1 package DBIx::Class::ResultSet;
9 use Carp::Clan qw/^DBIx::Class/;
12 use DBIx::Class::ResultSetColumn;
13 use DBIx::Class::ResultSourceHandle;
16 use base qw/DBIx::Class/;
18 __PACKAGE__->mk_group_accessors('simple' => qw/_result_class _source_handle/);
22 DBIx::Class::ResultSet - Represents a query used for fetching a set of results.
26 my $users_rs = $schema->resultset('User');
27 my $registered_users_rs = $schema->resultset('User')->search({ registered => 1 });
28 my @cds_in_2005 = $schema->resultset('CD')->search({ year => 2005 })->all();
32 A ResultSet is an object which stores a set of conditions representing
33 a query. It is the backbone of DBIx::Class (i.e. the really
34 important/useful bit).
36 No SQL is executed on the database when a ResultSet is created, it
37 just stores all the conditions needed to create the query.
39 A basic ResultSet representing the data of an entire table is returned
40 by calling C<resultset> on a L<DBIx::Class::Schema> and passing in a
41 L<Source|DBIx::Class::Manual::Glossary/Source> name.
43 my $users_rs = $schema->resultset('User');
45 A new ResultSet is returned from calling L</search> on an existing
46 ResultSet. The new one will contain all the conditions of the
47 original, plus any new conditions added in the C<search> call.
49 A ResultSet also incorporates an implicit iterator. L</next> and L</reset>
50 can be used to walk through all the L<DBIx::Class::Row>s the ResultSet
53 The query that the ResultSet represents is B<only> executed against
54 the database when these methods are called:
55 L</find> L</next> L</all> L</first> L</single> L</count>
59 =head2 Chaining resultsets
61 Let's say you've got a query that needs to be run to return some data
62 to the user. But, you have an authorization system in place that
63 prevents certain users from seeing certain information. So, you want
64 to construct the basic query in one method, but add constraints to it in
69 my $request = $self->get_request; # Get a request object somehow.
70 my $schema = $self->get_schema; # Get the DBIC schema object somehow.
72 my $cd_rs = $schema->resultset('CD')->search({
73 title => $request->param('title'),
74 year => $request->param('year'),
77 $self->apply_security_policy( $cd_rs );
82 sub apply_security_policy {
91 =head3 Resolving conditions and attributes
93 When a resultset is chained from another resultset, conditions and
94 attributes with the same keys need resolving.
96 L</join>, L</prefetch>, L</+select>, L</+as> attributes are merged
97 into the existing ones from the original resultset.
99 The L</where>, L</having> attribute, and any search conditions are
100 merged with an SQL C<AND> to the existing condition from the original
103 All other attributes are overridden by any new ones supplied in the
106 =head2 Multiple queries
108 Since a resultset just defines a query, you can do all sorts of
109 things with it with the same object.
111 # Don't hit the DB yet.
112 my $cd_rs = $schema->resultset('CD')->search({
113 title => 'something',
117 # Each of these hits the DB individually.
118 my $count = $cd_rs->count;
119 my $most_recent = $cd_rs->get_column('date_released')->max();
120 my @records = $cd_rs->all;
122 And it's not just limited to SELECT statements.
128 $cd_rs->create({ artist => 'Fred' });
130 Which is the same as:
132 $schema->resultset('CD')->create({
133 title => 'something',
138 See: L</search>, L</count>, L</get_column>, L</all>, L</create>.
142 If a resultset is used in a numeric context it returns the L</count>.
143 However, if it is used in a booleand context it is always true. So if
144 you want to check if a resultset has any results use C<if $rs != 0>.
145 C<if $rs> will always be true.
153 =item Arguments: $source, \%$attrs
155 =item Return Value: $rs
159 The resultset constructor. Takes a source object (usually a
160 L<DBIx::Class::ResultSourceProxy::Table>) and an attribute hash (see
161 L</ATTRIBUTES> below). Does not perform any queries -- these are
162 executed as needed by the other methods.
164 Generally you won't need to construct a resultset manually. You'll
165 automatically get one from e.g. a L</search> called in scalar context:
167 my $rs = $schema->resultset('CD')->search({ title => '100th Window' });
169 IMPORTANT: If called on an object, proxies to new_result instead so
171 my $cd = $schema->resultset('CD')->new({ title => 'Spoon' });
173 will return a CD object, not a ResultSet.
179 return $class->new_result(@_) if ref $class;
181 my ($source, $attrs) = @_;
182 $source = $source->handle
183 unless $source->isa('DBIx::Class::ResultSourceHandle');
184 $attrs = { %{$attrs||{}} };
186 if ($attrs->{page}) {
187 $attrs->{rows} ||= 10;
190 $attrs->{alias} ||= 'me';
192 # Creation of {} and bless separated to mitigate RH perl bug
193 # see https://bugzilla.redhat.com/show_bug.cgi?id=196836
195 _source_handle => $source,
196 cond => $attrs->{where},
205 $attrs->{result_class} || $source->resolve->result_class
215 =item Arguments: $cond, \%attrs?
217 =item Return Value: $resultset (scalar context), @row_objs (list context)
221 my @cds = $cd_rs->search({ year => 2001 }); # "... WHERE year = 2001"
222 my $new_rs = $cd_rs->search({ year => 2005 });
224 my $new_rs = $cd_rs->search([ { year => 2005 }, { year => 2004 } ]);
225 # year = 2005 OR year = 2004
227 If you need to pass in additional attributes but no additional condition,
228 call it as C<search(undef, \%attrs)>.
230 # "SELECT name, artistid FROM $artist_table"
231 my @all_artists = $schema->resultset('Artist')->search(undef, {
232 columns => [qw/name artistid/],
235 For a list of attributes that can be passed to C<search>, see
236 L</ATTRIBUTES>. For more examples of using this function, see
237 L<Searching|DBIx::Class::Manual::Cookbook/Searching>. For a complete
238 documentation for the first argument, see L<SQL::Abstract>.
240 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
246 my $rs = $self->search_rs( @_ );
247 return (wantarray ? $rs->all : $rs);
254 =item Arguments: $cond, \%attrs?
256 =item Return Value: $resultset
260 This method does the same exact thing as search() except it will
261 always return a resultset, even in list context.
268 # Special-case handling for (undef, undef).
269 if ( @_ == 2 && !defined $_[1] && !defined $_[0] ) {
274 $attrs = pop(@_) if @_ > 1 and ref $_[$#_] eq 'HASH';
275 my $our_attrs = { %{$self->{attrs}} };
276 my $having = delete $our_attrs->{having};
277 my $where = delete $our_attrs->{where};
281 my %safe = (alias => 1, cache => 1);
284 (@_ && defined($_[0])) # @_ == () or (undef)
286 (keys %$attrs # empty attrs or only 'safe' attrs
287 && List::Util::first { !$safe{$_} } keys %$attrs)
289 # no search, effectively just a clone
290 $rows = $self->get_cache;
293 my $new_attrs = { %{$our_attrs}, %{$attrs} };
295 # merge new attrs into inherited
296 foreach my $key (qw/join prefetch +select +as bind/) {
297 next unless exists $attrs->{$key};
298 $new_attrs->{$key} = $self->_merge_attr($our_attrs->{$key}, $attrs->{$key});
303 (@_ == 1 || ref $_[0] eq "HASH")
305 (ref $_[0] eq 'HASH')
307 (keys %{ $_[0] } > 0)
315 ? $self->throw_exception("Odd number of arguments to search")
322 if (defined $where) {
323 $new_attrs->{where} = (
324 defined $new_attrs->{where}
327 ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_
328 } $where, $new_attrs->{where}
335 $new_attrs->{where} = (
336 defined $new_attrs->{where}
339 ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_
340 } $cond, $new_attrs->{where}
346 if (defined $having) {
347 $new_attrs->{having} = (
348 defined $new_attrs->{having}
351 ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_
352 } $having, $new_attrs->{having}
358 my $rs = (ref $self)->new($self->result_source, $new_attrs);
360 $rs->set_cache($rows);
365 =head2 search_literal
369 =item Arguments: $sql_fragment, @bind_values
371 =item Return Value: $resultset (scalar context), @row_objs (list context)
375 my @cds = $cd_rs->search_literal('year = ? AND title = ?', qw/2001 Reload/);
376 my $newrs = $artist_rs->search_literal('name = ?', 'Metallica');
378 Pass a literal chunk of SQL to be added to the conditional part of the
381 CAVEAT: C<search_literal> is provided for Class::DBI compatibility and should
382 only be used in that context. C<search_literal> is a convenience method.
383 It is equivalent to calling $schema->search(\[]), but if you want to ensure
384 columns are bound correctly, use C<search>.
386 Example of how to use C<search> instead of C<search_literal>
388 my @cds = $cd_rs->search_literal('cdid = ? AND (artist = ? OR artist = ?)', (2, 1, 2));
389 my @cds = $cd_rs->search(\[ 'cdid = ? AND (artist = ? OR artist = ?)', [ 'cdid', 2 ], [ 'artist', 1 ], [ 'artist', 2 ] ]);
392 See L<DBIx::Class::Manual::Cookbook/Searching> and
393 L<DBIx::Class::Manual::FAQ/Searching> for searching techniques that do not
394 require C<search_literal>.
399 my ($self, $sql, @bind) = @_;
401 if ( @bind && ref($bind[-1]) eq 'HASH' ) {
404 return $self->search(\[ $sql, map [ __DUMMY__ => $_ ], @bind ], ($attr || () ));
411 =item Arguments: @values | \%cols, \%attrs?
413 =item Return Value: $row_object | undef
417 Finds a row based on its primary key or unique constraint. For example, to find
418 a row by its primary key:
420 my $cd = $schema->resultset('CD')->find(5);
422 You can also find a row by a specific unique constraint using the C<key>
423 attribute. For example:
425 my $cd = $schema->resultset('CD')->find('Massive Attack', 'Mezzanine', {
426 key => 'cd_artist_title'
429 Additionally, you can specify the columns explicitly by name:
431 my $cd = $schema->resultset('CD')->find(
433 artist => 'Massive Attack',
434 title => 'Mezzanine',
436 { key => 'cd_artist_title' }
439 If the C<key> is specified as C<primary>, it searches only on the primary key.
441 If no C<key> is specified, it searches on all unique constraints defined on the
442 source for which column data is provided, including the primary key.
