include_columns/} ] )
3071 # make copies of select/as and +select/+as
3073 $selection_pieces->{'select/as'}{select},
3074 $selection_pieces->{'select/as'}{as},
3075 $selection_pieces->{'+select/+as'}{select},
3076 $selection_pieces->{'+select/+as'}{as},
3078 { $_ ? [ ref $_ eq 'ARRAY' ? @$_ : $_ ] : [] }
3079 ( delete @{$attrs}{qw/select as +select +as/} )
3082 # default to * only when neither no non-plus selectors are available
3084 ! @{$selection_pieces->{'select/as'}{select}}
3086 ! @{$selection_pieces->{'columns'}{select}}
3088 for ($source->columns) {
3089 push @{$selection_pieces->{'select/as'}{select}}, $_;
3090 push @{$selection_pieces->{'select/as'}{as}}, $_;
3094 # final composition order (important)
3095 my @sel_pairs = grep {
3096 $selection_pieces->{$_}
3099 ( $selection_pieces->{$_}{select} && @{$selection_pieces->{$_}{select}} )
3101 ( $selection_pieces->{$_}{as} && @{$selection_pieces->{$_}{as}} )
3103 } qw|columns select/as +columns +select/+as|;
3105 # fill in missing as bits for each pair
3106 # if it's the last pair we can let things slide ( bare +select is sadly popular)
3109 for my $i (0 .. $#sel_pairs) {
3111 my $pairname = $sel_pairs[$i];
3113 my ($sel, $as) = @{$selection_pieces->{$pairname}}{qw/select as/};
3115 $self->throw_exception(
3116 "Unable to assemble final selection list: $pairname specified in addition to unbalanced $sel_pairs[$i-1]"
3117 ) if ($out_of_sync);
3119 if (@$sel == @$as) {
3122 elsif (@$sel < @$as) {
3123 $self->throw_exception(
3124 "More 'as' elements than 'select' elements for $pairname, unable to continue"
3128 # try to deduce the 'as' part, will work only if all the selectors are "plain", or contain an explicit -as
3129 # if we can not deduce something - stop right there and leave the rest of the selector un-as'ed
3130 # if there is an extra selection pair coming after that - it will die due to out_of_sync being set
3131 for my $j ($#$as+1 .. $#$sel) {
3132 if (my $ref = ref $sel->[$j]) {
3133 if ($ref eq 'HASH' and exists $sel->[$j]{-as}) {
3134 push @$as, $sel->[$j]{-as};
3142 push @$as, $sel->[$j];
3148 # assume all unqualified selectors to apply to the current alias (legacy stuff)
3149 # disqualify all $alias.col as-bits (collapser mandated)
3150 for (values %$selection_pieces) {
3151 $_->{select} = [ map { (ref $_ or $_ =~ /\./) ? $_ : "$alias.$_" } @{$_->{select}} ];
3152 $_->{as} = [ map { $_ =~ /^\Q$alias.\E(.+)$/ ? $1 : $_ } @{$_->{as}} ];
3156 # Blatant bugwardness encoded into multiple tests.
