DBIx-EAV
view release on metacpan or search on metacpan
lib/DBIx/EAV/Cursor.pm view on Meta::CPAN
package DBIx::EAV::Cursor;
use Moo;
use Carp qw/ croak confess /;
use Data::Dumper;
use SQL::Abstract;
use Scalar::Util qw/ blessed /;
my $sql = SQL::Abstract->new();
has 'eav', is => 'ro', required => 1;
has 'type', is => 'ro', required => 1;
has 'query', is => 'ro', default => sub { {} };
has 'options', is => 'ro', default => sub { {} };
has '_sth', is => 'ro', lazy => 1, builder => 1, predicate => '_has_sth', clearer => '_reset_sth';
sub _check_query_already_executed {
my $self = shift;
croak "Query already executed!" if defined $self->_sth;
}
sub _build__sth {
my $self = shift;
my ($sql_query, $bind) = $self->_build_sql_query();
my ($rv, $sth) = $self->eav->table('entities')->_do($sql_query, $bind);
$sth;
}
sub _build_sql_query {
my $self = shift;
my $opts = $self->options;
my $eav = $self->eav;
my $type = $self->type;
my $entities_table = $eav->table('entities');
my ($order_by, $group_by, $having, %parser_data, %replacements);
# selected field
my @select_fields = $opts->{select} ? @{$opts->{select}}
: @{$entities_table->columns};
# distinct (before normalizing @select_fields)
if ($opts->{distinct}) {
# if has group_by, warn and ignore distinct
if ($opts->{group_by}) {
}
else {
# exclude id from group by to make the distinct effect
$opts->{group_by} = [grep { !ref && $_ !~ /^(me\.|)id$/ } @select_fields];
}
}
# normalize select fields
for (my $i = 0; $i < @select_fields; $i++) {
# literal, dont touch
next if ref $select_fields[$i] eq 'SCALAR';
my $ident = $select_fields[$i];
my ($fn, $as);
# sql function
if (ref $ident) {
$as = delete $ident->{'-as'};
($fn, $ident) = each %$ident;
unless ($as) {
$as = lc($fn.'_'.$ident);
$as =~ s/\./_/g;
}
$parser_data{aliases}{$as} = 1;
}
my $info = $self->_parse_clause_identifier($ident, \%parser_data);
$select_fields[$i] = $fn ? \(sprintf "%s( %s ) AS %s", uc $fn, $info->{replacement}, $as)
: $info->{replacement};
}
# add type criteria unless we have a subselect ('from' option)
my $type_criteria = $opts->{from} ? [] : [ entity_type_id => $type->id ];
if ($opts->{subtype_depth}) {
push @$type_criteria, [ '_parent_type_'.$_, $type->id ]
for 1 .. $opts->{subtype_depth};
}
# parse WHERE
my ($where, @bind) = $sql->where({ -and => [ $type_criteria, $self->query] });
my $i = 0;
my $where_re = qr/ ([\w._]+) (?:=|!=|<>|>|<|>=|<=|IN|IS NULL|LIKE|NOT LIKE) \?/;
while ($where =~ /$where_re/g) {
my $ident = $1;
my $info = $self->_parse_clause_identifier($ident, \%parser_data, $bind[$i]);
$bind[$i] = $info->{bind}
if exists $info->{bind};
$replacements{$ident} = $info->{replacement};
$i++;
}
# replace identifiers in WHERE
while (my ($string, $replacement) = each %replacements) {
$where =~ s/\b$string\b/$replacement/g;
}
# parse ORDER BY
if (defined $opts->{order_by}) {
%replacements = ();
$order_by = $sql->where(undef, $opts->{order_by});
while ($order_by =~ / ([\w._]+)(?: ASC| DESC|,|$)/g) {
my $ident = $1;
my $info = $self->_parse_clause_identifier($ident, \%parser_data);
die "Cursor: query error: can't order by relationship! ($ident)'"
if $info->{is_relationship};
$replacements{$ident} = $info->{replacement};
}
# replace identifiers
while (my ($string, $replacement) = each %replacements) {
$order_by =~ s/\b$string\b/$replacement/g;
}
}
# prepare prefetch attributes
# if ($opts->{prefetch}) {
# foreach my $attr (ref $opts->{prefetch} eq 'ARRAY' ? @{$opts->{prefetch}} : ($opts->{prefetch})) {
# die "unknown attribute given to prefetch: '$attr'"
# unless $attr =~ /^(?:$possible_attrs)$/;
#
# $join_attr{$attr} = 1;
# push @select_fields, "$attr.value AS $attr";
# }
# }
# parse ORDER BY
if (defined $opts->{group_by}) {
my @fields;
foreach my $ident (ref $opts->{group_by} eq 'ARRAY' ? @{$opts->{group_by}} : $opts->{group_by}) {
my $info = $self->_parse_clause_identifier($ident, \%parser_data);
die "Cursor: query error: can't group by a relationship! ($ident)'"