444 If your table does not have a primary key, you B<must> provide a value for the
445 C<key> attribute matching one of the unique constraints on the source.
447 In addition to C<key>, L</find> recognizes and applies standard
448 L<resultset attributes|/ATTRIBUTES> in the same way as L</search> does.
450 Note: If your query does not return only one row, a warning is generated:
452 Query returned more than one row
454 See also L</find_or_create> and L</update_or_create>. For information on how to
455 declare unique constraints, see
456 L<DBIx::Class::ResultSource/add_unique_constraint>.
462 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
464 # Default to the primary key, but allow a specific key
465 my @cols = exists $attrs->{key}
466 ? $self->result_source->unique_constraint_columns($attrs->{key})
467 : $self->result_source->primary_columns;
468 $self->throw_exception(
469 "Can't find unless a primary key is defined or unique constraint is specified"
472 # Parse out a hashref from input
474 if (ref $_[0] eq 'HASH') {
475 $input_query = { %{$_[0]} };
477 elsif (@_ == @cols) {
479 @{$input_query}{@cols} = @_;
482 # Compatibility: Allow e.g. find(id => $value)
483 carp "Find by key => value deprecated; please use a hashref instead";
487 my (%related, $info);
489 KEY: foreach my $key (keys %$input_query) {
490 if (ref($input_query->{$key})
491 && ($info = $self->result_source->relationship_info($key))) {
492 my $val = delete $input_query->{$key};
493 next KEY if (ref($val) eq 'ARRAY'); # has_many for multi_create
494 my $rel_q = $self->result_source->_resolve_condition(
495 $info->{cond}, $val, $key
497 die "Can't handle OR join condition in find" if ref($rel_q) eq 'ARRAY';
498 @related{keys %$rel_q} = values %$rel_q;
501 if (my @keys = keys %related) {
502 @{$input_query}{@keys} = values %related;
506 # Build the final query: Default to the disjunction of the unique queries,
507 # but allow the input query in case the ResultSet defines the query or the
508 # user is abusing find
509 my $alias = exists $attrs->{alias} ? $attrs->{alias} : $self->{attrs}{alias};
511 if (exists $attrs->{key}) {
512 my @unique_cols = $self->result_source->unique_constraint_columns($attrs->{key});
513 my $unique_query = $self->_build_unique_query($input_query, \@unique_cols);
514 $query = $self->_add_alias($unique_query, $alias);
516 elsif ($self->{attrs}{accessor} and $self->{attrs}{accessor} eq 'single') {
517 # This means that we got here after a merger of relationship conditions
518 # in ::Relationship::Base::search_related (the row method), and furthermore
519 # the relationship is of the 'single' type. This means that the condition
520 # provided by the relationship (already attached to $self) is sufficient,
521 # as there can be only one row in the databse that would satisfy the
525 my @unique_queries = $self->_unique_queries($input_query, $attrs);
526 $query = @unique_queries
527 ? [ map { $self->_add_alias($_, $alias) } @unique_queries ]
528 : $self->_add_alias($input_query, $alias);
532 my $rs = $self->search ($query, $attrs);
533 if (keys %{$rs->_resolved_attrs->{collapse}}) {
535 carp "Query returned more than one row" if $rs->next;
545 # Add the specified alias to the specified query hash. A copy is made so the
546 # original query is not modified.
549 my ($self, $query, $alias) = @_;
551 my %aliased = %$query;
552 foreach my $col (grep { ! m/\./ } keys %aliased) {
553 $aliased{"$alias.$col"} = delete $aliased{$col};
561 # Build a list of queries which satisfy unique constraints.
563 sub _unique_queries {
564 my ($self, $query, $attrs) = @_;
566 my @constraint_names = exists $attrs->{key}
568 : $self->result_source->unique_constraint_names;
570 my $where = $self->_collapse_cond($self->{attrs}{where} || {});
571 my $num_where = scalar keys %$where;
574 foreach my $name (@constraint_names) {
575 my @unique_cols = $self->result_source->unique_constraint_columns($name);
576 my $unique_query = $self->_build_unique_query($query, \@unique_cols);
578 my $num_cols = scalar @unique_cols;
579 my $num_query = scalar keys %$unique_query;
581 my $total = $num_query + $num_where;
582 if ($num_query && ($num_query == $num_cols || $total == $num_cols)) {
583 # The query is either unique on its own or is unique in combination with
584 # the existing where clause
585 push @unique_queries, $unique_query;
589 return @unique_queries;
592 # _build_unique_query
594 # Constrain the specified query hash based on the specified column names.
596 sub _build_unique_query {
597 my ($self, $query, $unique_cols) = @_;
600 map { $_ => $query->{$_} }
601 grep { exists $query->{$_} }
606 =head2 search_related
610 =item Arguments: $rel, $cond, \%attrs?
612 =item Return Value: $new_resultset
616 $new_rs = $cd_rs->search_related('artist', {
620 Searches the specified relationship, optionally specifying a condition and
621 attributes for matching records. See L</ATTRIBUTES> for more information.
626 return shift->related_resultset(shift)->search(@_);
629 =head2 search_related_rs
631 This method works exactly the same as search_related, except that
632 it guarantees a restultset, even in list context.
636 sub search_related_rs {
637 return shift->related_resultset(shift)->search_rs(@_);
644 =item Arguments: none
646 =item Return Value: $cursor
650 Returns a storage-driven cursor to the given resultset. See
651 L<DBIx::Class::Cursor> for more information.
658 my $attrs = $self->_resolved_attrs_copy;
660 return $self->{cursor}
661 ||= $self->result_source->storage->select($attrs->{from}, $attrs->{select},
662 $attrs->{where},$attrs);
669 =item Arguments: $cond?
671 =item Return Value: $row_object?
675 my $cd = $schema->resultset('CD')->single({ year => 2001 });
677 Inflates the first result without creating a cursor if the resultset has
678 any records in it; if not returns nothing. Used by L</find> as a lean version of
681 While this method can take an optional search condition (just like L</search>)
682 being a fast-code-path it does not recognize search attributes. If you need to
683 add extra joins or similar, call L</search> and then chain-call L</single> on the
684 L<DBIx::Class::ResultSet> returned.
690 As of 0.08100, this method enforces the assumption that the preceeding
691 query returns only one row. If more than one row is returned, you will receive
694 Query returned more than one row
696 In this case, you should be using L</next> or L</find> instead, or if you really
697 know what you are doing, use the L</rows> attribute to explicitly limit the size
700 This method will also throw an exception if it is called on a resultset prefetching
701 has_many, as such a prefetch implies fetching multiple rows from the database in
702 order to assemble the resulting object.
709 my ($self, $where) = @_;
711 $self->throw_exception('single() only takes search conditions, no attributes. You want ->search( $cond, $attrs )->single()');
714 my $attrs = $self->_resolved_attrs_copy;
716 if (keys %{$attrs->{collapse}}) {
717 $self->throw_exception(
718 'single() can not be used on resultsets prefetching has_many. Use find( \%cond ) or next() instead'
723 if (defined $attrs->{where}) {
726 [ map { ref $_ eq 'ARRAY' ? [ -or => $_ ] : $_ }
727 $where, delete $attrs->{where} ]
730 $attrs->{where} = $where;
734 # XXX: Disabled since it doesn't infer uniqueness in all cases
735 # unless ($self->_is_unique_query($attrs->{where})) {
736 # carp "Query not guaranteed to return a single row"
737 # . "; please declare your unique constraints or use search instead";
740 my @data = $self->result_source->storage->select_single(
741 $attrs->{from}, $attrs->{select},
742 $attrs->{where}, $attrs
745 return (@data ? ($self->_construct_object(@data))[0] : undef);
751 # Try to determine if the specified query is guaranteed to be unique, based on
752 # the declared unique constraints.
754 sub _is_unique_query {
755 my ($self, $query) = @_;
757 my $collapsed = $self->_collapse_query($query);
758 my $alias = $self->{attrs}{alias};
760 foreach my $name ($self->result_source->unique_constraint_names) {
761 my @unique_cols = map {
763 } $self->result_source->unique_constraint_columns($name);
765 # Count the values for each unique column
766 my %seen = map { $_ => 0 } @unique_cols;
768 foreach my $key (keys %$collapsed) {
769 my $aliased = $key =~ /\./ ? $key : "$alias.$key";
770 next unless exists $seen{$aliased}; # Additional constraints are okay
771 $seen{$aliased} = scalar keys %{ $collapsed->{$key} };
774 # If we get 0 or more than 1 value for a column, it's not necessarily unique
775 return 1 unless grep { $_ != 1 } values %seen;
783 # Recursively collapse the query, accumulating values for each column.
785 sub _collapse_query {
786 my ($self, $query, $collapsed) = @_;
790 if (ref $query eq 'ARRAY') {
791 foreach my $subquery (@$query) {
792 next unless ref $subquery; # -or
793 $collapsed = $self->_collapse_query($subquery, $collapsed);
796 elsif (ref $query eq 'HASH') {
797 if (keys %$query and (keys %$query)[0] eq '-and') {
798 foreach my $subquery (@{$query->{-and}}) {
799 $collapsed = $self->_collapse_query($subquery, $collapsed);
803 foreach my $col (keys %$query) {
804 my $value = $query->{$col};
805 $collapsed->{$col}{$value}++;
817 =item Arguments: $cond?
819 =item Return Value: $resultsetcolumn
823 my $max_length = $rs->get_column('length')->max;
825 Returns a L<DBIx::Class::ResultSetColumn> instance for a column of the ResultSet.
830 my ($self, $column) = @_;
831 my $new = DBIx::Class::ResultSetColumn->new($self, $column);
839 =item Arguments: $cond, \%attrs?
841 =item Return Value: $resultset (scalar context), @row_objs (list context)
845 # WHERE title LIKE '%blue%'
846 $cd_rs = $rs->search_like({ title => '%blue%'});
848 Performs a search, but uses C<LIKE> instead of C<=> as the condition. Note
849 that this is simply a convenience method retained for ex Class::DBI users.
850 You most likely want to use L</search> with specific operators.
852 For more information, see L<DBIx::Class::Manual::Cookbook>.
854 This method is deprecated and will be removed in 0.09. Use L</search()>
855 instead. An example conversion is:
857 ->search_like({ foo => 'bar' });
861 ->search({ foo => { like => 'bar' } });
868 'search_like() is deprecated and will be removed in DBIC version 0.09.'
869 .' Instead use ->search({ x => { -like => "y%" } })'
870 .' (note the outer pair of {}s - they are important!)'
872 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
873 my $query = ref $_[0] eq 'HASH' ? { %{shift()} }: {@_};
874 $query->{$_} = { 'like' => $query->{$_} } for keys %$query;
875 return $class->search($query, { %$attrs });
882 =item Arguments: $first, $last
884 =item Return Value: $resultset (scalar context), @row_objs (list context)
888 Returns a resultset or object list representing a subset of elements from the
889 resultset slice is called on. Indexes are from 0, i.e., to get the first
892 my ($one, $two, $three) = $rs->slice(0, 2);
897 my ($self, $min, $max) = @_;
898 my $attrs = {}; # = { %{ $self->{attrs} || {} } };
899 $attrs->{offset} = $self->{attrs}{offset} || 0;
900 $attrs->{offset} += $min;
901 $attrs->{rows} = ($max ? ($max - $min + 1) : 1);
902 return $self->search(undef(), $attrs);
903 #my $slice = (ref $self)->new($self->result_source, $attrs);
904 #return (wantarray ? $slice->all : $slice);
911 =item Arguments: none
913 =item Return Value: $result?
917 Returns the next element in the resultset (C<undef> is there is none).
919 Can be used to efficiently iterate over records in the resultset:
921 my $rs = $schema->resultset('CD')->search;
922 while (my $cd = $rs->next) {
926 Note that you need to store the resultset object, and call C<next> on it.
927 Calling C<< resultset('Table')->next >> repeatedly will always return the
928 first record from the resultset.