3157 # While columns behaves sensibly, +columns is expected
3158 # to dump *any* foreign columns into the main object
3160 $selection_pieces->{'+columns'}{as} = [ map
3161 { (split /\./, $_)[-1] }
3162 @{$selection_pieces->{'+columns'}{as}}
3167 $attrs->{select} = $self->_merge_attr ($attrs->{select}, $selection_pieces->{$_}{select});
3168 $attrs->{as} = $self->_merge_attr ($attrs->{as}, $selection_pieces->{$_}{as});
3171 # de-duplicate the result (remove *identical* select/as pairs)
3172 # and also die on duplicate {as} pointing to different {select}s
3173 # not using a c-style for as the condition is prone to shrinkage
3176 while ($i <= $#{$attrs->{as}} ) {
3177 my ($sel, $as) = map { $attrs->{$_}[$i] } (qw/select as/);
3179 if ($seen->{"$sel \x00\x00 $as"}++) {
3181 for @{$attrs}{qw/select as/};
3183 elsif ($seen->{$as}++) {
3184 $self->throw_exception(
3185 "inflate_result() alias '$as' specified twice with different SQL-side {select}-ors"
3193 ## selector resolution done
3197 $attrs->{from} ||= [{
3198 -source_handle => $source->handle,
3199 -alias => $self->{attrs}{alias},
3200 $self->{attrs}{alias} => $source->from,
3203 if ( $attrs->{join} || $attrs->{prefetch} ) {
3205 $self->throw_exception ('join/prefetch can not be used with a custom {from}')
3206 if ref $attrs->{from} ne 'ARRAY';
3208 my $join = delete $attrs->{join} || {};
3210 if ( defined $attrs->{prefetch} ) {
3211 $join = $self->_merge_joinpref_attr( $join, $attrs->{prefetch} );
3214 $attrs->{from} = # have to copy here to avoid corrupting the original
3216 @{ $attrs->{from} },
3217 $source->_resolve_join(
3220 { %{ $attrs->{seen_join} || {} } },
3221 ( $attrs->{seen_join} && keys %{$attrs->{seen_join}})
3222 ? $attrs->{from}[-1][0]{-join_path}
3229 if ( defined $attrs->{order_by} ) {
3230 $attrs->{order_by} = (
3231 ref( $attrs->{order_by} ) eq 'ARRAY'
3232 ? [ @{ $attrs->{order_by} } ]
3233 : [ $attrs->{order_by} || () ]
3237 if ($attrs->{group_by} and ref $attrs->{group_by} ne 'ARRAY') {
3238 $attrs->{group_by} = [ $attrs->{group_by} ];
3241 # generate the distinct induced group_by early, as prefetch will be carried via a
3242 # subquery (since a group_by is present)
3243 if (delete $attrs->{distinct}) {
3244 if ($attrs->{group_by}) {
3245 carp ("Useless use of distinct on a grouped resultset ('distinct' is ignored when a 'group_by' is present)");
3248 $attrs->{group_by} = $source->storage->_group_over_selection (
3249 @{$attrs}{qw/from select order_by/}
3254 $attrs->{collapse} ||= {};
3255 if ( my $prefetch = delete $attrs->{prefetch} ) {
3256 $prefetch = $self->_merge_joinpref_attr( {}, $prefetch );
3258 my $prefetch_ordering = [];
3260 # this is a separate structure (we don't look in {from} directly)
3261 # as the resolver needs to shift things off the lists to work
3262 # properly (identical-prefetches on different branches)
3264 if (ref $attrs->{from} eq 'ARRAY') {
3266 my $start_depth = $attrs->{seen_join}{-relation_chain_depth} || 0;
3268 for my $j ( @{$attrs->{from}}[1 .. $#{$attrs->{from}} ] ) {
3269 next unless $j->[0]{-alias};
3270 next unless $j->[0]{-join_path};
3271 next if ($j->[0]{-relation_chain_depth} || 0) < $start_depth;
3273 my @jpath = map { keys %$_ } @{$j->[0]{-join_path}};
3276 $p = $p->{$_} ||= {} for @jpath[ ($start_depth/2) .. $#jpath]; #only even depths are actual jpath boundaries
3277 push @{$p->{-join_aliases} }, $j->[0]{-alias};
3282 $source->_resolve_prefetch( $prefetch, $alias, $join_map, $prefetch_ordering, $attrs->{collapse} );
3284 # we need to somehow mark which columns came from prefetch
3285 $attrs->{_prefetch_select} = [ map { $_->[0] } @prefetch ];
3287 push @{ $attrs->{select} }, @{$attrs->{_prefetch_select}};
3288 push @{ $attrs->{as} }, (map { $_->[1] } @prefetch);