if $info->{is_relationship};
push @fields, $info->{replacement};
}
$group_by .= 'GROUP BY '. join(', ', @fields);
}
# parse HAVING
if (defined $opts->{having}) {
my @having_bind;
($having, @having_bind) = $sql->where($opts->{having});
push @bind, @having_bind;
$having =~ s/^\s*WHERE/HAVING/;
%replacements = ();
while ($having =~ /$where_re/g) {
my $ident = $1;
my $info = $self->_parse_clause_identifier($ident, \%parser_data);
$replacements{$ident} = $info->{replacement};
}
# replace identifiers
while (my ($string, $replacement) = each %replacements) {
$having =~ s/\b$string\b/$replacement/g;
}
}
# build sql statement
# SELECT ... FROM
# from subselect
my $from = $entities_table->name;
if (my $subquery = $opts->{from}) {
my ($sub_select, $sub_bind) = @$$subquery;
$from = "($sub_select)";
push @bind, @$sub_bind;
}
my $sql_query = $sql->select("$from AS me", \@select_fields);
# JOINs
if (my $depth = $opts->{subtype_depth}) {
my $hierarchy_table = $eav->table("type_hierarchy")->name;
$sql_query .= " LEFT JOIN $hierarchy_table AS _parent_type_1 ON (_parent_type_1.child_type_id = me.entity_type_id)";
my $i = 2;
while ($depth > 1) {
$sql_query .= sprintf(" LEFT JOIN $hierarchy_table AS _parent_type_%d ON (_parent_type_%d.child_type_id = _parent_type_%d.parent_type_id)",
$i, $i, $i - 1);
$depth--;
$i++;
}
}
$sql_query .= " $_" for @{$parser_data{joins} || []};
# WHERE, GROUP BY, HAVING, ORDER BY
$sql_query .= " $where";
$sql_query .= " $group_by" if defined $group_by;
$sql_query .= " $having" if defined $having;
$sql_query .= " $order_by" if defined $order_by;
# LIMIT / OFFSET
if ($opts->{limit}) {
die "invalid limit" unless $opts->{limit} =~ /^\d+$/;
$sql_query .= " LIMIT $opts->{limit}";
if (defined $opts->{offset}) {
die "invalid offset" unless $opts->{offset} =~ /^\d+$/;
$sql_query .= " OFFSET $opts->{offset}";
}
}
# return query and bind values
($sql_query, \@bind);
}
sub _parse_clause_identifier {
my ($self, $identifier, $parser_data, $bind_value) = @_;
# cached
return $parser_data->{cache}->{$identifier}
if exists $parser_data->{cache}->{$identifier};
my $type = $self->type;
my $eav = $self->eav;
# special case: parent_type
return $parser_data->{cache}->{$identifier} = { replacement => $identifier.'.parent_type_id' }
if $identifier =~ /^_parent_type_\d+$/;
# special case: alias
return { replacement => $identifier }
if exists $parser_data->{aliases}{$identifier};
# remove me.
$identifier =~ s/^me\.//;
# parse possibly deep related identifier
# valid formats:
# - <attr>
# - <rel>
# - <rel>+.<attr>
my @parts = split /\./, $identifier;
my @joins;
my $current_type = $type;
my $current_entity_alias = 'me';
my @rels;
for (my $i = 0; $i < @parts; $i++) {
my $id_part = $parts[$i];
if ($current_type->has_relationship($id_part)) {
my $rel = $current_type->relationship($id_part);
my ($our_side, $their_side) = $rel->{is_right_entity} ? qw/ right left / : qw/ left right /;
my $related_type = $self->eav->type_by_id($rel->{"${their_side}_entity_type_id"});
push @rels, $rel->{name};
my $current_rel_alias = join '_', @rels, 'link';
# join relationship table
unless ($parser_data->{joined}{$current_rel_alias}) {
push @{$parser_data->{joins}}, sprintf "INNER JOIN %sentity_relationships AS %s ON %s.id = %s.%s_entity_id AND %s.relationship_id = %d",
$eav->schema->table_prefix,
$current_rel_alias,
$current_entity_alias,
$current_rel_alias,
$our_side,
$current_rel_alias,
$rel->{id};
$parser_data->{joined}{$current_rel_alias} = 1;
}
# endpart is the relationship itself
if ($i == $#parts) {
if (defined $bind_value) {
die sprintf('Cursor: query error: the entity given to "%s" is not an entity of type %s.', $identifier, $related_type->name)
unless blessed $bind_value
&& $bind_value->isa('DBIx::EAV::Entity')
&& $bind_value->is_type($related_type->name);
die "Cursor: query error: the '".$related_type->name."' instance given to '$identifier' is not in storage."