934 if (my $cache = $self->get_cache) {
935 $self->{all_cache_position} ||= 0;
936 return $cache->[$self->{all_cache_position}++];
938 if ($self->{attrs}{cache}) {
939 $self->{all_cache_position} = 1;
940 return ($self->all)[0];
942 if ($self->{stashed_objects}) {
943 my $obj = shift(@{$self->{stashed_objects}});
944 delete $self->{stashed_objects} unless @{$self->{stashed_objects}};
948 exists $self->{stashed_row}
949 ? @{delete $self->{stashed_row}}
950 : $self->cursor->next
952 return undef unless (@row);
953 my ($row, @more) = $self->_construct_object(@row);
954 $self->{stashed_objects} = \@more if @more;
958 sub _construct_object {
959 my ($self, @row) = @_;
961 my $info = $self->_collapse_result($self->{_attrs}{as}, \@row)
963 my @new = $self->result_class->inflate_result($self->result_source, @$info);
964 @new = $self->{_attrs}{record_filter}->(@new)
965 if exists $self->{_attrs}{record_filter};
969 sub _collapse_result {
970 my ($self, $as_proto, $row) = @_;
972 # if the first row that ever came in is totally empty - this means we got
973 # hit by a smooth^Wempty left-joined resultset. Just noop in that case
974 # instead of producing a {}
983 return undef unless $has_def;
987 # 'foo' => [ undef, 'foo' ]
988 # 'foo.bar' => [ 'foo', 'bar' ]
989 # 'foo.bar.baz' => [ 'foo.bar', 'baz' ]
991 my @construct_as = map { [ (/^(?:(.*)\.)?([^.]+)$/) ] } @$as_proto;
993 my %collapse = %{$self->{_attrs}{collapse}||{}};
997 # if we're doing collapsing (has_many prefetch) we need to grab records
998 # until the PK changes, so fill @pri_index. if not, we leave it empty so
999 # we know we don't have to bother.
1001 # the reason for not using the collapse stuff directly is because if you
1002 # had for e.g. two artists in a row with no cds, the collapse info for
1003 # both would be NULL (undef) so you'd lose the second artist
1005 # store just the index so we can check the array positions from the row
1006 # without having to contruct the full hash
1008 if (keys %collapse) {
1009 my %pri = map { ($_ => 1) } $self->result_source->primary_columns;
1010 foreach my $i (0 .. $#construct_as) {
1011 next if defined($construct_as[$i][0]); # only self table
1012 if (delete $pri{$construct_as[$i][1]}) {
1013 push(@pri_index, $i);
1015 last unless keys %pri; # short circuit (Johnny Five Is Alive!)
1019 # no need to do an if, it'll be empty if @pri_index is empty anyway
1021 my %pri_vals = map { ($_ => $copy[$_]) } @pri_index;
1025 do { # no need to check anything at the front, we always want the first row
1029 foreach my $this_as (@construct_as) {
1030 $const{$this_as->[0]||''}{$this_as->[1]} = shift(@copy);
1033 push(@const_rows, \%const);
1035 } until ( # no pri_index => no collapse => drop straight out
1038 do { # get another row, stash it, drop out if different PK
1040 @copy = $self->cursor->next;
1041 $self->{stashed_row} = \@copy;
1043 # last thing in do block, counts as true if anything doesn't match
1045 # check xor defined first for NULL vs. NOT NULL then if one is
1046 # defined the other must be so check string equality
1049 (defined $pri_vals{$_} ^ defined $copy[$_])
1050 || (defined $pri_vals{$_} && ($pri_vals{$_} ne $copy[$_]))
1055 my $alias = $self->{attrs}{alias};
1062 foreach my $const (@const_rows) {
1063 scalar @const_keys or do {
1064 @const_keys = sort { length($a) <=> length($b) } keys %$const;
1066 foreach my $key (@const_keys) {
1069 my @parts = split(/\./, $key);
1071 my $data = $const->{$key};
1072 foreach my $p (@parts) {
1073 $target = $target->[1]->{$p} ||= [];
1075 if ($cur eq ".${key}" && (my @ckey = @{$collapse{$cur}||[]})) {
1076 # collapsing at this point and on final part
1077 my $pos = $collapse_pos{$cur};
1078 CK: foreach my $ck (@ckey) {
1079 if (!defined $pos->{$ck} || $pos->{$ck} ne $data->{$ck}) {
1080 $collapse_pos{$cur} = $data;
1081 delete @collapse_pos{ # clear all positioning for sub-entries
1082 grep { m/^\Q${cur}.\E/ } keys %collapse_pos
1089 if (exists $collapse{$cur}) {
1090 $target = $target->[-1];
1093 $target->[0] = $data;
1095 $info->[0] = $const->{$key};
1103 =head2 result_source
1107 =item Arguments: $result_source?
1109 =item Return Value: $result_source
1113 An accessor for the primary ResultSource object from which this ResultSet
1120 =item Arguments: $result_class?
1122 =item Return Value: $result_class
1126 An accessor for the class to use when creating row objects. Defaults to
1127 C<< result_source->result_class >> - which in most cases is the name of the
1128 L<"table"|DBIx::Class::Manual::Glossary/"ResultSource"> class.
1130 Note that changing the result_class will also remove any components
1131 that were originally loaded in the source class via
1132 L<DBIx::Class::ResultSource/load_components>. Any overloaded methods
1133 in the original source class will not run.
1138 my ($self, $result_class) = @_;
1139 if ($result_class) {
1140 $self->ensure_class_loaded($result_class);
1141 $self->_result_class($result_class);
1143 $self->_result_class;
1150 =item Arguments: $cond, \%attrs??
1152 =item Return Value: $count
1156 Performs an SQL C<COUNT> with the same query as the resultset was built
1157 with to find the number of elements. Passing arguments is equivalent to
1158 C<< $rs->search ($cond, \%attrs)->count >>
1164 return $self->search(@_)->count if @_ and defined $_[0];
1165 return scalar @{ $self->get_cache } if $self->get_cache;
1167 my $attrs = $self->_resolved_attrs_copy;
1169 # this is a little optimization - it is faster to do the limit
1170 # adjustments in software, instead of a subquery
1171 my $rows = delete $attrs->{rows};
1172 my $offset = delete $attrs->{offset};
1175 if ($self->_has_resolved_attr (qw/collapse group_by/)) {
1176 $crs = $self->_count_subq_rs ($attrs);
1179 $crs = $self->_count_rs ($attrs);
1181 my $count = $crs->next;
1183 $count -= $offset if $offset;
1184 $count = $rows if $rows and $rows < $count;
1185 $count = 0 if ($count < 0);
1194 =item Arguments: $cond, \%attrs??
1196 =item Return Value: $count_rs
1200 Same as L</count> but returns a L<DBIx::Class::ResultSetColumn> object.
1201 This can be very handy for subqueries:
1203 ->search( { amount => $some_rs->count_rs->as_query } )
1205 As with regular resultsets the SQL query will be executed only after
1206 the resultset is accessed via L</next> or L</all>. That would return
1207 the same single value obtainable via L</count>.
1213 return $self->search(@_)->count_rs if @_;
1215 # this may look like a lack of abstraction (count() does about the same)
1216 # but in fact an _rs *must* use a subquery for the limits, as the
1217 # software based limiting can not be ported if this $rs is to be used
1218 # in a subquery itself (i.e. ->as_query)
1219 if ($self->_has_resolved_attr (qw/collapse group_by offset rows/)) {
1220 return $self->_count_subq_rs;
1223 return $self->_count_rs;
1228 # returns a ResultSetColumn object tied to the count query
1231 my ($self, $attrs) = @_;
1233 my $rsrc = $self->result_source;
1234 $attrs ||= $self->_resolved_attrs;
1236 my $tmp_attrs = { %$attrs };
1238 # take off any limits, record_filter is cdbi, and no point of ordering a count
1239 delete $tmp_attrs->{$_} for (qw/select as rows offset order_by record_filter/);
1241 # overwrite the selector (supplied by the storage)
1242 $tmp_attrs->{select} = $rsrc->storage->_count_select ($rsrc, $tmp_attrs);
1243 $tmp_attrs->{as} = 'count';
1245 # read the comment on top of the actual function to see what this does
1246 $tmp_attrs->{from} = $self->_switch_to_inner_join_if_needed (
1247 $tmp_attrs->{from}, $tmp_attrs->{alias}
1250 my $tmp_rs = $rsrc->resultset_class->new($rsrc, $tmp_attrs)->get_column ('count');
1256 # same as above but uses a subquery
1258 sub _count_subq_rs {
1259 my ($self, $attrs) = @_;
1261 my $rsrc = $self->result_source;
1262 $attrs ||= $self->_resolved_attrs_copy;
1264 my $sub_attrs = { %$attrs };
1266 # extra selectors do not go in the subquery and there is no point of ordering it
1267 delete $sub_attrs->{$_} for qw/collapse select _prefetch_select as order_by/;
1269 # if we prefetch, we group_by primary keys only as this is what we would get out of the rs via ->next/->all
1270 # clobber old group_by regardless
1271 if ( keys %{$attrs->{collapse}} ) {
1272 $sub_attrs->{group_by} = [ map { "$attrs->{alias}.$_" } ($rsrc->primary_columns) ]
1275 $sub_attrs->{select} = $rsrc->storage->_subq_count_select ($rsrc, $sub_attrs);
1277 # read the comment on top of the actual function to see what this does
1278 $sub_attrs->{from} = $self->_switch_to_inner_join_if_needed (
1279 $sub_attrs->{from}, $sub_attrs->{alias}
1282 # this is so that ordering can be thrown away in things like Top limit
1283 $sub_attrs->{-for_count_only} = 1;
1285 my $sub_rs = $rsrc->resultset_class->new ($rsrc, $sub_attrs);
1288 -alias => 'count_subq',
1289 -source_handle => $rsrc->handle,
1290 count_subq => $sub_rs->as_query,
1293 # the subquery replaces this
1294 delete $attrs->{$_} for qw/where bind collapse group_by having having_bind rows offset/;
1296 return $self->_count_rs ($attrs);
1300 # The DBIC relationship chaining implementation is pretty simple - every
1301 # new related_relationship is pushed onto the {from} stack, and the {select}
1302 # window simply slides further in. This means that when we count somewhere
1303 # in the middle, we got to make sure that everything in the join chain is an
1304 # actual inner join, otherwise the count will come back with unpredictable
1305 # results (a resultset may be generated with _some_ rows regardless of if
1306 # the relation which the $rs currently selects has rows or not). E.g.
1307 # $artist_rs->cds->count - normally generates:
1308 # SELECT COUNT( * ) FROM artist me LEFT JOIN cd cds ON cds.artist = me.artistid
1309 # which actually returns the number of artists * (number of cds || 1)
1311 # So what we do here is crawl {from}, determine if the current alias is at
1312 # the top of the stack, and if not - make sure the chain is inner-joined down
1315 sub _switch_to_inner_join_if_needed {
1316 my ($self, $from, $alias) = @_;
1318 # subqueries and other oddness is naturally not supported
1320 ref $from ne 'ARRAY'
1324 ref $from->[0] ne 'HASH'
1326 ! $from->[0]{-alias}
1328 $from->[0]{-alias} eq $alias
1333 for my $j (@{$from}[1 .. $#$from]) {
1334 if ($j->[0]{-alias} eq $alias) {
1335 $switch_branch = $j->[0]{-join_path};
1340 # something else went wrong
1341 return $from unless $switch_branch;
1343 # So it looks like we will have to switch some stuff around.
1344 # local() is useless here as we will be leaving the scope
1345 # anyway, and deep cloning is just too fucking expensive
1346 # So replace the inner hashref manually
1347 my @new_from = ($from->[0]);
1348 my $sw_idx = { map { $_ => 1 } @$switch_branch };
1350 for my $j (@{$from}[1 .. $#$from]) {
1351 my $jalias = $j->[0]{-alias};
1353 if ($sw_idx->{$jalias}) {
1354 my %attrs = %{$j->[0]};
1355 delete $attrs{-join_type};
1374 =head2 count_literal
1378 =item Arguments: $sql_fragment, @bind_values
1380 =item Return Value: $count
1384 Counts the results in a literal query. Equivalent to calling L</search_literal>
1385 with the passed arguments, then L</count>.