3290 push( @{$attrs->{order_by}}, @$prefetch_ordering );
3291 $attrs->{_collapse_order_by} = \@$prefetch_ordering;
3294 # if both page and offset are specified, produce a combined offset
3295 # even though it doesn't make much sense, this is what pre 081xx has
3297 if (my $page = delete $attrs->{page}) {
3299 ($attrs->{rows} * ($page - 1))
3301 ($attrs->{offset} || 0)
3305 return $self->{_attrs} = $attrs;
3309 my ($self, $attr) = @_;
3311 if (ref $attr eq 'HASH') {
3312 return $self->_rollout_hash($attr);
3313 } elsif (ref $attr eq 'ARRAY') {
3314 return $self->_rollout_array($attr);
3320 sub _rollout_array {
3321 my ($self, $attr) = @_;
3324 foreach my $element (@{$attr}) {
3325 if (ref $element eq 'HASH') {
3326 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
3327 } elsif (ref $element eq 'ARRAY') {
3328 # XXX - should probably recurse here
3329 push( @rolled_array, @{$self->_rollout_array($element)} );
3331 push( @rolled_array, $element );
3334 return \@rolled_array;
3338 my ($self, $attr) = @_;
3341 foreach my $key (keys %{$attr}) {
3342 push( @rolled_array, { $key => $attr->{$key} } );
3344 return \@rolled_array;
3347 sub _calculate_score {
3348 my ($self, $a, $b) = @_;
3350 if (defined $a xor defined $b) {
3353 elsif (not defined $a) {
3357 if (ref $b eq 'HASH') {
3358 my ($b_key) = keys %{$b};
3359 if (ref $a eq 'HASH') {
3360 my ($a_key) = keys %{$a};
3361 if ($a_key eq $b_key) {
3362 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3367 return ($a eq $b_key) ? 1 : 0;
3370 if (ref $a eq 'HASH') {
3371 my ($a_key) = keys %{$a};
3372 return ($b eq $a_key) ? 1 : 0;
3374 return ($b eq $a) ? 1 : 0;
3379 sub _merge_joinpref_attr {
3380 my ($self, $orig, $import) = @_;
3382 return $import unless defined($orig);
3383 return $orig unless defined($import);
3385 $orig = $self->_rollout_attr($orig);
3386 $import = $self->_rollout_attr($import);
3389 foreach my $import_element ( @{$import} ) {
3390 # find best candidate from $orig to merge $b_element into
3391 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3392 foreach my $orig_element ( @{$orig} ) {
3393 my $score = $self->_calculate_score( $orig_element, $import_element );
3394 if ($score > $best_candidate->{score}) {
3395 $best_candidate->{position} = $position;
3396 $best_candidate->{score} = $score;
3400 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3402 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3403 push( @{$orig}, $import_element );
3405 my $orig_best = $orig->[$best_candidate->{position}];
3406 # merge orig_best and b_element together and replace original with merged
3407 if (ref $orig_best ne 'HASH') {
3408 $orig->[$best_candidate->{position}] = $import_element;
3409 } elsif (ref $import_element eq 'HASH') {
3410 my ($key) = keys %{$orig_best};
3411 $orig->[$best_candidate->{position}] = { $key => $self->_merge_joinpref_attr($orig_best->{$key}, $import_element->{$key}) };
3414 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3425 my $hm = Hash::Merge->new;
3427 $hm->specify_behavior({
3430 my ($defl, $defr) = map { defined $_ } (@_[0,1]);
3432 if ($defl xor $defr) {
3433 return $defl ? $_[0] : $_[1];
3438 elsif (__HM_DEDUP and $_[0] eq $_[1]) {
3442 return [$_[0], $_[1]];
3446 return $_[1] if !defined $_[0];
3447 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3448 return [$_[0], @{$_[1]}]
3451 return $_[1] if !defined $_[0];
3452 return $_[0] if !keys %{$_[1]};
3453 return [$_[0], $_[1]]
3458 return $_[0] if !defined $_[1];
3459 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3460 return [@{$_[0]}, $_[1]]
3463 my @ret = @{$_[0]} or return $_[1];
3464 return [ @ret, @{$_[1]} ] unless __HM_DEDUP;
3465 my %idx = map { $_ => 1 } @ret;
