| 1 |
package Class::DBI::Iterator; |
| 2 |
|
| 3 |
=head1 NAME |
| 4 |
|
| 5 |
Class::DBI::Iterator - Iterate over Class::DBI search results |
| 6 |
|
| 7 |
=head1 SYNOPSIS |
| 8 |
|
| 9 |
my $it = My::Class->search(foo => 'bar'); |
| 10 |
|
| 11 |
my $results = $it->count; |
| 12 |
|
| 13 |
my $first_result = $it->first; |
| 14 |
while ($it->next) { ... } |
| 15 |
|
| 16 |
my @slice = $it->slice(10,19); |
| 17 |
my $slice = $it->slice(10,19); |
| 18 |
|
| 19 |
$it->reset; |
| 20 |
|
| 21 |
$it->delete_all; |
| 22 |
|
| 23 |
=head1 DESCRIPTION |
| 24 |
|
| 25 |
Any Class::DBI search (including a has_many method) which returns multiple |
| 26 |
objects can be made to return an iterator instead simply by executing |
| 27 |
the search in scalar context. |
| 28 |
|
| 29 |
Then, rather than having to fetch all the results at the same time, you |
| 30 |
can fetch them one at a time, potentially saving a considerable amount |
| 31 |
of processing time and memory. |
| 32 |
|
| 33 |
=head1 CAVEAT |
| 34 |
|
| 35 |
Note that there is no provision for the data changing (or even being |
| 36 |
deleted) in the database inbetween performing the search and retrieving |
| 37 |
the next result. |
| 38 |
|
| 39 |
=cut |
| 40 |
|
| 41 |
use strict; |
| 42 |
use overload |
| 43 |
'0+' => 'count', |
| 44 |
fallback => 1; |
| 45 |
|
| 46 |
sub new { |
| 47 |
my ($me, $them, $data, @mapper) = @_; |
| 48 |
bless { |
| 49 |
_class => $them, |
| 50 |
_data => $data, |
| 51 |
_mapper => [@mapper], |
| 52 |
_place => 0, |
| 53 |
}, |
| 54 |
ref $me || $me; |
| 55 |
} |
| 56 |
|
| 57 |
sub set_mapping_method { |
| 58 |
my ($self, @mapper) = @_; |
| 59 |
$self->{_mapper} = [@mapper]; |
| 60 |
$self; |
| 61 |
} |
| 62 |
|
| 63 |
sub class { shift->{_class} } |
| 64 |
sub data { @{ shift->{_data} } } |
| 65 |
sub mapper { @{ shift->{_mapper} } } |
| 66 |
|
| 67 |
sub count { |
| 68 |
my $self = shift; |
| 69 |
$self->{_count} ||= scalar $self->data; |
| 70 |
} |
| 71 |
|
| 72 |
sub next { |
| 73 |
my $self = shift; |
| 74 |
my $use = $self->{_data}->[ $self->{_place}++ ] or return; |
| 75 |
my @obj = ($self->class->construct($use)); |
| 76 |
foreach my $meth ($self->mapper) { |
| 77 |
@obj = map $_->$meth(), @obj; |
| 78 |
} |
| 79 |
warn "Discarding extra inflated objects" if @obj > 1; |
| 80 |
return $obj[0]; |
| 81 |
} |
| 82 |
|
| 83 |
sub first { |
| 84 |
my $self = shift; |
| 85 |
$self->reset; |
| 86 |
return $self->next; |
| 87 |
} |
| 88 |
|
| 89 |
sub slice { |
| 90 |
my ($self, $start, $end) = @_; |
| 91 |
$end ||= $start; |
| 92 |
$self->{_place} = $start; |
| 93 |
my @return; |
| 94 |
while ($self->{_place} <= $end) { |
| 95 |
push @return, $self->next || last; |
| 96 |
} |
| 97 |
return @return if wantarray; |
| 98 |
|
| 99 |
my $slice = $self->new($self->class, \@return, $self->mapper,); |
| 100 |
return $slice; |
| 101 |
} |
| 102 |
|
| 103 |
sub delete_all { |
| 104 |
my $self = shift; |
| 105 |
my $count = $self->count or return; |
| 106 |
$self->first->delete; # to reset counter |
| 107 |
while (my $obj = $self->next) { |
| 108 |
$obj->delete; |
| 109 |
} |
| 110 |
$self->{_data} = []; |
| 111 |
1; |
| 112 |
} |
| 113 |
|
| 114 |
sub reset { shift->{_place} = 0 } |
| 115 |
|
| 116 |
1; |