unless $bind_value->in_storage;
}
# set replacement for WHERE, and change bind value to the entity id
# note: dont cache this result because bindvalue can change
return {
replacement => $current_rel_alias .'.'. $their_side.'_entity_id',
bind => $bind_value ? $bind_value->id : '',
is_relationship => 1
}
}
# step into the related type
else {
$current_type = $related_type;
$current_entity_alias = $current_entity_alias eq 'me' ? $rel->{name}
: $current_entity_alias.'_'.$rel->{name};
unless ($parser_data->{joined}{$current_entity_alias}) {
push @{$parser_data->{joins}}, sprintf "INNER JOIN %sentities AS %s ON %s.id = %s.%s_entity_id",
$eav->schema->table_prefix,
$current_entity_alias,
$current_entity_alias,
$current_rel_alias,
$their_side;
$parser_data->{joined}{$current_entity_alias} = 1;
}
}
}
elsif ($current_type->has_static_attribute($id_part)) {
# attribute allowed only at the and
confess "Cursor: query error: invalid identifier '$identifier': attribute only allowed at the and of identifier."
if $i < $#parts;
return $parser_data->{cache}->{$identifier} = {
replacement => $current_entity_alias.'.'.$id_part,
};
}
elsif ($current_type->has_attribute($id_part)) {
# attribute allowed only at the and
confess "Cursor: query error: invalid identifier '$identifier': attribute only allowed at the and of identifier."
if $i < $#parts;
my $attr = $current_type->attribute($id_part);
my $join_alias = $current_entity_alias eq 'me' ? $attr->{name}
: $current_entity_alias.'_'.$attr->{name};
unless ($parser_data->{joined}{$join_alias}) {
push @{$parser_data->{joins}}, sprintf "LEFT JOIN %svalue_%s AS %s ON (%s.entity_id = %s.id AND %s.attribute_id = %s)",
$eav->schema->table_prefix,
$attr->{data_type},
$join_alias,
$join_alias,
$current_entity_alias,
$join_alias,
$attr->{id};
$parser_data->{joined}{$join_alias} = 1;
}
return { replacement => $join_alias.'.value' }
}
else {
die sprintf "Cursor: query error: invalid identifier '%s': '%s' is not a valid attribute/relationship for '%s'\n",
$identifier,
$id_part,
$current_type->name;
}
}
}
sub as_query {
\[shift->_build_sql_query];
}
sub reset {
my $self = shift;
$self->_reset_sth;
$self;
}
sub first {
$_[0]->reset->next;
}
sub next {
my $self = shift;
$self->_sth->fetchrow_hashref;
}
sub all {
my $self = shift;
my @rows;
$self->reset;
while (my $row = $self->next) {
push @rows, $row;
}
$self->reset;
return wantarray ? @rows : \@rows;
}
1;
__END__
=head1 NAME
DBIx::EAV::Cursor - Represents a query used for fetching entities.
=head1 SYNOPSIS
# get cursor from resultset
my $cursor = $eav->resultset('CD')->search(\%query)->cursor;
while (my $cd = $cursor->next) {
# $cd is the raw hashref returned from database
printf "CD id: %d\n", $cd->{id};
}
=head1 DESCRIPTION
A cursor is used to build, execute and iterate over a SQL query for the entities
table. A a cursor instance is returned from L<find|DBIx::EAV::Collection> and from
L<get|DBIx::EAV::Entity> (when you get() related data). You will never need
to create a instance of this class yourself.
=head1 METHODS
=head2 all
=over 4
=item Arguments: none
=item Return Value: L<@entities|DBIx::EAV::Entity>
=back
Returns all entities in the result.
=head2 next
=over 4
=item Arguments: none
=item Return Value: L<$result|DBIx::EAV::Entity> | undef
=back
Returns the next element in the resultset (C<undef> if there is none).
Can be used to efficiently iterate over records in the resultset:
my $cursor = $eav->resultset('CD')->find;
while (my $cd = $cursor->next) {
print $cd->get('title');
}
Note that you need to store the cursor object, and call C<next> on it.
Calling C<< resultset('CD')->next >> repeatedly will always return the
first record from the cursor.
=head2 first
=over 4
=item Arguments: none
=item Return Value: L<$result|DBIx::EAV::Entity> | undef
=back
L<Resets|/reset> the cursor (causing a fresh query to storage) and returns
an object for the first result (or C<undef> if the resultset is empty).
=head2 reset
Deletes the current statement handle, if any. Next data fetching will trigger a
new database query.
=head1 LICENSE
Copyright (C) Carlos Fernando Avila Gratz.
This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself.
=head1 AUTHOR
Carlos Fernando Avila Gratz E<lt>cafe@kreato.com.brE<gt>
=cut
( run in 0.555 second using v1.01-cache-2.11-cpan-73692580452 )