3466 push @ret, grep { ! defined $idx{$_} } (@{$_[1]});
3470 return [ $_[1] ] if ! @{$_[0]};
3471 return $_[0] if !keys %{$_[1]};
3472 return $_[0] if __HM_DEDUP and List::Util::first { $_ eq $_[1] } @{$_[0]};
3473 return [ @{$_[0]}, $_[1] ];
3478 return $_[0] if !defined $_[1];
3479 return $_[1] if !keys %{$_[0]};
3480 return [$_[0], $_[1]]
3483 return $_[0] if !@{$_[1]};
3484 return $_[1] if !keys %{$_[0]};
3485 return $_[1] if __HM_DEDUP and List::Util::first { $_ eq $_[0] } @{$_[1]};
3486 return [ $_[0], @{$_[1]} ];
3489 return $_[0] if !keys %{$_[1]};
3490 return $_[1] if !keys %{$_[0]};
3491 return $_[0] if $_[0] eq $_[1];
3492 return [ $_[0], $_[1] ];
3495 } => 'DBIC_RS_ATTR_MERGER');
3499 return $hm->merge ($_[1], $_[2]);
3507 $self->_source_handle($_[0]->handle);
3509 $self->_source_handle->resolve;
3514 sub STORABLE_freeze {
3515 my ($self, $cloning) = @_;
3516 my $to_serialize = { %$self };
3518 # A cursor in progress can't be serialized (and would make little sense anyway)
3519 delete $to_serialize->{cursor};
3521 return nfreeze($to_serialize);
3524 # need this hook for symmetry
3526 my ($self, $cloning, $serialized) = @_;
3528 %$self = %{ thaw($serialized) };
3534 =head2 throw_exception
3536 See L<DBIx::Class::Schema/throw_exception> for details.
3540 sub throw_exception {
3543 if (ref $self && $self->_source_handle->schema) {
3544 $self->_source_handle->schema->throw_exception(@_)
3547 DBIx::Class::Exception->throw(@_);
3551 # XXX: FIXME: Attributes docs need clearing up
3555 Attributes are used to refine a ResultSet in various ways when
3556 searching for data. They can be passed to any method which takes an
3557 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3560 These are in no particular order:
3566 =item Value: ( $order_by | \@order_by | \%order_by )
3570 Which column(s) to order the results by.
3572 [The full list of suitable values is documented in
3573 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3576 If a single column name, or an arrayref of names is supplied, the
3577 argument is passed through directly to SQL. The hashref syntax allows
3578 for connection-agnostic specification of ordering direction:
3580 For descending order:
3582 order_by => { -desc => [qw/col1 col2 col3/] }
3584 For explicit ascending order:
3586 order_by => { -asc => 'col' }
3588 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3589 supported, although you are strongly encouraged to use the hashref
3590 syntax as outlined above.
3596 =item Value: \@columns
3600 Shortcut to request a particular set of columns to be retrieved. Each
3601 column spec may be a string (a table column name), or a hash (in which
3602 case the key is the C<as> value, and the value is used as the C<select>
3603 expression). Adds C<me.> onto the start of any column without a C<.> in
3604 it and sets C<select> from that, then auto-populates C<as> from
3605 C<select> as normal. (You may also use the C<cols> attribute, as in
3606 earlier versions of DBIC.)
3608 Essentially C<columns> does the same as L</select> and L</as>.
3610 columns => [ 'foo', { bar => 'baz' } ]
3614 select => [qw/foo baz/],
3621 =item Value: \@columns
3625 Indicates additional columns to be selected from storage. Works the same
3626 as L</columns> but adds columns to the selection. (You may also use the
3627 C<include_columns> attribute, as in earlier versions of DBIC). For
3630 $schema->resultset('CD')->search(undef, {
3631 '+columns' => ['artist.name'],
3635 would return all CDs and include a 'name' column to the information
3636 passed to object inflation. Note that the 'artist' is the name of the
3637 column (or relationship) accessor, and 'name' is the name of the column
3638 accessor in the related table.