1389 sub count_literal { shift->search_literal(@_)->count; }
1395 =item Arguments: none
1397 =item Return Value: @objects
1401 Returns all elements in the resultset. Called implicitly if the resultset
1402 is returned in list context.
1409 $self->throw_exception("all() doesn't take any arguments, you probably wanted ->search(...)->all()");
1412 return @{ $self->get_cache } if $self->get_cache;
1416 if (keys %{$self->_resolved_attrs->{collapse}}) {
1417 # Using $self->cursor->all is really just an optimisation.
1418 # If we're collapsing has_many prefetches it probably makes
1419 # very little difference, and this is cleaner than hacking
1420 # _construct_object to survive the approach
1421 $self->cursor->reset;
1422 my @row = $self->cursor->next;
1424 push(@obj, $self->_construct_object(@row));
1425 @row = (exists $self->{stashed_row}
1426 ? @{delete $self->{stashed_row}}
1427 : $self->cursor->next);
1430 @obj = map { $self->_construct_object(@$_) } $self->cursor->all;
1433 $self->set_cache(\@obj) if $self->{attrs}{cache};
1442 =item Arguments: none
1444 =item Return Value: $self
1448 Resets the resultset's cursor, so you can iterate through the elements again.
1449 Implicitly resets the storage cursor, so a subsequent L</next> will trigger
1456 delete $self->{_attrs} if exists $self->{_attrs};
1457 $self->{all_cache_position} = 0;
1458 $self->cursor->reset;
1466 =item Arguments: none
1468 =item Return Value: $object?
1472 Resets the resultset and returns an object for the first result (if the
1473 resultset returns anything).
1478 return $_[0]->reset->next;
1484 # Determines whether and what type of subquery is required for the $rs operation.
1485 # If grouping is necessary either supplies its own, or verifies the current one
1486 # After all is done delegates to the proper storage method.
1488 sub _rs_update_delete {
1489 my ($self, $op, $values) = @_;
1491 my $rsrc = $self->result_source;
1493 my $needs_group_by_subq = $self->_has_resolved_attr (qw/collapse group_by -join/);
1494 my $needs_subq = $self->_has_resolved_attr (qw/row offset/);
1496 if ($needs_group_by_subq or $needs_subq) {
1498 # make a new $rs selecting only the PKs (that's all we really need)
1499 my $attrs = $self->_resolved_attrs_copy;
1501 delete $attrs->{$_} for qw/collapse select as/;
1502 $attrs->{columns} = [ map { "$attrs->{alias}.$_" } ($self->result_source->primary_columns) ];
1504 if ($needs_group_by_subq) {
1505 # make sure no group_by was supplied, or if there is one - make sure it matches
1506 # the columns compiled above perfectly. Anything else can not be sanely executed
1507 # on most databases so croak right then and there
1509 if (my $g = $attrs->{group_by}) {
1510 my @current_group_by = map
1511 { $_ =~ /\./ ? $_ : "$attrs->{alias}.$_" }
1512 (ref $g eq 'ARRAY' ? @$g : $g );
1515 join ("\x00", sort @current_group_by)
1517 join ("\x00", sort @{$attrs->{columns}} )
1519 $self->throw_exception (
1520 "You have just attempted a $op operation on a resultset which does group_by"
1521 . ' on columns other than the primary keys, while DBIC internally needs to retrieve'
1522 . ' the primary keys in a subselect. All sane RDBMS engines do not support this'
1523 . ' kind of queries. Please retry the operation with a modified group_by or'
1524 . ' without using one at all.'
1529 $attrs->{group_by} = $attrs->{columns};
1533 my $subrs = (ref $self)->new($rsrc, $attrs);
1535 return $self->result_source->storage->_subq_update_delete($subrs, $op, $values);
1538 return $rsrc->storage->$op(
1540 $op eq 'update' ? $values : (),
1541 $self->_cond_for_update_delete,
1547 # _cond_for_update_delete
1549 # update/delete require the condition to be modified to handle
1550 # the differing SQL syntax available. This transforms the $self->{cond}
1551 # appropriately, returning the new condition.
1553 sub _cond_for_update_delete {
1554 my ($self, $full_cond) = @_;
1557 $full_cond ||= $self->{cond};
1558 # No-op. No condition, we're updating/deleting everything
1559 return $cond unless ref $full_cond;
1561 if (ref $full_cond eq 'ARRAY') {
1565 foreach my $key (keys %{$_}) {
1567 $hash{$1} = $_->{$key};
1573 elsif (ref $full_cond eq 'HASH') {
1574 if ((keys %{$full_cond})[0] eq '-and') {
1576 my @cond = @{$full_cond->{-and}};
1577 for (my $i = 0; $i < @cond; $i++) {
1578 my $entry = $cond[$i];
1580 if (ref $entry eq 'HASH') {
1581 $hash = $self->_cond_for_update_delete($entry);
1584 $entry =~ /([^.]+)$/;
1585 $hash->{$1} = $cond[++$i];
1587 push @{$cond->{-and}}, $hash;
1591 foreach my $key (keys %{$full_cond}) {
1593 $cond->{$1} = $full_cond->{$key};
1598 $self->throw_exception("Can't update/delete on resultset with condition unless hash or array");
1609 =item Arguments: \%values
1611 =item Return Value: $storage_rv
1615 Sets the specified columns in the resultset to the supplied values in a
1616 single query. Return value will be true if the update succeeded or false
1617 if no records were updated; exact type of success value is storage-dependent.
1622 my ($self, $values) = @_;
1623 $self->throw_exception('Values for update must be a hash')
1624 unless ref $values eq 'HASH';
1626 return $self->_rs_update_delete ('update', $values);
1633 =item Arguments: \%values
1635 =item Return Value: 1
1639 Fetches all objects and updates them one at a time. Note that C<update_all>
1640 will run DBIC cascade triggers, while L</update> will not.
1645 my ($self, $values) = @_;
1646 $self->throw_exception('Values for update_all must be a hash')
1647 unless ref $values eq 'HASH';
1648 foreach my $obj ($self->all) {
1649 $obj->set_columns($values)->update;
1658 =item Arguments: none
1660 =item Return Value: $storage_rv
1664 Deletes the contents of the resultset from its result source. Note that this
1665 will not run DBIC cascade triggers. See L</delete_all> if you need triggers
1666 to run. See also L<DBIx::Class::Row/delete>.
1668 Return value will be the amount of rows deleted; exact type of return value
1669 is storage-dependent.
1675 $self->throw_exception('delete does not accept any arguments')
1678 return $self->_rs_update_delete ('delete');
1685 =item Arguments: none
1687 =item Return Value: 1
1691 Fetches all objects and deletes them one at a time. Note that C<delete_all>
1692 will run DBIC cascade triggers, while L</delete> will not.
1698 $self->throw_exception('delete_all does not accept any arguments')
1701 $_->delete for $self->all;
1709 =item Arguments: \@data;
1713 Accepts either an arrayref of hashrefs or alternatively an arrayref of arrayrefs.
1714 For the arrayref of hashrefs style each hashref should be a structure suitable
1715 forsubmitting to a $resultset->create(...) method.
1717 In void context, C<insert_bulk> in L<DBIx::Class::Storage::DBI> is used
1718 to insert the data, as this is a faster method.
1720 Otherwise, each set of data is inserted into the database using
1721 L<DBIx::Class::ResultSet/create>, and the resulting objects are
1722 accumulated into an array. The array itself, or an array reference
1723 is returned depending on scalar or list context.
1725 Example: Assuming an Artist Class that has many CDs Classes relating:
1727 my $Artist_rs = $schema->resultset("Artist");
1729 ## Void Context Example
1730 $Artist_rs->populate([
1731 { artistid => 4, name => 'Manufactured Crap', cds => [
1732 { title => 'My First CD', year => 2006 },
1733 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
1736 { artistid => 5, name => 'Angsty-Whiny Girl', cds => [
1737 { title => 'My parents sold me to a record company' ,year => 2005 },
1738 { title => 'Why Am I So Ugly?', year => 2006 },
1739 { title => 'I Got Surgery and am now Popular', year => 2007 }
1744 ## Array Context Example
1745 my ($ArtistOne, $ArtistTwo, $ArtistThree) = $Artist_rs->populate([
1746 { name => "Artist One"},
1747 { name => "Artist Two"},
1748 { name => "Artist Three", cds=> [
1749 { title => "First CD", year => 2007},
1750 { title => "Second CD", year => 2008},
1754 print $ArtistOne->name; ## response is 'Artist One'
1755 print $ArtistThree->cds->count ## reponse is '2'
1757 For the arrayref of arrayrefs style, the first element should be a list of the
1758 fieldsnames to which the remaining elements are rows being inserted. For
1761 $Arstist_rs->populate([
1762 [qw/artistid name/],
1763 [100, 'A Formally Unknown Singer'],
1764 [101, 'A singer that jumped the shark two albums ago'],
1765 [102, 'An actually cool singer.'],
1768 Please note an important effect on your data when choosing between void and
1769 wantarray context. Since void context goes straight to C<insert_bulk> in
1770 L<DBIx::Class::Storage::DBI> this will skip any component that is overriding
1771 C<insert>. So if you are using something like L<DBIx-Class-UUIDColumns> to
1772 create primary keys for you, you will find that your PKs are empty. In this
1773 case you will have to use the wantarray context in order to create those
1779 my $self = shift @_;
1780 my $data = ref $_[0][0] eq 'HASH'
1781 ? $_[0] : ref $_[0][0] eq 'ARRAY' ? $self->_normalize_populate_args($_[0]) :
1782 $self->throw_exception('Populate expects an arrayref of hashes or arrayref of arrayrefs');
1784 if(defined wantarray) {
1786 foreach my $item (@$data) {
1787 push(@created, $self->create($item));
1789 return wantarray ? @created : \@created;
1791 my ($first, @rest) = @$data;
1793 my @names = grep {!ref $first->{$_}} keys %$first;
1794 my @rels = grep { $self->result_source->has_relationship($_) } keys %$first;
1795 my @pks = $self->result_source->primary_columns;
1797 ## do the belongs_to relationships
1798 foreach my $index (0..$#$data) {
1800 # delegate to create() for any dataset without primary keys with specified relationships
1801 if (grep { !defined $data->[$index]->{$_} } @pks ) {
1803 if (grep { ref $data->[$index]{$r} eq $_ } qw/HASH ARRAY/) { # a related set must be a HASH or AoH
1804 my @ret = $self->populate($data);
1810 foreach my $rel (@rels) {
1811 next unless ref $data->[$index]->{$rel} eq "HASH";
1812 my $result = $self->related_resultset($rel)->create($data->[$index]->{$rel});
1813 my ($reverse) = keys %{$self->result_source->reverse_relationship_info($rel)};
1814 my $related = $result->result_source->_resolve_condition(
1815 $result->result_source->relationship_info($reverse)->{cond},
1820 delete $data->[$index]->{$rel};
1821 $data->[$index] = {%{$data->[$index]}, %$related};
1823 push @names, keys %$related if $index == 0;
1827 ## do bulk insert on current row
1828 my @values = map { [ @$_{@names} ] } @$data;
1830 $self->result_source->storage->insert_bulk(
1831 $self->result_source,
1836 ## do the has_many relationships
1837 foreach my $item (@$data) {
1839 foreach my $rel (@rels) {
1840 next unless $item->{$rel} && ref $item->{$rel} eq "ARRAY";
1842 my $parent = $self->find(map {{$_=>$item->{$_}} } @pks)
1843 || $self->throw_exception('Cannot find the relating object.');
1845 my $child = $parent->$rel;
1847 my $related = $child->result_source->_resolve_condition(
1848 $parent->result_source->relationship_info($rel)->{cond},
1853 my @rows_to_add = ref $item->{$rel} eq 'ARRAY' ? @{$item->{$rel}} : ($item->{$rel});
1854 my @populate = map { {%$_, %$related} } @rows_to_add;
1856 $child->populate( \@populate );
1862 =head2 _normalize_populate_args ($args)
1864 Private method used by L</populate> to normalize its incoming arguments. Factored
1865 out in case you want to subclass and accept new argument structures to the
1866 L</populate> method.
1870 sub _normalize_populate_args {
1871 my ($self, $data) = @_;
1872 my @names = @{shift(@$data)};
1873 my @results_to_create;
1874 foreach my $datum (@$data) {
1875 my %result_to_create;
1876 foreach my $index (0..$#names) {
1877 $result_to_create{$names[$index]} = $$datum[$index];
1879 push @results_to_create, \%result_to_create;
1881 return \@results_to_create;
1888 =item Arguments: none
1890 =item Return Value: $pager
1894 Return Value a L<Data::Page> object for the current resultset. Only makes
1895 sense for queries with a C<page> attribute.
1897 To get the full count of entries for a paged resultset, call
1898 C<total_entries> on the L<Data::Page> object.
1905 return $self->{pager} if $self->{pager};
1907 my $attrs = $self->{attrs};
1908 $self->throw_exception("Can't create pager for non-paged rs")
1909 unless $self->{attrs}{page};
1910 $attrs->{rows} ||= 10;
1912 # throw away the paging flags and re-run the count (possibly
1913 # with a subselect) to get the real total count
1914 my $count_attrs = { %$attrs };
1915 delete $count_attrs->{$_} for qw/rows offset page pager/;
1916 my $total_count = (ref $self)->new($self->result_source, $count_attrs)->count;
1918 return $self->{pager} = Data::Page->new(
1921 $self->{attrs}{page}
1929 =item Arguments: $page_number
1931 =item Return Value: $rs
1935 Returns a resultset for the $page_number page of the resultset on which page
1936 is called, where each page contains a number of rows equal to the 'rows'
1937 attribute set on the resultset (10 by default).
1942 my ($self, $page) = @_;
1943 return (ref $self)->new($self->result_source, { %{$self->{attrs}}, page => $page });
1950 =item Arguments: \%vals
1952 =item Return Value: $rowobject
1956 Creates a new row object in the resultset's result class and returns
1957 it. The row is not inserted into the database at this point, call
1958 L<DBIx::Class::Row/insert> to do that. Calling L<DBIx::Class::Row/in_storage>
1959 will tell you whether the row object has been inserted or not.
1961 Passes the hashref of input on to L<DBIx::Class::Row/new>.
1966 my ($self, $values) = @_;
1967 $self->throw_exception( "new_result needs a hash" )
1968 unless (ref $values eq 'HASH');
1971 my $alias = $self->{attrs}{alias};
1974 defined $self->{cond}
1975 && $self->{cond} eq $DBIx::Class::ResultSource::UNRESOLVABLE_CONDITION
1977 %new = %{ $self->{attrs}{related_objects} || {} }; # nothing might have been inserted yet
1978 $new{-from_resultset} = [ keys %new ] if keys %new;
1980 $self->throw_exception(
1981 "Can't abstract implicit construct, condition not a hash"
1982 ) if ($self->{cond} && !(ref $self->{cond} eq 'HASH'));
1984 my $collapsed_cond = (
1986 ? $self->_collapse_cond($self->{cond})
1990 # precendence must be given to passed values over values inherited from
1991 # the cond, so the order here is important.
1992 my %implied = %{$self->_remove_alias($collapsed_cond, $alias)};
1993 while( my($col,$value) = each %implied ){
1994 if(ref($value) eq 'HASH' && keys(%$value) && (keys %$value)[0] eq '='){
1995 $new{$col} = $value->{'='};
1998 $new{$col} = $value if $self->_is_deterministic_value($value);
2004 %{ $self->_remove_alias($values, $alias) },
2005 -source_handle => $self->_source_handle,
2006 -result_source => $self->result_source, # DO NOT REMOVE THIS, REQUIRED
2009 return $self->result_class->new(\%new);
2012 # _is_deterministic_value
2014 # Make an effor to strip non-deterministic values from the condition,
2015 # to make sure new_result chokes less
2017 sub _is_deterministic_value {
2020 my $ref_type = ref $value;
2021 return 1 if $ref_type eq '' || $ref_type eq 'SCALAR';
2022 return 1 if Scalar::Util::blessed($value);
2026 # _has_resolved_attr
2028 # determines if the resultset defines at least one
2029 # of the attributes supplied
2031 # used to determine if a subquery is neccessary
2033 # supports some virtual attributes:
2035 # This will scan for any joins being present on the resultset.
2036 # It is not a mere key-search but a deep inspection of {from}
2039 sub _has_resolved_attr {
2040 my ($self, @attr_names) = @_;
2042 my $attrs = $self->_resolved_attrs;
2046 for my $n (@attr_names) {
2047 if (grep { $n eq $_ } (qw/-join/) ) {
2048 $extra_checks{$n}++;
2052 my $attr = $attrs->{$n};
2054 next if not defined $attr;
2056 if (ref $attr eq 'HASH') {
2057 return 1 if keys %$attr;
2059 elsif (ref $attr eq 'ARRAY') {
2067 # a resolved join is expressed as a multi-level from
2069 $extra_checks{-join}
2071 ref $attrs->{from} eq 'ARRAY'
2073 @{$attrs->{from}} > 1
2081 # Recursively collapse the condition.
2083 sub _collapse_cond {
2084 my ($self, $cond, $collapsed) = @_;
2088 if (ref $cond eq 'ARRAY') {
2089 foreach my $subcond (@$cond) {
2090 next unless ref $subcond; # -or
2091 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2094 elsif (ref $cond eq 'HASH') {
2095 if (keys %$cond and (keys %$cond)[0] eq '-and') {
2096 foreach my $subcond (@{$cond->{-and}}) {
2097 $collapsed = $self->_collapse_cond($subcond, $collapsed);
2101 foreach my $col (keys %$cond) {
2102 my $value = $cond->{$col};
2103 $collapsed->{$col} = $value;
2113 # Remove the specified alias from the specified query hash. A copy is made so
2114 # the original query is not modified.
2117 my ($self, $query, $alias) = @_;
2119 my %orig = %{ $query || {} };
2122 foreach my $key (keys %orig) {
2124 $unaliased{$key} = $orig{$key};
2127 $unaliased{$1} = $orig{$key}
2128 if $key =~ m/^(?:\Q$alias\E\.)?([^.]+)$/;
2134 =head2 as_query (EXPERIMENTAL)
2138 =item Arguments: none
2140 =item Return Value: \[ $sql, @bind ]
2144 Returns the SQL query and bind vars associated with the invocant.
2146 This is generally used as the RHS for a subquery.
2148 B<NOTE>: This feature is still experimental.
2155 my $attrs = $self->_resolved_attrs_copy;
2160 # my ($sql, \@bind, \%dbi_bind_attrs) = _select_args_to_query (...)
2161 # $sql also has no wrapping parenthesis in list ctx
2163 my $sqlbind = $self->result_source->storage
2164 ->_select_args_to_query ($attrs->{from}, $attrs->{select}, $attrs->{where}, $attrs);
2173 =item Arguments: \%vals, \%attrs?
2175 =item Return Value: $rowobject
2179 my $artist = $schema->resultset('Artist')->find_or_new(
2180 { artist => 'fred' }, { key => 'artists' });
2182 $cd->cd_to_producer->find_or_new({ producer => $producer },
2183 { key => 'primary });
2185 Find an existing record from this resultset, based on its primary
2186 key, or a unique constraint. If none exists, instantiate a new result
2187 object and return it. The object will not be saved into your storage
2188 until you call L<DBIx::Class::Row/insert> on it.
2190 You most likely want this method when looking for existing rows using
2191 a unique constraint that is not the primary key, or looking for
2194 If you want objects to be saved immediately, use L</find_or_create> instead.
2196 B<Note>: C<find_or_new> is probably not what you want when creating a
2197 new row in a table that uses primary keys supplied by the
2198 database. Passing in a primary key column with a value of I<undef>
2199 will cause L</find> to attempt to search for a row with a value of
2206 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2207 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2208 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2211 return $self->new_result($hash);
2218 =item Arguments: \%vals
2220 =item Return Value: a L<DBIx::Class::Row> $object
2224 Attempt to create a single new row or a row with multiple related rows
2225 in the table represented by the resultset (and related tables). This
2226 will not check for duplicate rows before inserting, use
2227 L</find_or_create> to do that.
2229 To create one row for this resultset, pass a hashref of key/value
2230 pairs representing the columns of the table and the values you wish to
2231 store. If the appropriate relationships are set up, foreign key fields
2232 can also be passed an object representing the foreign row, and the
2233 value will be set to its primary key.
2235 To create related objects, pass a hashref of related-object column values
2236 B<keyed on the relationship name>. If the relationship is of type C<multi>
2237 (L<DBIx::Class::Relationship/has_many>) - pass an arrayref of hashrefs.
2238 The process will correctly identify columns holding foreign keys, and will
2239 transparrently populate them from the keys of the corresponding relation.
2240 This can be applied recursively, and will work correctly for a structure
2241 with an arbitrary depth and width, as long as the relationships actually
2242 exists and the correct column data has been supplied.
2245 Instead of hashrefs of plain related data (key/value pairs), you may
2246 also pass new or inserted objects. New objects (not inserted yet, see
2247 L</new>), will be inserted into their appropriate tables.
2249 Effectively a shortcut for C<< ->new_result(\%vals)->insert >>.
2251 Example of creating a new row.
2253 $person_rs->create({
2254 name=>"Some Person",
2255 email=>"somebody@someplace.com"
2258 Example of creating a new row and also creating rows in a related C<has_many>
2259 or C<has_one> resultset. Note Arrayref.
2262 { artistid => 4, name => 'Manufactured Crap', cds => [
2263 { title => 'My First CD', year => 2006 },
2264 { title => 'Yet More Tweeny-Pop crap', year => 2007 },
2269 Example of creating a new row and also creating a row in a related
2270 C<belongs_to>resultset. Note Hashref.
2273 title=>"Music for Silly Walks",
2276 name=>"Silly Musician",
2283 my ($self, $attrs) = @_;
2284 $self->throw_exception( "create needs a hashref" )
2285 unless ref $attrs eq 'HASH';
2286 return $self->new_result($attrs)->insert;
2289 =head2 find_or_create
2293 =item Arguments: \%vals, \%attrs?
2295 =item Return Value: $rowobject
2299 $cd->cd_to_producer->find_or_create({ producer => $producer },
2300 { key => 'primary' });
2302 Tries to find a record based on its primary key or unique constraints; if none
2303 is found, creates one and returns that instead.
2305 my $cd = $schema->resultset('CD')->find_or_create({
2307 artist => 'Massive Attack',
2308 title => 'Mezzanine',
2312 Also takes an optional C<key> attribute, to search by a specific key or unique
2313 constraint. For example:
2315 my $cd = $schema->resultset('CD')->find_or_create(
2317 artist => 'Massive Attack',
2318 title => 'Mezzanine',
2320 { key => 'cd_artist_title' }
2323 B<Note>: Because find_or_create() reads from the database and then
2324 possibly inserts based on the result, this method is subject to a race
2325 condition. Another process could create a record in the table after
2326 the find has completed and before the create has started. To avoid
2327 this problem, use find_or_create() inside a transaction.
2329 B<Note>: C<find_or_create> is probably not what you want when creating
2330 a new row in a table that uses primary keys supplied by the
2331 database. Passing in a primary key column with a value of I<undef>
2332 will cause L</find> to attempt to search for a row with a value of
2335 See also L</find> and L</update_or_create>. For information on how to declare
2336 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2340 sub find_or_create {
2342 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2343 my $hash = ref $_[0] eq 'HASH' ? shift : {@_};
2344 if (keys %$hash and my $row = $self->find($hash, $attrs) ) {
2347 return $self->create($hash);
2350 =head2 update_or_create
2354 =item Arguments: \%col_values, { key => $unique_constraint }?
2356 =item Return Value: $rowobject
2360 $resultset->update_or_create({ col => $val, ... });
2362 First, searches for an existing row matching one of the unique constraints
2363 (including the primary key) on the source of this resultset. If a row is
2364 found, updates it with the other given column values. Otherwise, creates a new
2367 Takes an optional C<key> attribute to search on a specific unique constraint.
2370 # In your application
2371 my $cd = $schema->resultset('CD')->update_or_create(
2373 artist => 'Massive Attack',
2374 title => 'Mezzanine',
2377 { key => 'cd_artist_title' }
2380 $cd->cd_to_producer->update_or_create({
2381 producer => $producer,
2388 If no C<key> is specified, it searches on all unique constraints defined on the
2389 source, including the primary key.
2391 If the C<key> is specified as C<primary>, it searches only on the primary key.
2393 See also L</find> and L</find_or_create>. For information on how to declare
2394 unique constraints, see L<DBIx::Class::ResultSource/add_unique_constraint>.
2396 B<Note>: C<update_or_create> is probably not what you want when
2397 looking for a row in a table that uses primary keys supplied by the
2398 database, unless you actually have a key value. Passing in a primary
2399 key column with a value of I<undef> will cause L</find> to attempt to
2400 search for a row with a value of I<NULL>.
2404 sub update_or_create {
2406 my $attrs = (@_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {});
2407 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2409 my $row = $self->find($cond, $attrs);
2411 $row->update($cond);
2415 return $self->create($cond);
2418 =head2 update_or_new
2422 =item Arguments: \%col_values, { key => $unique_constraint }?
2424 =item Return Value: $rowobject
2428 $resultset->update_or_new({ col => $val, ... });
2430 First, searches for an existing row matching one of the unique constraints
2431 (including the primary key) on the source of this resultset. If a row is
2432 found, updates it with the other given column values. Otherwise, instantiate
2433 a new result object and return it. The object will not be saved into your storage
2434 until you call L<DBIx::Class::Row/insert> on it.
2436 Takes an optional C<key> attribute to search on a specific unique constraint.
2439 # In your application
2440 my $cd = $schema->resultset('CD')->update_or_new(
2442 artist => 'Massive Attack',
2443 title => 'Mezzanine',
2446 { key => 'cd_artist_title' }
2449 if ($cd->in_storage) {
2450 # the cd was updated
2453 # the cd is not yet in the database, let's insert it
2457 See also L</find>, L</find_or_create> and L<find_or_new>.
2463 my $attrs = ( @_ > 1 && ref $_[$#_] eq 'HASH' ? pop(@_) : {} );
2464 my $cond = ref $_[0] eq 'HASH' ? shift : {@_};
2466 my $row = $self->find( $cond, $attrs );
2467 if ( defined $row ) {
2468 $row->update($cond);
2472 return $self->new_result($cond);
2479 =item Arguments: none
2481 =item Return Value: \@cache_objects?
2485 Gets the contents of the cache for the resultset, if the cache is set.
2487 The cache is populated either by using the L</prefetch> attribute to
2488 L</search> or by calling L</set_cache>.
2500 =item Arguments: \@cache_objects
2502 =item Return Value: \@cache_objects
2506 Sets the contents of the cache for the resultset. Expects an arrayref
2507 of objects of the same class as those produced by the resultset. Note that
2508 if the cache is set the resultset will return the cached objects rather
2509 than re-querying the database even if the cache attr is not set.
2511 The contents of the cache can also be populated by using the
2512 L</prefetch> attribute to L</search>.
2517 my ( $self, $data ) = @_;
2518 $self->throw_exception("set_cache requires an arrayref")
2519 if defined($data) && (ref $data ne 'ARRAY');
2520 $self->{all_cache} = $data;
2527 =item Arguments: none
2529 =item Return Value: []
2533 Clears the cache for the resultset.
2538 shift->set_cache(undef);
2541 =head2 related_resultset
2545 =item Arguments: $relationship_name
2547 =item Return Value: $resultset
2551 Returns a related resultset for the supplied relationship name.
2553 $artist_rs = $schema->resultset('CD')->related_resultset('Artist');
2557 sub related_resultset {
2558 my ($self, $rel) = @_;
2560 $self->{related_resultsets} ||= {};
2561 return $self->{related_resultsets}{$rel} ||= do {
2562 my $rel_info = $self->result_source->relationship_info($rel);
2564 $self->throw_exception(
2565 "search_related: result source '" . $self->result_source->source_name .
2566 "' has no such relationship $rel")
2569 my ($from,$seen) = $self->_chain_relationship($rel);
2571 my $join_count = $seen->{$rel};
2572 my $alias = ($join_count > 1 ? join('_', $rel, $join_count) : $rel);
2574 #XXX - temp fix for result_class bug. There likely is a more elegant fix -groditi
2575 my %attrs = %{$self->{attrs}||{}};
2576 delete @attrs{qw(result_class alias)};
2580 if (my $cache = $self->get_cache) {
2581 if ($cache->[0] && $cache->[0]->related_resultset($rel)->get_cache) {
2582 $new_cache = [ map { @{$_->related_resultset($rel)->get_cache} }
2587 my $rel_source = $self->result_source->related_source($rel);
2591 # The reason we do this now instead of passing the alias to the
2592 # search_rs below is that if you wrap/overload resultset on the
2593 # source you need to know what alias it's -going- to have for things
2594 # to work sanely (e.g. RestrictWithObject wants to be able to add
2595 # extra query restrictions, and these may need to be $alias.)
2597 my $attrs = $rel_source->resultset_attributes;
2598 local $attrs->{alias} = $alias;
2600 $rel_source->resultset
2608 where => $self->{cond},
2613 $new->set_cache($new_cache) if $new_cache;
2618 =head2 current_source_alias
2622 =item Arguments: none
2624 =item Return Value: $source_alias
2628 Returns the current table alias for the result source this resultset is built
2629 on, that will be used in the SQL query. Usually it is C<me>.
2631 Currently the source alias that refers to the result set returned by a
2632 L</search>/L</find> family method depends on how you got to the resultset: it's
2633 C<me> by default, but eg. L</search_related> aliases it to the related result
2634 source name (and keeps C<me> referring to the original result set). The long
2635 term goal is to make L<DBIx::Class> always alias the current resultset as C<me>
2636 (and make this method unnecessary).
2638 Thus it's currently necessary to use this method in predefined queries (see
2639 L<DBIx::Class::Manual::Cookbook/Predefined searches>) when referring to the
2640 source alias of the current result set:
2642 # in a result set class
2644 my ($self, $user) = @_;
2646 my $me = $self->current_source_alias;
2648 return $self->search(
2649 "$me.modified" => $user->id,
2655 sub current_source_alias {
2658 return ($self->{attrs} || {})->{alias} || 'me';
2661 # This code is called by search_related, and makes sure there
2662 # is clear separation between the joins before, during, and
2663 # after the relationship. This information is needed later
2664 # in order to properly resolve prefetch aliases (any alias
2665 # with a relation_chain_depth less than the depth of the
2666 # current prefetch is not considered)
2668 # The increments happen in 1/2s to make it easier to correlate the
2669 # join depth with the join path. An integer means a relationship
2670 # specified via a search_related, whereas a fraction means an added
2671 # join/prefetch via attributes
2672 sub _chain_relationship {
2673 my ($self, $rel) = @_;
2674 my $source = $self->result_source;
2675 my $attrs = $self->{attrs};
2681 -source_handle => $source->handle,
2682 -alias => $attrs->{alias},
2683 $attrs->{alias} => $source->from,
2687 my $seen = { %{$attrs->{seen_join} || {} } };
2688 my $jpath = ($attrs->{seen_join} && keys %{$attrs->{seen_join}})
2689 ? $from->[-1][0]{-join_path}
2693 # we need to take the prefetch the attrs into account before we
2694 # ->_resolve_join as otherwise they get lost - captainL
2695 my $merged = $self->_merge_attr( $attrs->{join}, $attrs->{prefetch} );
2697 my @requested_joins = $source->_resolve_join(
2704 push @$from, @requested_joins;
2706 $seen->{-relation_chain_depth} += 0.5;
2708 # if $self already had a join/prefetch specified on it, the requested
2709 # $rel might very well be already included. What we do in this case
2710 # is effectively a no-op (except that we bump up the chain_depth on
2711 # the join in question so we could tell it *is* the search_related)
2715 # we consider the last one thus reverse
2716 for my $j (reverse @requested_joins) {
2717 if ($rel eq $j->[0]{-join_path}[-1]) {
2718 $j->[0]{-relation_chain_depth} += 0.5;
2724 # alternative way to scan the entire chain - not backwards compatible
2725 # for my $j (reverse @$from) {
2726 # next unless ref $j eq 'ARRAY';
2727 # if ($j->[0]{-join_path} && $j->[0]{-join_path}[-1] eq $rel) {
2728 # $j->[0]{-relation_chain_depth} += 0.5;
2729 # $already_joined++;
2734 unless ($already_joined) {
2735 push @$from, $source->_resolve_join(
2743 $seen->{-relation_chain_depth} += 0.5;
2745 return ($from,$seen);
2748 # too many times we have to do $attrs = { %{$self->_resolved_attrs} }
2749 sub _resolved_attrs_copy {
2751 return { %{$self->_resolved_attrs (@_)} };
2754 sub _resolved_attrs {
2756 return $self->{_attrs} if $self->{_attrs};
2758 my $attrs = { %{ $self->{attrs} || {} } };
2759 my $source = $self->result_source;
2760 my $alias = $attrs->{alias};
2762 $attrs->{columns} ||= delete $attrs->{cols} if exists $attrs->{cols};
2765 # build columns (as long as select isn't set) into a set of as/select hashes
2766 unless ( $attrs->{select} ) {
2768 ( ref($_) eq 'HASH' )
2772 /^\Q${alias}.\E(.+)$/
2783 } ( ref($attrs->{columns}) eq 'ARRAY' ) ?