3640 =head2 include_columns
3644 =item Value: \@columns
3648 Deprecated. Acts as a synonym for L</+columns> for backward compatibility.
3654 =item Value: \@select_columns
3658 Indicates which columns should be selected from the storage. You can use
3659 column names, or in the case of RDBMS back ends, function or stored procedure
3662 $rs = $schema->resultset('Employee')->search(undef, {
3665 { count => 'employeeid' },
3666 { max => { length => 'name' }, -as => 'longest_name' }
3671 SELECT name, COUNT( employeeid ), MAX( LENGTH( name ) ) AS longest_name FROM employee
3673 B<NOTE:> You will almost always need a corresponding L</as> attribute when you
3674 use L</select>, to instruct DBIx::Class how to store the result of the column.
3675 Also note that the L</as> attribute has nothing to do with the SQL-side 'AS'
3676 identifier aliasing. You can however alias a function, so you can use it in
3677 e.g. an C<ORDER BY> clause. This is done via the C<-as> B<select function
3678 attribute> supplied as shown in the example above.
3684 Indicates additional columns to be selected from storage. Works the same as
3685 L</select> but adds columns to the default selection, instead of specifying
3694 Indicates additional column names for those added via L</+select>. See L</as>.
3702 =item Value: \@inflation_names
3706 Indicates column names for object inflation. That is L</as> indicates the
3707 slot name in which the column value will be stored within the
3708 L<Row|DBIx::Class::Row> object. The value will then be accessible via this
3709 identifier by the C<get_column> method (or via the object accessor B<if one
3710 with the same name already exists>) as shown below. The L</as> attribute has
3711 B<nothing to do> with the SQL-side C<AS>. See L</select> for details.
3713 $rs = $schema->resultset('Employee')->search(undef, {
3716 { count => 'employeeid' },
3717 { max => { length => 'name' }, -as => 'longest_name' }
3726 If the object against which the search is performed already has an accessor
3727 matching a column name specified in C<as>, the value can be retrieved using
3728 the accessor as normal:
3730 my $name = $employee->name();
3732 If on the other hand an accessor does not exist in the object, you need to
3733 use C<get_column> instead:
3735 my $employee_count = $employee->get_column('employee_count');
3737 You can create your own accessors if required - see
3738 L<DBIx::Class::Manual::Cookbook> for details.
3744 =item Value: ($rel_name | \@rel_names | \%rel_names)
3748 Contains a list of relationships that should be joined for this query. For
3751 # Get CDs by Nine Inch Nails
3752 my $rs = $schema->resultset('CD')->search(
3753 { 'artist.name' => 'Nine Inch Nails' },
3754 { join => 'artist' }
3757 Can also contain a hash reference to refer to the other relation's relations.
3760 package MyApp::Schema::Track;
3761 use base qw/DBIx::Class/;
3762 __PACKAGE__->table('track');
3763 __PACKAGE__->add_columns(qw/trackid cd position title/);
3764 __PACKAGE__->set_primary_key('trackid');
3765 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
3768 # In your application
3769 my $rs = $schema->resultset('Artist')->search(
3770 { 'track.title' => 'Teardrop' },
3772 join => { cd => 'track' },
3773 order_by => 'artist.name',
3777 You need to use the relationship (not the table) name in conditions,
3778 because they are aliased as such. The current table is aliased as "me", so
3779 you need to use me.column_name in order to avoid ambiguity. For example:
3781 # Get CDs from 1984 with a 'Foo' track
3782 my $rs = $schema->resultset('CD')->search(
3785 'tracks.name' => 'Foo'
3787 { join => 'tracks' }
3790 If the same join is supplied twice, it will be aliased to <rel>_2 (and
3791 similarly for a third time). For e.g.
3793 my $rs = $schema->resultset('Artist')->search({
3794 'cds.title' => 'Down to Earth',
3795 'cds_2.title' => 'Popular',
3797 join => [ qw/cds cds/ ],
3800 will return a set of all artists that have both a cd with title 'Down
3801 to Earth' and a cd with title 'Popular'.