2784 @{ delete $attrs->{columns}} :
2785 (delete $attrs->{columns} ||
2786 $source->storage->order_columns_for_select(
2788 [ $source->columns ]
2792 # add the additional columns on
2793 foreach ( 'include_columns', '+columns' ) {
2794 push @colbits, map {
2795 ( ref($_) eq 'HASH' )
2797 : { ( split( /\./, $_ ) )[-1] => ( /\./ ? $_ : "${alias}.$_" ) }
2798 } ( ref($attrs->{$_}) eq 'ARRAY' ) ? @{ delete $attrs->{$_} } : delete $attrs->{$_} if ( $attrs->{$_} );
2801 # start with initial select items
2802 if ( $attrs->{select} ) {
2804 ( ref $attrs->{select} eq 'ARRAY' )
2805 ? [ @{ $attrs->{select} } ]
2806 : [ $attrs->{select} ];
2810 ref $attrs->{as} eq 'ARRAY'
2811 ? [ @{ $attrs->{as} } ]
2814 : [ map { m/^\Q${alias}.\E(.+)$/ ? $1 : $_ } @{ $attrs->{select} } ]
2819 # otherwise we intialise select & as to empty
2820 $attrs->{select} = [];
2824 # now add colbits to select/as
2825 push( @{ $attrs->{select} }, map { values( %{$_} ) } @colbits );
2826 push( @{ $attrs->{as} }, map { keys( %{$_} ) } @colbits );
2829 if ( $adds = delete $attrs->{'+select'} ) {
2830 $adds = [$adds] unless ref $adds eq 'ARRAY';
2832 @{ $attrs->{select} },
2833 map { /\./ || ref $_ ? $_ : "${alias}.$_" } @$adds
2836 if ( $adds = delete $attrs->{'+as'} ) {
2837 $adds = [$adds] unless ref $adds eq 'ARRAY';
2838 push( @{ $attrs->{as} }, @$adds );
2841 $attrs->{from} ||= [ {
2842 -source_handle => $source->handle,
2843 -alias => $self->{attrs}{alias},
2844 $self->{attrs}{alias} => $source->from,
2847 if ( $attrs->{join} || $attrs->{prefetch} ) {
2849 $self->throw_exception ('join/prefetch can not be used with a literal scalarref {from}')
2850 if ref $attrs->{from} ne 'ARRAY';
2852 my $join = delete $attrs->{join} || {};
2854 if ( defined $attrs->{prefetch} ) {
2855 $join = $self->_merge_attr( $join, $attrs->{prefetch} );
2858 $attrs->{from} = # have to copy here to avoid corrupting the original
2860 @{ $attrs->{from} },
2861 $source->_resolve_join(
2864 { %{ $attrs->{seen_join} || {} } },
2865 ($attrs->{seen_join} && keys %{$attrs->{seen_join}})
2866 ? $attrs->{from}[-1][0]{-join_path}
2873 if ( defined $attrs->{order_by} ) {
2874 $attrs->{order_by} = (
2875 ref( $attrs->{order_by} ) eq 'ARRAY'
2876 ? [ @{ $attrs->{order_by} } ]
2877 : [ $attrs->{order_by} || () ]
2881 if ($attrs->{group_by} and ! ref $attrs->{group_by}) {
2882 $attrs->{group_by} = [ $attrs->{group_by} ];
2885 # generate the distinct induced group_by early, as prefetch will be carried via a
2886 # subquery (since a group_by is present)
2887 if (delete $attrs->{distinct}) {
2888 $attrs->{group_by} ||= [ grep { !ref($_) || (ref($_) ne 'HASH') } @{$attrs->{select}} ];
2891 $attrs->{collapse} ||= {};
2892 if ( my $prefetch = delete $attrs->{prefetch} ) {
2893 $prefetch = $self->_merge_attr( {}, $prefetch );
2895 my $prefetch_ordering = [];
2897 my $join_map = $self->_joinpath_aliases ($attrs->{from}, $attrs->{seen_join});
2900 $source->_resolve_prefetch( $prefetch, $alias, $join_map, $prefetch_ordering, $attrs->{collapse} );
2902 # we need to somehow mark which columns came from prefetch
2903 $attrs->{_prefetch_select} = [ map { $_->[0] } @prefetch ];
2905 push @{ $attrs->{select} }, @{$attrs->{_prefetch_select}};
2906 push @{ $attrs->{as} }, (map { $_->[1] } @prefetch);
2908 push( @{$attrs->{order_by}}, @$prefetch_ordering );
2909 $attrs->{_collapse_order_by} = \@$prefetch_ordering;
2912 # if both page and offset are specified, produce a combined offset
2913 # even though it doesn't make much sense, this is what pre 081xx has
2915 if (my $page = delete $attrs->{page}) {
2917 ($attrs->{rows} * ($page - 1))
2919 ($attrs->{offset} || 0)
2923 return $self->{_attrs} = $attrs;
2926 sub _joinpath_aliases {
2927 my ($self, $fromspec, $seen) = @_;
2930 return $paths unless ref $fromspec eq 'ARRAY';
2932 my $cur_depth = $seen->{-relation_chain_depth} || 0;
2934 if (int ($cur_depth) != $cur_depth) {
2935 $self->throw_exception ("-relation_chain_depth is not an integer, something went horribly wrong ($cur_depth)");
2938 for my $j (@$fromspec) {
2940 next if ref $j ne 'ARRAY';
2941 next if ($j->[0]{-relation_chain_depth} || 0) < $cur_depth;
2943 my $jpath = $j->[0]{-join_path};
2946 $p = $p->{$_} ||= {} for @{$jpath}[$cur_depth .. $#$jpath];
2947 push @{$p->{-join_aliases} }, $j->[0]{-alias};
2954 my ($self, $attr) = @_;
2956 if (ref $attr eq 'HASH') {
2957 return $self->_rollout_hash($attr);
2958 } elsif (ref $attr eq 'ARRAY') {
2959 return $self->_rollout_array($attr);
2965 sub _rollout_array {
2966 my ($self, $attr) = @_;
2969 foreach my $element (@{$attr}) {
2970 if (ref $element eq 'HASH') {
2971 push( @rolled_array, @{ $self->_rollout_hash( $element ) } );
2972 } elsif (ref $element eq 'ARRAY') {
2973 # XXX - should probably recurse here
2974 push( @rolled_array, @{$self->_rollout_array($element)} );
2976 push( @rolled_array, $element );
2979 return \@rolled_array;
2983 my ($self, $attr) = @_;
2986 foreach my $key (keys %{$attr}) {
2987 push( @rolled_array, { $key => $attr->{$key} } );
2989 return \@rolled_array;
2992 sub _calculate_score {
2993 my ($self, $a, $b) = @_;
2995 if (ref $b eq 'HASH') {
2996 my ($b_key) = keys %{$b};
2997 if (ref $a eq 'HASH') {
2998 my ($a_key) = keys %{$a};
2999 if ($a_key eq $b_key) {
3000 return (1 + $self->_calculate_score( $a->{$a_key}, $b->{$b_key} ));
3005 return ($a eq $b_key) ? 1 : 0;
3008 if (ref $a eq 'HASH') {
3009 my ($a_key) = keys %{$a};
3010 return ($b eq $a_key) ? 1 : 0;
3012 return ($b eq $a) ? 1 : 0;
3018 my ($self, $orig, $import) = @_;
3020 return $import unless defined($orig);
3021 return $orig unless defined($import);
3023 $orig = $self->_rollout_attr($orig);
3024 $import = $self->_rollout_attr($import);
3027 foreach my $import_element ( @{$import} ) {
3028 # find best candidate from $orig to merge $b_element into
3029 my $best_candidate = { position => undef, score => 0 }; my $position = 0;
3030 foreach my $orig_element ( @{$orig} ) {
3031 my $score = $self->_calculate_score( $orig_element, $import_element );
3032 if ($score > $best_candidate->{score}) {
3033 $best_candidate->{position} = $position;
3034 $best_candidate->{score} = $score;
3038 my ($import_key) = ( ref $import_element eq 'HASH' ) ? keys %{$import_element} : ($import_element);
3040 if ($best_candidate->{score} == 0 || exists $seen_keys->{$import_key}) {
3041 push( @{$orig}, $import_element );
3043 my $orig_best = $orig->[$best_candidate->{position}];
3044 # merge orig_best and b_element together and replace original with merged
3045 if (ref $orig_best ne 'HASH') {
3046 $orig->[$best_candidate->{position}] = $import_element;
3047 } elsif (ref $import_element eq 'HASH') {
3048 my ($key) = keys %{$orig_best};
3049 $orig->[$best_candidate->{position}] = { $key => $self->_merge_attr($orig_best->{$key}, $import_element->{$key}) };
3052 $seen_keys->{$import_key} = 1; # don't merge the same key twice
3062 $self->_source_handle($_[0]->handle);
3064 $self->_source_handle->resolve;
3068 =head2 throw_exception
3070 See L<DBIx::Class::Schema/throw_exception> for details.
3074 sub throw_exception {
3076 if (ref $self && $self->_source_handle->schema) {
3077 $self->_source_handle->schema->throw_exception(@_)
3084 # XXX: FIXME: Attributes docs need clearing up
3088 Attributes are used to refine a ResultSet in various ways when
3089 searching for data. They can be passed to any method which takes an
3090 C<\%attrs> argument. See L</search>, L</search_rs>, L</find>,
3093 These are in no particular order:
3099 =item Value: ( $order_by | \@order_by | \%order_by )
3103 Which column(s) to order the results by.
3105 [The full list of suitable values is documented in
3106 L<SQL::Abstract/"ORDER BY CLAUSES">; the following is a summary of
3109 If a single column name, or an arrayref of names is supplied, the
3110 argument is passed through directly to SQL. The hashref syntax allows
3111 for connection-agnostic specification of ordering direction:
3113 For descending order:
3115 order_by => { -desc => [qw/col1 col2 col3/] }
3117 For explicit ascending order:
3119 order_by => { -asc => 'col' }
3121 The old scalarref syntax (i.e. order_by => \'year DESC') is still
3122 supported, although you are strongly encouraged to use the hashref
3123 syntax as outlined above.
3129 =item Value: \@columns
3133 Shortcut to request a particular set of columns to be retrieved. Each
3134 column spec may be a string (a table column name), or a hash (in which
3135 case the key is the C<as> value, and the value is used as the C<select>
3136 expression). Adds C<me.> onto the start of any column without a C<.> in
3137 it and sets C<select> from that, then auto-populates C<as> from
3138 C<select> as normal. (You may also use the C<cols> attribute, as in
3139 earlier versions of DBIC.)
3145 =item Value: \@columns
3149 Indicates additional columns to be selected from storage. Works the same
3150 as L</columns> but adds columns to the selection. (You may also use the
3151 C<include_columns> attribute, as in earlier versions of DBIC). For
3154 $schema->resultset('CD')->search(undef, {
3155 '+columns' => ['artist.name'],
3159 would return all CDs and include a 'name' column to the information
3160 passed to object inflation. Note that the 'artist' is the name of the
3161 column (or relationship) accessor, and 'name' is the name of the column
3162 accessor in the related table.
3164 =head2 include_columns
3168 =item Value: \@columns
3172 Deprecated. Acts as a synonym for L</+columns> for backward compatibility.
3178 =item Value: \@select_columns
3182 Indicates which columns should be selected from the storage. You can use
3183 column names, or in the case of RDBMS back ends, function or stored procedure
3186 $rs = $schema->resultset('Employee')->search(undef, {
3189 { count => 'employeeid' },
3194 When you use function/stored procedure names and do not supply an C<as>
3195 attribute, the column names returned are storage-dependent. E.g. MySQL would
3196 return a column named C<count(employeeid)> in the above example.
3202 Indicates additional columns to be selected from storage. Works the same as
3203 L</select> but adds columns to the selection.
3211 Indicates additional column names for those added via L</+select>. See L</as>.
3219 =item Value: \@inflation_names
3223 Indicates column names for object inflation. That is, C<as>
3224 indicates the name that the column can be accessed as via the
3225 C<get_column> method (or via the object accessor, B<if one already
3226 exists>). It has nothing to do with the SQL code C<SELECT foo AS bar>.
3228 The C<as> attribute is used in conjunction with C<select>,
3229 usually when C<select> contains one or more function or stored
3232 $rs = $schema->resultset('Employee')->search(undef, {
3235 { count => 'employeeid' }
3237 as => ['name', 'employee_count'],
3240 my $employee = $rs->first(); # get the first Employee
3242 If the object against which the search is performed already has an accessor
3243 matching a column name specified in C<as>, the value can be retrieved using
3244 the accessor as normal:
3246 my $name = $employee->name();
3248 If on the other hand an accessor does not exist in the object, you need to
3249 use C<get_column> instead:
3251 my $employee_count = $employee->get_column('employee_count');
3253 You can create your own accessors if required - see
3254 L<DBIx::Class::Manual::Cookbook> for details.
3256 Please note: This will NOT insert an C<AS employee_count> into the SQL
3257 statement produced, it is used for internal access only. Thus
3258 attempting to use the accessor in an C<order_by> clause or similar
3259 will fail miserably.
3261 To get around this limitation, you can supply literal SQL to your
3262 C<select> attibute that contains the C<AS alias> text, eg:
3264 select => [\'myfield AS alias']
3270 =item Value: ($rel_name | \@rel_names | \%rel_names)
3274 Contains a list of relationships that should be joined for this query. For
3277 # Get CDs by Nine Inch Nails
3278 my $rs = $schema->resultset('CD')->search(
3279 { 'artist.name' => 'Nine Inch Nails' },
3280 { join => 'artist' }
3283 Can also contain a hash reference to refer to the other relation's relations.
3286 package MyApp::Schema::Track;
3287 use base qw/DBIx::Class/;
3288 __PACKAGE__->table('track');
3289 __PACKAGE__->add_columns(qw/trackid cd position title/);
3290 __PACKAGE__->set_primary_key('trackid');
3291 __PACKAGE__->belongs_to(cd => 'MyApp::Schema::CD');
3294 # In your application
3295 my $rs = $schema->resultset('Artist')->search(
3296 { 'track.title' => 'Teardrop' },
3298 join => { cd => 'track' },
3299 order_by => 'artist.name',
3303 You need to use the relationship (not the table) name in conditions,
3304 because they are aliased as such. The current table is aliased as "me", so
3305 you need to use me.column_name in order to avoid ambiguity. For example:
3307 # Get CDs from 1984 with a 'Foo' track
3308 my $rs = $schema->resultset('CD')->search(
3311 'tracks.name' => 'Foo'
3313 { join => 'tracks' }
3316 If the same join is supplied twice, it will be aliased to <rel>_2 (and
3317 similarly for a third time). For e.g.
3319 my $rs = $schema->resultset('Artist')->search({
3320 'cds.title' => 'Down to Earth',
3321 'cds_2.title' => 'Popular',
3323 join => [ qw/cds cds/ ],
3326 will return a set of all artists that have both a cd with title 'Down
3327 to Earth' and a cd with title 'Popular'.
3329 If you want to fetch related objects from other tables as well, see C<prefetch>
3332 For more help on using joins with search, see L<DBIx::Class::Manual::Joining>.
3338 =item Value: ($rel_name | \@rel_names | \%rel_names)
3342 Contains one or more relationships that should be fetched along with
3343 the main query (when they are accessed afterwards the data will
3344 already be available, without extra queries to the database). This is
3345 useful for when you know you will need the related objects, because it
3346 saves at least one query:
3348 my $rs = $schema->resultset('Tag')->search(
3357 The initial search results in SQL like the following:
3359 SELECT tag.*, cd.*, artist.* FROM tag
3360 JOIN cd ON tag.cd = cd.cdid
3361 JOIN artist ON cd.artist = artist.artistid
3363 L<DBIx::Class> has no need to go back to the database when we access the
3364 C<cd> or C<artist> relationships, which saves us two SQL statements in this
3367 Simple prefetches will be joined automatically, so there is no need
3368 for a C<join> attribute in the above search.
3370 C<prefetch> can be used with the following relationship types: C<belongs_to>,
3371 C<has_one> (or if you're using C<add_relationship>, any relationship declared
3372 with an accessor type of 'single' or 'filter'). A more complex example that
3373 prefetches an artists cds, the tracks on those cds, and the tags associted
3374 with that artist is given below (assuming many-to-many from artists to tags):
3376 my $rs = $schema->resultset('Artist')->search(
3380 { cds => 'tracks' },
3381 { artist_tags => 'tags' }
3387 B<NOTE:> If you specify a C<prefetch> attribute, the C<join> and C<select>
3388 attributes will be ignored.
3390 B<CAVEATs>: Prefetch does a lot of deep magic. As such, it may not behave
3391 exactly as you might expect.
3397 Prefetch uses the L</cache> to populate the prefetched relationships. This
3398 may or may not be what you want.
3402 If you specify a condition on a prefetched relationship, ONLY those
3403 rows that match the prefetched condition will be fetched into that relationship.
3404 This means that adding prefetch to a search() B<may alter> what is returned by
3405 traversing a relationship. So, if you have C<< Artist->has_many(CDs) >> and you do
3407 my $artist_rs = $schema->resultset('Artist')->search({
3413 my $count = $artist_rs->first->cds->count;
3415 my $artist_rs_prefetch = $artist_rs->search( {}, { prefetch => 'cds' } );
3417 my $prefetch_count = $artist_rs_prefetch->first->cds->count;
3419 cmp_ok( $count, '==', $prefetch_count, "Counts should be the same" );
3421 that cmp_ok() may or may not pass depending on the datasets involved. This
3422 behavior may or may not survive the 0.09 transition.
3434 Makes the resultset paged and specifies the page to retrieve. Effectively
3435 identical to creating a non-pages resultset and then calling ->page($page)
3438 If L<rows> attribute is not specified it defaults to 10 rows per page.
3440 When you have a paged resultset, L</count> will only return the number
3441 of rows in the page. To get the total, use the L</pager> and call
3442 C<total_entries> on it.
3452 Specifes the maximum number of rows for direct retrieval or the number of
3453 rows per page if the page attribute or method is used.
3459 =item Value: $offset
3463 Specifies the (zero-based) row number for the first row to be returned, or the
3464 of the first row of the first page if paging is used.
3470 =item Value: \@columns
3474 A arrayref of columns to group by. Can include columns of joined tables.
3476 group_by => [qw/ column1 column2 ... /]
3482 =item Value: $condition
3486 HAVING is a select statement attribute that is applied between GROUP BY and
3487 ORDER BY. It is applied to the after the grouping calculations have been
3490 having => { 'count(employee)' => { '>=', 100 } }
3496 =item Value: (0 | 1)
3500 Set to 1 to group by all columns.
3506 Adds to the WHERE clause.
3508 # only return rows WHERE deleted IS NULL for all searches
3509 __PACKAGE__->resultset_attributes({ where => { deleted => undef } }); )
3511 Can be overridden by passing C<{ where => undef }> as an attribute
3518 Set to 1 to cache search results. This prevents extra SQL queries if you
3519 revisit rows in your ResultSet:
3521 my $resultset = $schema->resultset('Artist')->search( undef, { cache => 1 } );
3523 while( my $artist = $resultset->next ) {
3527 $rs->first; # without cache, this would issue a query
3529 By default, searches are not cached.
3531 For more examples of using these attributes, see
3532 L<DBIx::Class::Manual::Cookbook>.
3538 =item Value: \@from_clause
3542 The C<from> attribute gives you manual control over the C<FROM> clause of SQL
3543 statements generated by L<DBIx::Class>, allowing you to express custom C<JOIN>
3546 NOTE: Use this on your own risk. This allows you to shoot off your foot!
3548 C<join> will usually do what you need and it is strongly recommended that you
3549 avoid using C<from> unless you cannot achieve the desired result using C<join>.
3550 And we really do mean "cannot", not just tried and failed. Attempting to use
3551 this because you're having problems with C<join> is like trying to use x86
3552 ASM because you've got a syntax error in your C. Trust us on this.
3554 Now, if you're still really, really sure you need to use this (and if you're
3555 not 100% sure, ask the mailing list first), here's an explanation of how this
3558 The syntax is as follows -
3561 { <alias1> => <table1> },
3563 { <alias2> => <table2>, -join_type => 'inner|left|right' },
3564 [], # nested JOIN (optional)
3565 { <table1.column1> => <table2.column2>, ... (more conditions) },
3567 # More of the above [ ] may follow for additional joins
3574 ON <table1.column1> = <table2.column2>
3575 <more joins may follow>
3577 An easy way to follow the examples below is to remember the following:
3579 Anything inside "[]" is a JOIN
3580 Anything inside "{}" is a condition for the enclosing JOIN
3582 The following examples utilize a "person" table in a family tree application.
3583 In order to express parent->child relationships, this table is self-joined:
3585 # Person->belongs_to('father' => 'Person');
3586 # Person->belongs_to('mother' => 'Person');
3588 C<from> can be used to nest joins. Here we return all children with a father,
3589 then search against all mothers of those children:
3591 $rs = $schema->resultset('Person')->search(
3594 alias => 'mother', # alias columns in accordance with "from"
3596 { mother => 'person' },
3599 { child => 'person' },
3601 { father => 'person' },
3602 { 'father.person_id' => 'child.father_id' }
3605 { 'mother.person_id' => 'child.mother_id' }
3612 # SELECT mother.* FROM person mother
3615 # JOIN person father
3616 # ON ( father.person_id = child.father_id )
3618 # ON ( mother.person_id = child.mother_id )
3620 The type of any join can be controlled manually. To search against only people
3621 with a father in the person table, we could explicitly use C<INNER JOIN>:
3623 $rs = $schema->resultset('Person')->search(
3626 alias => 'child', # alias columns in accordance with "from"
3628 { child => 'person' },
3630 { father => 'person', -join_type => 'inner' },
3631 { 'father.id' => 'child.father_id' }
3638 # SELECT child.* FROM person child
3639 # INNER JOIN person father ON child.father_id = father.id
3641 You can select from a subquery by passing a resultset to from as follows.
3643 $schema->resultset('Artist')->search(
3645 { alias => 'artist2',
3646 from => [ { artist2 => $artist_rs->as_query } ],
3649 # and you'll get sql like this..
3650 # SELECT artist2.artistid, artist2.name, artist2.rank, artist2.charfield FROM
3651 # ( SELECT me.artistid, me.name, me.rank, me.charfield FROM artists me ) artist2
3653 If you need to express really complex joins, you
3654 can supply literal SQL to C<from> via a scalar reference. In this case
3655 the contents of the scalar will replace the table name associated with the
3658 WARNING: This technique might very well not work as expected on chained
3659 searches - you have been warned.
3661 # Assuming the Event resultsource is defined as:
3663 MySchema::Event->add_columns (
3666 is_auto_increment => 1,
3675 MySchema::Event->set_primary_key ('sequence');
3677 # This will get back the latest event for every location. The column
3678 # selector is still provided by DBIC, all we do is add a JOIN/WHERE
3679 # combo to limit the resultset
3681 $rs = $schema->resultset('Event');
3682 $table = $rs->result_source->name;
3683 $latest = $rs->search (
3686 (SELECT e1.* FROM $table e1
3688 ON e1.location = e2.location
3689 AND e1.sequence < e2.sequence
3690 WHERE e2.sequence is NULL
3695 # Equivalent SQL (with the DBIC chunks added):
3697 SELECT me.sequence, me.location, me.type FROM
3698 (SELECT e1.* FROM events e1
3700 ON e1.location = e2.location
3701 AND e1.sequence < e2.sequence
3702 WHERE e2.sequence is NULL
3709 =item Value: ( 'update' | 'shared' )
3713 Set to 'update' for a SELECT ... FOR UPDATE or 'shared' for a SELECT