3803 If you want to fetch related objects from other tables as well, see C<prefetch>
3806 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
3812 =item Value: ($rel_name | \@rel_names | \%rel_names)
3816 Contains one or more relationships that should be fetched along with
3817 the main query (when they are accessed afterwards the data will
3818 already be available, without extra queries to the database). This is
3819 useful for when you know you will need the related objects, because it
3820 saves at least one query:
3822 my $rs = $schema->resultset('Tag')->search(
3831 The initial search results in SQL like the following:
3833 SELECT tag.*, cd.*, artist.* FROM tag
3834 JOIN cd ON tag.cd = cd.cdid
3835 JOIN artist ON cd.artist = artist.artistid
3837 L<DBIx::Class> has no need to go back to the database when we access the
3838 C<cd> or C<artist> relationships, which saves us two SQL statements in this
3841 Simple prefetches will be joined automatically, so there is no need
3842 for a C<join> attribute in the above search.
3844 C<prefetch> can be used with the following relationship types: C<belongs_to>,
3845 C<has_one> (or if you're using C<add_relationship>, any relationship declared
3846 with an accessor type of 'single' or 'filter'). A more complex example that
3847 prefetches an artists cds, the tracks on those cds, and the tags associated
3848 with that artist is given below (assuming many-to-many from artists to tags):
3850 my $rs = $schema->resultset('Artist')->search(
3854 { cds => 'tracks' },
3855 { artist_tags => 'tags' }
3861 B<NOTE:> If you specify a C<prefetch> attribute, the C<join> and C<select>
3862 attributes will be ignored.
3864 B<CAVEATs>: Prefetch does a lot of deep magic. As such, it may not behave
3865 exactly as you might expect.
3871 Prefetch uses the L</cache> to populate the prefetched relationships. This
3872 may or may not be what you want.
3876 If you specify a condition on a prefetched relationship, ONLY those
3877 rows that match the prefetched condition will be fetched into that relationship.
3878 This means that adding prefetch to a search() B<may alter> what is returned by
3879 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
3881 my $artist_rs = $schema->resultset('Artist')->search({
3887 my $count = $artist_rs->first->cds->count;
3889 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
3891 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
3893 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
3895 that cmp_ok() may or may not pass depending on the datasets involved. This
3896 behavior may or may not survive the 0.09 transition.
3908 Makes the resultset paged and specifies the page to retrieve. Effectively
3909 identical to creating a non-pages resultset and then calling ->page($page)
3912 If L<rows> attribute is not specified it defaults to 10 rows per page.
3914 When you have a paged resultset, L</count> will only return the number
3915 of rows in the page. To get the total, use the L</pager> and call
3916 C<total_entries> on it.
3926 Specifies the maximum number of rows for direct retrieval or the number of
3927 rows per page if the page attribute or method is used.
3933 =item Value: $offset
3937 Specifies the (zero-based) row number for the first row to be returned, or the
3938 of the first row of the first page if paging is used.
3944 =item Value: \@columns
3948 A arrayref of columns to group by. Can include columns of joined tables.
3950 group_by => [qw/ column1 column2 ... /]
3956 =item Value: $condition
3960 HAVING is a select statement attribute that is applied between GROUP BY and
3961 ORDER BY. It is applied to the after the grouping calculations have been
3964 having => { 'count(employee)' => { '>=', 100 } }
3970 =item Value: (0 | 1)
3974 Set to 1 to group by all columns. If the resultset already has a group_by
3975 attribute, this setting is ignored and an appropriate warning is issued.
3981 Adds to the WHERE clause.
3983 # only return rows WHERE deleted IS NULL for all searches
3984 __PACKAGE__->resultset_attributes({ where => { deleted => undef } }); )
3986 Can be overridden by passing C<< { where => undef } >> as an attribute
3993 Set to 1 to cache search results. This prevents extra SQL queries if you
3994 revisit rows in your ResultSet:
3996 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
3998 while( my $artist = $resultset->next ) {
4002 $rs->first; # without cache, this would issue a query
4004 By default, searches are not cached.
4006 For more examples of using these attributes, see
4007 L<DBIx::Class::Manual::Cookbook>.
4013 =item Value: ( 'update' | 'shared' )
4017 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT