1 package SQL::Abstract; # see doc at end of file
5 use Module::Runtime ();
10 use Exporter 'import';
11 our @EXPORT_OK = qw(is_plain_value is_literal_value is_undef_value);
21 *SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION = $ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}
27 #======================================================================
29 #======================================================================
31 our $VERSION = '2.000000';
33 # This would confuse some packagers
34 $VERSION = eval $VERSION if $VERSION =~ /_/; # numify for warning-free dev releases
38 # special operators (-in, -between). May be extended/overridden by user.
39 # See section WHERE: BUILTIN SPECIAL OPERATORS below for implementation
40 my @BUILTIN_SPECIAL_OPS = (
41 {regex => qr/^ (?: not \s )? between $/ix, handler => sub { die "NOPE" }},
42 {regex => qr/^ is (?: \s+ not )? $/ix, handler => sub { die "NOPE" }},
43 {regex => qr/^ (?: not \s )? in $/ix, handler => sub { die "NOPE" }},
44 {regex => qr/^ ident $/ix, handler => sub { die "NOPE" }},
45 {regex => qr/^ value $/ix, handler => sub { die "NOPE" }},
48 #======================================================================
49 # DEBUGGING AND ERROR REPORTING
50 #======================================================================
53 return unless $_[0]->{debug}; shift; # a little faster
54 my $func = (caller(1))[3];
55 warn "[$func] ", @_, "\n";
59 my($func) = (caller(1))[3];
60 Carp::carp "[$func] Warning: ", @_;
64 my($func) = (caller(1))[3];
65 Carp::croak "[$func] Fatal: ", @_;
68 sub is_literal_value ($) {
69 ref $_[0] eq 'SCALAR' ? [ ${$_[0]} ]
70 : ( ref $_[0] eq 'REF' and ref ${$_[0]} eq 'ARRAY' ) ? [ @${ $_[0] } ]
74 sub is_undef_value ($) {
78 and exists $_[0]->{-value}
79 and not defined $_[0]->{-value}
83 # FIXME XSify - this can be done so much more efficiently
84 sub is_plain_value ($) {
86 ! length ref $_[0] ? \($_[0])
88 ref $_[0] eq 'HASH' and keys %{$_[0]} == 1
90 exists $_[0]->{-value}
91 ) ? \($_[0]->{-value})
93 # reuse @_ for even moar speedz
94 defined ( $_[1] = Scalar::Util::blessed $_[0] )
96 # deliberately not using Devel::OverloadInfo - the checks we are
97 # intersted in are much more limited than the fullblown thing, and
98 # this is a very hot piece of code
100 # simply using ->can('(""') can leave behind stub methods that
101 # break actually using the overload later (see L<perldiag/Stub
102 # found while resolving method "%s" overloading "%s" in package
103 # "%s"> and the source of overload::mycan())
105 # either has stringification which DBI SHOULD prefer out of the box
106 grep { *{ (qq[${_}::(""]) }{CODE} } @{ $_[2] = mro::get_linear_isa( $_[1] ) }
108 # has nummification or boolification, AND fallback is *not* disabled
110 SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION
113 grep { *{"${_}::(0+"}{CODE} } @{$_[2]}
115 grep { *{"${_}::(bool"}{CODE} } @{$_[2]}
119 # no fallback specified at all
120 ! ( ($_[3]) = grep { *{"${_}::()"}{CODE} } @{$_[2]} )
122 # fallback explicitly undef
123 ! defined ${"$_[3]::()"}
136 #======================================================================
138 #======================================================================
142 bool => '_expand_bool',
143 row => '_expand_row',
145 func => '_expand_func',
146 values => '_expand_values',
147 list => '_expand_list',
150 (map +($_ => __PACKAGE__->make_binop_expander('_expand_between')),
151 qw(between not_between)),
152 (map +($_ => __PACKAGE__->make_binop_expander('_expand_in')),
154 (map +($_ => '_expand_op_andor'), ('and', 'or')),
155 (map +($_ => '_expand_op_is'), ('is', 'is_not')),
156 (map +($_ => __PACKAGE__->make_unop_expander("_expand_${_}")),
157 qw(ident value nest)),
158 bind => __PACKAGE__->make_unop_expander(sub { +{ -bind => $_[2] } }),
161 (map +($_, "_render_$_"),
162 qw(op func bind ident literal row values keyword)),
165 (map +($_ => '_render_op_between'), 'between', 'not_between'),
166 (map +($_ => '_render_op_in'), 'in', 'not_in'),
167 (map +($_ => '_render_unop_postfix'),
168 'is_null', 'is_not_null', 'asc', 'desc',
170 (not => '_render_unop_paren'),
171 (map +($_ => '_render_op_andor'), qw(and or)),
172 ',' => '_render_op_multop',
175 delete => [ qw(target where returning) ],
176 update => [ qw(target set where returning) ],
177 insert => [ qw(target fields from returning) ],
178 select => [ qw(select from where order_by) ],
181 'delete.from' => '_expand_delete_clause_target',
182 'update.update' => '_expand_update_clause_target',
183 'insert.into' => '_expand_insert_clause_target',
184 'insert.values' => '_expand_insert_clause_from',
187 'delete.target' => '_render_delete_clause_target',
188 'update.target' => '_render_update_clause_target',
189 'insert.target' => '_render_insert_clause_target',
190 'insert.fields' => '_render_insert_clause_fields',
191 'insert.from' => '_render_insert_clause_from',
195 foreach my $stmt (keys %{$Defaults{clauses_of}}) {
196 $Defaults{expand}{$stmt} = '_expand_statement';
197 $Defaults{render}{$stmt} = '_render_statement';
198 foreach my $clause (@{$Defaults{clauses_of}{$stmt}}) {
199 $Defaults{expand_clause}{"${stmt}.${clause}"}
200 = "_expand_${stmt}_clause_${clause}";
206 my $class = ref($self) || $self;
207 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
209 # choose our case by keeping an option around
210 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
212 # default logic for interpreting arrayrefs
213 $opt{logic} = $opt{logic} ? uc $opt{logic} : 'OR';
215 # how to return bind vars
216 $opt{bindtype} ||= 'normal';
218 # default comparison is "=", but can be overridden
221 # try to recognize which are the 'equality' and 'inequality' ops
222 # (temporary quickfix (in 2007), should go through a more seasoned API)
223 $opt{equality_op} = qr/^( \Q$opt{cmp}\E | \= )$/ix;
224 $opt{inequality_op} = qr/^( != | <> )$/ix;
226 $opt{like_op} = qr/^ (is_)?r?like $/xi;
227 $opt{not_like_op} = qr/^ (is_)?not_r?like $/xi;
230 $opt{sqltrue} ||= '1=1';
231 $opt{sqlfalse} ||= '0=1';
234 $opt{special_ops} ||= [];
237 $opt{unary_ops} ||= [];
239 # rudimentary sanity-check for user supplied bits treated as functions/operators
240 # If a purported function matches this regular expression, an exception is thrown.
241 # Literal SQL is *NOT* subject to this check, only functions (and column names
242 # when quoting is not in effect)
245 # need to guard against ()'s in column names too, but this will break tons of
246 # hacks... ideas anyone?
247 $opt{injection_guard} ||= qr/
253 $opt{expand_unary} = {};
255 foreach my $name (sort keys %Defaults) {
256 $opt{$name} = { %{$Defaults{$name}}, %{$opt{$name}||{}} };
259 if ($class ne __PACKAGE__) {
261 # check for overriden methods
263 foreach my $type (qw(insert update delete)) {
264 my $method = "_${type}_returning";
265 if (__PACKAGE__->can($method) ne $class->can($method)) {
266 my $clause = "${type}.returning";
267 $opt{expand_clause}{$clause} = sub { $_[2] },
268 $opt{render_clause}{$clause}
269 = sub { [ $_[0]->$method($_[3]) ] };
272 if (__PACKAGE__->can('_table') ne $class->can('_table')) {
273 $opt{expand_clause}{'select.from'} = sub {
274 return +{ -literal => [ $_[0]->_table($_[2]) ] };
277 if (__PACKAGE__->can('_order_by') ne $class->can('_order_by')) {
278 $opt{expand_clause}{'select.order_by'} = sub { $_[2] };
279 $opt{render_clause}{'select.order_by'} = sub {
280 [ $_[0]->_order_by($_[2]) ];
283 if (__PACKAGE__->can('_select_fields') ne $class->can('_select_fields')) {
284 $opt{expand_clause}{'select.select'} = sub { $_[2] };
285 $opt{render_clause}{'select.select'} = sub {
286 my @super = $_[0]->_select_fields($_[2]);
288 ref($super[0]) eq 'HASH'
289 ? $_[0]->render_expr($super[0])
292 return $_[0]->join_query_parts(
293 ' ', { -keyword => 'select' }, $effort
297 foreach my $type (qw(in between)) {
298 my $meth = "_where_field_".uc($type);
299 if (__PACKAGE__->can($meth) ne $class->can($meth)) {
301 my ($self, $op, $v, $k) = @_;
302 $op = join ' ', split '_', $op;
303 return +{ -literal => [
304 $self->$meth($k, $op, $v)
307 $opt{expand_op}{$_} = $exp for $type, "not_${type}";
310 if ($class->isa('DBIx::Class::SQLMaker')) {
311 $opt{warn_once_on_nest} = 1;
312 $opt{disable_old_special_ops} = 1;
313 $opt{render_clause}{'select.where'} = sub {
314 my ($sql, @bind) = $_[0]->where($_[2]);
315 s/\A\s+//, s/\s+\Z// for $sql;
316 return [ $sql, @bind ];
318 $opt{expand_op}{ident} = $class->make_unop_expander(sub {
319 my ($self, undef, $body) = @_;
320 $body = $body->from if Scalar::Util::blessed($body);
321 $self->_expand_ident(ident => $body);
324 if ($class->isa('SQL::Abstract::More')) {
325 my $orig = $opt{expand_op}{or};
326 $opt{expand_op}{or} = sub {
327 my ($self, $logop, $v, $k) = @_;
328 if ($k and ref($v) eq 'ARRAY') {
329 my ($type, $val) = @$v;
332 ref($type) eq 'HASH' and ref($val) eq 'HASH'
333 and keys %$type == 1 and keys %$val == 1
334 and (keys %$type)[0] eq (keys %$val)[0]
337 ($type) = values %$type;
338 ($val) = values %$val;
340 if ($self->is_bind_value_with_type(my $v = [ $type, $val ])) {
341 return $self->expand_expr(
342 { $k, map +($op ? { $op => $_ } : $_), { -bind => $v } }
346 return $self->$orig($logop, $v, $k);
348 $opt{render}{bind} = sub {
349 return [ '?', map +(ref($_->[0]) ? $_ : $_->[1]), $_[2] ]
354 if ($opt{lazy_join_sql_parts}) {
355 my $mod = Module::Runtime::use_module('SQL::Abstract::Parts');
356 $opt{join_sql_parts} ||= sub { $mod->new(@_) };
359 $opt{join_sql_parts} ||= sub { join $_[0], @_[1..$#_] };
361 return bless \%opt, $class;
365 my ($self, $name, $key, $value) = @_;
366 return $self->{$name}{$key} unless @_ > 3;
367 $self->{$name}{$key} = $value;
371 sub make_unop_expander {
372 my (undef, $exp) = @_;
374 my ($self, $name, $body, $k) = @_;
375 return $self->_expand_hashpair_cmp($k, { "-${name}" => $body })
377 return $self->$exp($name, $body);
381 sub make_binop_expander {
382 my (undef, $exp) = @_;
384 my ($self, $name, $body, $k) = @_;
385 $k = shift @{$body = [ @$body ]} unless defined $k;
386 $k = ref($k) ? $k : { -ident => $k };
387 return $self->$exp($name, $body, $k);
392 my ($self, $plugin, @args) = @_;
393 unless (ref $plugin) {
394 $plugin =~ s/\A\+/${\__PACKAGE__}::Plugin::/;
395 require(join('/', split '::', $plugin).'.pm');
397 $plugin->apply_to($self, @args);
402 foreach my $type (qw(
403 expand op_expand render op_render clause_expand clause_render
405 my $name = join '_', reverse split '_', $type;
406 my $singular = "${type}er";
408 eval qq{sub ${singular} {
410 return \$self->_ext_rw('${name}', \@_) if \@_ == 1;
411 return \$self->${singular}s(\@_)
412 }; 1 } or die "Method builder failed for ${singular}: $@";
413 eval qq{sub wrap_${singular} {
414 shift->wrap_${singular}s(\@_)
415 }; 1 } or die "Method builder failed for wrap_${singular}: $@";
417 eval qq{sub ${singular}s {
418 my (\$self, \@args) = \@_;
419 while (my (\$this_key, \$this_value) = splice(\@args, 0, 2)) {
420 \$self->_ext_rw('${name}', \$this_key, \$this_value);
423 }; 1 } or die "Method builder failed for ${singular}s: $@";
424 eval qq{sub wrap_${singular}s {
425 my (\$self, \@args) = \@_;
426 while (my (\$this_key, \$this_builder) = splice(\@args, 0, 2)) {
427 my \$orig = \$self->_ext_rw('${name}', \$this_key);
429 '${name}', \$this_key,
430 \$this_builder->(\$orig, '${name}', \$this_key),
434 }; 1 } or die "Method builder failed for wrap_${singular}s: $@";
435 eval qq{sub ${singular}_list { sort keys %{\$_[0]->{\$name}} }; 1; }
436 or die "Method builder failed for ${singular}_list: $@";
438 foreach my $singular (qw(unop_expander binop_expander)) {
439 eval qq{sub ${singular} { shift->${singular}s(\@_) }; 1 }
440 or die "Method builder failed for ${singular}: $@";
441 eval qq{sub ${singular}s {
442 my (\$self, \@args) = \@_;
443 while (my (\$this_key, \$this_value) = splice(\@args, 0, 2)) {
445 expand_op => \$this_key,
446 \$self->make_${singular}(\$this_value),
450 }; 1 } or die "Method builder failed for ${singular}s: $@";
454 #sub register_op { $_[0]->{is_op}{$_[1]} = 1; $_[0] }
456 sub statement_list { sort keys %{$_[0]->{clauses_of}} }
459 my ($self, $of, @clauses) = @_;
461 return @{$self->{clauses_of}{$of}||[]};
463 if (ref($clauses[0]) eq 'CODE') {
464 @clauses = $self->${\($clauses[0])}(@{$self->{clauses_of}{$of}||[]});
466 $self->{clauses_of}{$of} = \@clauses;
475 ref($self->{$_}) eq 'HASH'
484 sub sqltrue { +{ -literal => [ $_[0]->{sqltrue} ] } }
485 sub sqlfalse { +{ -literal => [ $_[0]->{sqlfalse} ] } }
487 sub _assert_pass_injection_guard {
488 if ($_[1] =~ $_[0]->{injection_guard}) {
489 my $class = ref $_[0];
490 puke "Possible SQL injection attempt '$_[1]'. If this is indeed a part of the "
491 . "desired SQL use literal SQL ( \'...' or \[ '...' ] ) or supply your own "
492 . "{injection_guard} attribute to ${class}->new()"
497 #======================================================================
499 #======================================================================
502 my ($self, $table, $data, $options) = @_;
505 if (ref($table) eq 'HASH') {
508 my %clauses = (target => $table, values => $data, %{$options||{}});
512 my @rendered = $self->render_statement({ -insert => $stmt });
513 return wantarray ? @rendered : $rendered[0];
516 sub _expand_insert_clause_target {
517 +(target => $_[0]->expand_expr($_[2], -ident));
520 sub _expand_insert_clause_fields {
522 $_[0]->expand_expr({ -list => $_[2] }, -ident)
523 ] } if ref($_[2]) eq 'ARRAY';
524 return $_[2]; # should maybe still expand somewhat?
527 sub _expand_insert_clause_from {
528 my ($self, undef, $data) = @_;
529 if (ref($data) eq 'HASH' and (keys(%$data))[0] =~ /^-/) {
530 return $self->expand_expr($data);
532 my ($f_aqt, $v_aqt) = $self->_expand_insert_values($data);
534 from => { -values => [ $v_aqt ] },
535 ($f_aqt ? (fields => $f_aqt) : ()),
539 sub _expand_insert_clause_returning {
540 +(returning => $_[0]->expand_expr({ -list => $_[2] }, -ident));
543 sub _expand_insert_values {
544 my ($self, $data) = @_;
545 if (is_literal_value($data)) {
546 (undef, $self->expand_expr($data));
548 my ($fields, $values) = (
549 ref($data) eq 'HASH' ?
550 ([ sort keys %$data ], [ @{$data}{sort keys %$data} ])
554 # no names (arrayref) means can't generate bindtype
555 !($fields) && $self->{bindtype} eq 'columns'
556 && belch "can't do 'columns' bindtype when called with arrayref";
560 ? $self->expand_expr({ -row => $fields }, -ident)
565 local our $Cur_Col_Meta = $fields->[$_];
566 $self->_expand_insert_value($values->[$_])
573 sub _render_insert_clause_fields {
574 return $_[0]->render_aqt($_[2]);
577 sub _render_insert_clause_target {
578 my ($self, undef, $from) = @_;
579 $self->join_query_parts(' ', { -keyword => 'insert into' }, $from);
582 sub _render_insert_clause_from {
583 return $_[0]->render_aqt($_[2], 1);
586 # So that subclasses can override INSERT ... RETURNING separately from
587 # UPDATE and DELETE (e.g. DBIx::Class::SQLMaker::Oracle does this)
588 sub _insert_returning { shift->_returning(@_) }
590 sub _redispatch_returning {
591 my ($self, $type, undef, $returning) = @_;
592 [ $self->${\"_${type}_returning"}({ returning => $returning }) ];
596 my ($self, $options) = @_;
598 my $f = $options->{returning};
600 my ($sql, @bind) = @{ $self->render_aqt(
601 $self->expand_expr({ -list => $f }, -ident)
603 my $rsql = $self->_sqlcase(' returning ').$sql;
604 return wantarray ? ($rsql, @bind) : $rsql;
607 sub _expand_insert_value {
610 my $k = our $Cur_Col_Meta;
612 if (ref($v) eq 'ARRAY') {
613 if ($self->{array_datatypes}) {
614 return +{ -bind => [ $k, $v ] };
616 my ($sql, @bind) = @$v;
617 $self->_assert_bindval_matches_bindtype(@bind);
618 return +{ -literal => $v };
620 if (ref($v) eq 'HASH') {
621 if (grep !/^-/, keys %$v) {
622 belch "HASH ref as bind value in insert is not supported";
623 return +{ -bind => [ $k, $v ] };
627 return +{ -bind => [ $k, undef ] };
629 return $self->expand_expr($v);
634 #======================================================================
636 #======================================================================
639 my ($self, $table, $set, $where, $options) = @_;
642 if (ref($table) eq 'HASH') {
646 @clauses{qw(target set where)} = ($table, $set, $where);
647 puke "Unsupported data type specified to \$sql->update"
648 unless ref($clauses{set}) eq 'HASH';
649 @clauses{keys %$options} = values %$options;
653 my @rendered = $self->render_statement({ -update => $stmt });
654 return wantarray ? @rendered : $rendered[0];
657 sub _render_update_clause_target {
658 my ($self, undef, $target) = @_;
659 $self->join_query_parts(' ', { -keyword => 'update' }, $target);
662 sub _update_set_values {
663 my ($self, $data) = @_;
665 return @{ $self->render_aqt(
666 $self->_expand_update_set_values(undef, $data),
670 sub _expand_update_set_values {
671 my ($self, undef, $data) = @_;
672 $self->expand_expr({ -list => [
675 $set = { -bind => $_ } unless defined $set;
676 +{ -op => [ '=', { -ident => $k }, $set ] };
682 ? ($self->{array_datatypes}
683 ? [ $k, +{ -bind => [ $k, $v ] } ]
684 : [ $k, +{ -literal => $v } ])
686 local our $Cur_Col_Meta = $k;
687 [ $k, $self->_expand_expr($v) ]
694 sub _expand_update_clause_target {
695 my ($self, undef, $target) = @_;
696 +(target => $self->expand_expr({ -list => $target }, -ident));
699 sub _expand_update_clause_set {
700 return $_[2] if ref($_[2]) eq 'HASH' and ($_[2]->{-op}||[''])->[0] eq ',';
701 +(set => $_[0]->_expand_update_set_values($_[1], $_[2]));
704 sub _expand_update_clause_where {
705 +(where => $_[0]->expand_expr($_[2]));
708 sub _expand_update_clause_returning {
709 +(returning => $_[0]->expand_expr({ -list => $_[2] }, -ident));
712 # So that subclasses can override UPDATE ... RETURNING separately from
714 sub _update_returning { shift->_returning(@_) }
718 #======================================================================
720 #======================================================================
723 my ($self, @args) = @_;
725 if (ref(my $sel = $args[0]) eq 'HASH') {
729 @clauses{qw(from select where order_by)} = @args;
731 # This oddity is to literalify since historically SQLA doesn't quote
732 # a single identifier argument, so we convert it into a literal
734 $clauses{select} = { -literal => [ $clauses{select}||'*' ] }
735 unless ref($clauses{select});
740 my @rendered = $self->render_statement({ -select => $stmt });
741 return wantarray ? @rendered : $rendered[0];
744 sub _expand_select_clause_select {
745 my ($self, undef, $select) = @_;
746 +(select => $self->expand_expr({ -list => $select }, -ident));
749 sub _expand_select_clause_from {
750 my ($self, undef, $from) = @_;
751 +(from => $self->expand_expr({ -list => $from }, -ident));
754 sub _expand_select_clause_where {
755 my ($self, undef, $where) = @_;
758 if (my $conv = $self->{convert}) {
770 ->wrap_expander(bind => $_wrap)
771 ->wrap_op_expanders(map +($_ => $_wrap), qw(ident value))
772 ->wrap_expander(func => sub {
775 my ($self, $type, $thing) = @_;
776 if (ref($thing) eq 'ARRAY' and $thing->[0] eq $conv
777 and @$thing == 2 and ref($thing->[1]) eq 'HASH'
780 or $thing->[1]{-value}
781 or $thing->[1]{-bind})
783 return { -func => $thing }; # already went through our expander
785 return $self->$orig($type, $thing);
793 return +(where => $sqla->expand_expr($where));
796 sub _expand_select_clause_order_by {
797 my ($self, undef, $order_by) = @_;
798 +(order_by => $self->_expand_order_by($order_by));
802 my ($self, $fields) = @_;
803 return $fields unless ref($fields);
804 my ($sql, @bind) = @{ $self->render_aqt(
805 $self->expand_expr({ -list => $fields }, '-ident')
807 return wantarray ? ($sql, @bind) : $sql;
810 #======================================================================
812 #======================================================================
815 my ($self, $table, $where, $options) = @_;
818 if (ref($table) eq 'HASH') {
821 my %clauses = (target => $table, where => $where, %{$options||{}});
825 my @rendered = $self->render_statement({ -delete => $stmt });
826 return wantarray ? @rendered : $rendered[0];
829 # So that subclasses can override DELETE ... RETURNING separately from
831 sub _delete_returning { shift->_returning(@_) }
833 sub _expand_delete_clause_target {
834 +(target => $_[0]->expand_expr({ -list => $_[2] }, -ident));
837 sub _expand_delete_clause_where { +(where => $_[0]->expand_expr($_[2])); }
839 sub _expand_delete_clause_returning {
840 +(returning => $_[0]->expand_expr({ -list => $_[2] }, -ident));
843 sub _render_delete_clause_target {
844 my ($self, undef, $from) = @_;
845 $self->join_query_parts(' ', { -keyword => 'delete from' }, $from);
848 #======================================================================
850 #======================================================================
854 # Finally, a separate routine just to handle WHERE clauses
856 my ($self, $where, $order) = @_;
858 local $self->{convert_where} = $self->{convert};
861 my ($sql, @bind) = defined($where)
862 ? $self->_recurse_where($where)
864 $sql = (defined $sql and length $sql) ? $self->_sqlcase(' where ') . "( $sql )" : '';
868 my ($order_sql, @order_bind) = $self->_order_by($order);
870 push @bind, @order_bind;
873 return wantarray ? ($sql, @bind) : $sql;
876 { our $Default_Scalar_To = -value }
879 my ($self, $expr, $default_scalar_to) = @_;
880 local our $Default_Scalar_To = $default_scalar_to if $default_scalar_to;
881 $self->_expand_expr($expr);
885 my ($self, $aqt, $top_level) = @_;
886 my ($k, $v, @rest) = %$aqt;
888 die "Not a node type: $k" unless $k =~ s/^-//;
889 if (my $meth = $self->{render}{$k}) {
890 local our $Render_Top_Level = $top_level;
891 return $self->$meth($k, $v)||[];
893 die "notreached: $k";
897 my ($self, $expr, $default_scalar_to) = @_;
898 return @{ $self->render_aqt(
899 $self->expand_expr($expr, $default_scalar_to)
903 sub render_statement {
904 my ($self, $expr, $default_scalar_to) = @_;
906 $self->expand_expr($expr, $default_scalar_to), 1
910 sub _expand_statement {
911 my ($self, $type, $args) = @_;
912 my $ec = $self->{expand_clause};
915 $args->{$type} = delete $args->{_}
917 my %has_clause = map +($_ => 1), @{$self->{clauses_of}{$type}};
918 return +{ "-${type}" => +{
920 my $val = $args->{$_};
921 if (defined($val) and my $exp = $ec->{"${type}.$_"}) {
922 if ((my (@exp) = $self->$exp($_ => $val)) == 1) {
927 } elsif ($has_clause{$_}) {
928 ($_ => $self->expand_expr($val))
936 sub _render_statement {
937 my ($self, $type, $args) = @_;
939 foreach my $clause (@{$self->{clauses_of}{$type}}) {
940 next unless my $clause_expr = $args->{$clause};
942 if (my $rdr = $self->{render_clause}{"${type}.${clause}"}) {
943 $self->$rdr($clause, $clause_expr, $args);
945 my $r = $self->render_aqt($clause_expr, 1);
946 next unless defined $r->[0] and length $r->[0];
947 $self->join_query_parts(' ',
948 { -keyword => $clause },
955 my $q = $self->join_query_parts(' ', @parts);
956 return $self->join_query_parts('',
957 (our $Render_Top_Level ? $q : ('(', $q, ')'))
962 my ($self, $raw) = @_;
964 return $op if grep $_->{$op}, @{$self}{qw(expand_op render_op)};
965 s/^-(?=.)//, s/\s+/_/g for $op;
970 my ($self, $expr) = @_;
971 our $Expand_Depth ||= 0; local $Expand_Depth = $Expand_Depth + 1;
972 return undef unless defined($expr);
973 if (ref($expr) eq 'HASH') {
974 return undef unless my $kc = keys %$expr;
976 return $self->_expand_logop(and => $expr);
978 my ($key, $value) = %$expr;
979 if ($key =~ /^-/ and $key =~ s/ [_\s]? \d+ $//x ) {
980 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
981 . "You probably wanted ...-and => [ $key => COND1, $key => COND2 ... ]";
983 return $self->_expand_hashpair($key, $value);
985 if (ref($expr) eq 'ARRAY') {
986 return $self->_expand_logop(lc($self->{logic}), $expr);
988 if (my $literal = is_literal_value($expr)) {
989 return +{ -literal => $literal };
991 if (!ref($expr) or Scalar::Util::blessed($expr)) {
992 return $self->_expand_scalar($expr);
997 sub _expand_hashpair {
998 my ($self, $k, $v) = @_;
999 unless (defined($k) and length($k)) {
1000 if (defined($k) and my $literal = is_literal_value($v)) {
1001 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
1002 return { -literal => $literal };
1004 puke "Supplying an empty left hand side argument is not supported";
1007 return $self->_expand_hashpair_op($k, $v);
1008 } elsif ($k =~ /^\W+$/) {
1009 my ($lhs, @rhs) = ref($v) eq 'ARRAY' ? @$v : $v;
1010 return $self->_expand_op(
1011 -op, [ $k, $self->expand_expr($lhs, -ident), @rhs ]
1014 return $self->_expand_hashpair_ident($k, $v);
1017 sub _expand_hashpair_ident {
1018 my ($self, $k, $v) = @_;
1020 local our $Cur_Col_Meta = $k;
1022 # hash with multiple or no elements is andor
1024 if (ref($v) eq 'HASH' and keys %$v != 1) {
1025 return $self->_expand_logop(and => $v, $k);
1028 # undef needs to be re-sent with cmp to achieve IS/IS NOT NULL
1030 if (is_undef_value($v)) {
1031 return $self->_expand_hashpair_cmp($k => undef);
1034 # scalars and objects get expanded as whatever requested or values
1036 if (!ref($v) or Scalar::Util::blessed($v)) {
1037 return $self->_expand_hashpair_scalar($k, $v);
1040 # single key hashref is a hashtriple
1042 if (ref($v) eq 'HASH') {
1043 return $self->_expand_hashtriple($k, %$v);
1046 # arrayref needs re-engineering over the elements
1048 if (ref($v) eq 'ARRAY') {
1049 return $self->sqlfalse unless @$v;
1050 $self->_debug("ARRAY($k) means distribute over elements");
1052 ($v->[0]||'') =~ /^-(and|or)$/i
1053 ? (shift(@{$v = [ @$v ]}), $1)
1054 : lc($self->{logic} || 'OR')
1056 return $self->_expand_logop(
1061 if (my $literal = is_literal_value($v)) {
1062 unless (length $k) {
1063 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
1066 my ($sql, @bind) = @$literal;
1067 if ($self->{bindtype} eq 'columns') {
1069 $self->_assert_bindval_matches_bindtype($_);
1072 return +{ -literal => [ $self->_quote($k).' '.$sql, @bind ] };
1077 sub _expand_scalar {
1078 my ($self, $expr) = @_;
1080 return $self->_expand_expr({ (our $Default_Scalar_To) => $expr });
1083 sub _expand_hashpair_scalar {
1084 my ($self, $k, $v) = @_;
1086 return $self->_expand_hashpair_cmp(
1087 $k, $self->_expand_scalar($v),
1091 sub _expand_hashpair_op {
1092 my ($self, $k, $v) = @_;
1094 $self->_assert_pass_injection_guard($k =~ /\A-(.*)\Z/s);
1096 my $op = $self->_normalize_op($k);
1098 my $wsop = join(' ', split '_', $op);
1100 my $is_special = List::Util::first { $wsop =~ $_->{regex} }
1101 @{$self->{special_ops}};
1105 # the old special op system requires illegality for top-level use
1108 (our $Expand_Depth) == 1
1112 $self->{disable_old_special_ops}
1113 and List::Util::first { $wsop =~ $_->{regex} } @BUILTIN_SPECIAL_OPS
1117 puke "Illegal use of top-level '-$wsop'"
1121 if (my $exp = $self->{expand}{$op}||$self->{expand_op}{$op}) {
1122 return $self->$exp($op, $v);
1125 if ($self->{render}{$op}) {
1126 return { "-${op}" => $v };
1129 # Ops prefixed with -not_ get converted
1131 if (my ($rest) = $op =~/^not_(.*)$/) {
1134 $self->_expand_expr({ "-${rest}", $v })
1140 # the old unary op system means we should touch nothing and let it work
1142 my $op = join(' ', split '_', $op);
1144 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1145 return { -op => [ $op, $v ] };
1149 my $type = $is_special || $self->{render_op}{$op} ? -op : -func;
1151 if ($self->{restore_old_unop_handling}) {
1158 and (keys %$v)[0] =~ /^-/
1159 and not $self->{render_op}{$op}
1168 if ($type eq -func and ref($v) eq 'ARRAY') {
1169 return $self->_expand_expr({ -func => [ $op, @$v ] });
1172 return $self->_expand_expr({ $type => [ $op, $v ] });
1175 sub _expand_hashpair_cmp {
1176 my ($self, $k, $v) = @_;
1177 $self->_expand_hashtriple($k, $self->{cmp}, $v);
1180 sub _expand_hashtriple {
1181 my ($self, $k, $vk, $vv) = @_;
1183 my $ik = $self->_expand_expr({ -ident => $k });
1185 my $op = $self->_normalize_op($vk);
1186 $self->_assert_pass_injection_guard($op);
1188 if ($op =~ s/ _? \d+ $//x ) {
1189 return $self->_expand_expr($k, { $vk, $vv });
1191 if (my $x = $self->{expand_op}{$op}) {
1192 local our $Cur_Col_Meta = $k;
1193 return $self->$x($op, $vv, $k);
1197 my $op = join(' ', split '_', $op);
1199 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}) {
1200 return { -op => [ $op, $ik, $vv ] };
1202 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1206 { -op => [ $op, $vv ] }
1210 if (ref($vv) eq 'ARRAY') {
1212 my $logic = (defined($raw[0]) and $raw[0] =~ /^-(and|or)$/i)
1213 ? (shift(@raw), lc $1) : 'or';
1214 my @values = map +{ $vk => $_ }, @raw;
1216 $op =~ $self->{inequality_op}
1217 or $op =~ $self->{not_like_op}
1219 if (lc($logic) eq 'or' and @values > 1) {
1220 belch "A multi-element arrayref as an argument to the inequality op '${\uc(join ' ', split '_', $op)}' "
1221 . 'is technically equivalent to an always-true 1=1 (you probably wanted '
1222 . "to say ...{ \$inequality_op => [ -and => \@values ] }... instead)"
1227 # try to DWIM on equality operators
1228 return ($self->_dwim_op_to_is($op,
1229 "Supplying an empty arrayref to '%s' is deprecated",
1230 "operator '%s' applied on an empty array (field '$k')"
1231 ) ? $self->sqlfalse : $self->sqltrue);
1233 return $self->_expand_logop($logic => \@values, $k);
1235 if (is_undef_value($vv)) {
1236 my $is = ($self->_dwim_op_to_is($op,
1237 "Supplying an undefined argument to '%s' is deprecated",
1238 "unexpected operator '%s' with undef operand",
1239 ) ? 'is' : 'is not');
1241 return $self->_expand_hashpair($k => { $is, undef });
1243 local our $Cur_Col_Meta = $k;
1247 $self->_expand_expr($vv)
1251 sub _dwim_op_to_is {
1252 my ($self, $raw, $empty, $fail) = @_;
1254 my $op = $self->_normalize_op($raw);
1256 if ($op =~ /^not$/i) {
1259 if ($op =~ $self->{equality_op}) {
1262 if ($op =~ $self->{like_op}) {
1263 belch(sprintf $empty, uc(join ' ', split '_', $op));
1266 if ($op =~ $self->{inequality_op}) {
1269 if ($op =~ $self->{not_like_op}) {
1270 belch(sprintf $empty, uc(join ' ', split '_', $op));
1273 puke(sprintf $fail, $op);
1277 my ($self, undef, $args) = @_;
1278 my ($func, @args) = @$args;
1279 return +{ -func => [ $func, map $self->expand_expr($_), @args ] };
1283 my ($self, undef, $body) = @_;
1284 unless (defined($body) or (ref($body) and ref($body) eq 'ARRAY')) {
1285 puke "-ident requires a single plain scalar argument (a quotable identifier) or an arrayref of identifier parts";
1287 my ($sep) = map +(defined() ? $_ : '.') , $self->{name_sep};
1288 my @parts = map +($sep
1289 ? map split(/\Q${sep}\E/, $_), @$_
1291 ), ref($body) ? $body : [ $body ];
1292 return { -ident => $parts[-1] } if $self->{_dequalify_idents};
1293 unless ($self->{quote_char}) {
1294 $self->_assert_pass_injection_guard($_) for @parts;
1296 return +{ -ident => \@parts };
1300 +{ -bind => [ our $Cur_Col_Meta, $_[2] ] };
1304 my ($self, undef, $args) = @_;
1305 +{ -row => [ map $self->expand_expr($_), @$args ] };
1309 my ($self, undef, $args) = @_;
1310 my ($op, @opargs) = @$args;
1311 if (my $exp = $self->{expand_op}{$op}) {
1312 return $self->$exp($op, \@opargs);
1314 if (List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1315 return { -op => [ $op, @opargs ] };
1317 +{ -op => [ $op, map $self->expand_expr($_), @opargs ] };
1321 my ($self, undef, $v) = @_;
1323 return $self->_expand_expr($v);
1325 puke "-bool => undef not supported" unless defined($v);
1326 return $self->_expand_expr({ -ident => $v });
1330 my ($self, undef, $expr) = @_;
1332 ',', map $self->expand_expr($_),
1333 @{$expr->{-op}}[1..$#{$expr->{-op}}]
1334 ] } if ref($expr) eq 'HASH' and ($expr->{-op}||[''])->[0] eq ',';
1335 return +{ -op => [ ',',
1336 map $self->expand_expr($_),
1337 ref($expr) eq 'ARRAY' ? @$expr : $expr
1342 my ($self, $logop, $v, $k) = @_;
1343 $self->${\$self->{expand_op}{$logop}}($logop, $v, $k);
1346 sub _expand_op_andor {
1347 my ($self, $logop, $v, $k) = @_;
1349 $v = [ map +{ $k, $_ },
1351 ? (map +{ $_ => $v->{$_} }, sort keys %$v)
1355 if (ref($v) eq 'HASH') {
1356 return undef unless keys %$v;
1359 map $self->_expand_expr({ $_ => $v->{$_} }),
1363 if (ref($v) eq 'ARRAY') {
1364 $logop eq 'and' or $logop eq 'or' or puke "unknown logic: $logop";
1367 (ref($_) eq 'ARRAY' and @$_)
1368 or (ref($_) eq 'HASH' and %$_)
1374 while (my ($el) = splice @expr, 0, 1) {
1375 puke "Supplying an empty left hand side argument is not supported in array-pairs"
1376 unless defined($el) and length($el);
1377 my $elref = ref($el);
1379 local our $Expand_Depth = 0;
1380 push(@res, grep defined, $self->_expand_expr({ $el, shift(@expr) }));
1381 } elsif ($elref eq 'ARRAY') {
1382 push(@res, grep defined, $self->_expand_expr($el)) if @$el;
1383 } elsif (my $l = is_literal_value($el)) {
1384 push @res, { -literal => $l };
1385 } elsif ($elref eq 'HASH') {
1386 local our $Expand_Depth = 0;
1387 push @res, grep defined, $self->_expand_expr($el) if %$el;
1393 # return $res[0] if @res == 1;
1394 return { -op => [ $logop, @res ] };
1400 my ($self, $op, $vv, $k) = @_;
1401 ($k, $vv) = @$vv unless defined $k;
1402 puke "$op can only take undef as argument"
1406 and exists($vv->{-value})
1407 and !defined($vv->{-value})
1409 return +{ -op => [ $op.'_null', $self->expand_expr($k, -ident) ] };
1412 sub _expand_between {
1413 my ($self, $op, $vv, $k) = @_;
1414 my @rhs = map $self->_expand_expr($_),
1415 ref($vv) eq 'ARRAY' ? @$vv : $vv;
1417 (@rhs == 1 and ref($rhs[0]) eq 'HASH' and $rhs[0]->{-literal})
1419 (@rhs == 2 and defined($rhs[0]) and defined($rhs[1]))
1421 puke "Operator '${\uc($op)}' requires either an arrayref with two defined values or expressions, or a single literal scalarref/arrayref-ref";
1425 $self->expand_expr($k),
1426 map $self->expand_expr($_, -value), @rhs
1431 my ($self, $op, $vv, $k) = @_;
1432 if (my $literal = is_literal_value($vv)) {
1433 my ($sql, @bind) = @$literal;
1434 my $opened_sql = $self->_open_outer_paren($sql);
1436 $op, $self->expand_expr($k, -ident),
1437 { -literal => [ $opened_sql, @bind ] }
1441 'SQL::Abstract before v1.75 used to generate incorrect SQL when the '
1442 . "-${\uc($op)} operator was given an undef-containing list: !!!AUDIT YOUR CODE "
1443 . 'AND DATA!!! (the upcoming Data::Query-based version of SQL::Abstract '
1444 . 'will emit the logically correct SQL instead of raising this exception)'
1446 puke("Argument passed to the '${\uc($op)}' operator can not be undefined")
1448 my @rhs = map $self->expand_expr($_, -value),
1449 map { defined($_) ? $_: puke($undef_err) }
1450 (ref($vv) eq 'ARRAY' ? @$vv : $vv);
1451 return $self->${\($op =~ /^not/ ? 'sqltrue' : 'sqlfalse')} unless @rhs;
1455 $self->expand_expr($k, -ident),
1461 my ($self, undef, $v) = @_;
1462 # DBIx::Class requires a nest warning to be emitted once but the private
1463 # method it overrode to do so no longer exists
1464 if ($self->{warn_once_on_nest}) {
1465 unless (our $Nest_Warned) {
1467 "-nest in search conditions is deprecated, you most probably wanted:\n"
1468 .q|{..., -and => [ \%cond0, \@cond1, \'cond2', \[ 'cond3', [ col => bind ] ], etc. ], ... }|
1473 return $self->_expand_expr($v);
1476 sub _expand_values {
1477 my ($self, undef, $values) = @_;
1478 return { -values => [
1481 ? $self->expand_expr($_)
1482 : +{ -row => [ map $self->expand_expr($_), @$_ ] }
1483 ), ref($values) eq 'ARRAY' ? @$values : $values
1487 sub _recurse_where {
1488 my ($self, $where) = @_;
1490 # Special case: top level simple string treated as literal
1492 my $where_exp = (ref($where)
1493 ? $self->_expand_select_clause_where(undef, $where)
1494 : { -literal => [ $where ] });
1496 # dispatch expanded expression
1498 my ($sql, @bind) = defined($where_exp) ? @{ $self->render_aqt($where_exp) || [] } : ();
1499 # DBIx::Class used to call _recurse_where in scalar context
1500 # something else might too...
1502 return ($sql, @bind);
1505 belch "Calling _recurse_where in scalar context is deprecated and will go away before 2.0";
1511 my ($self, undef, $ident) = @_;
1513 return [ $self->_quote($ident) ];
1517 my ($self, undef, $values) = @_;
1518 return $self->join_query_parts('',
1520 $self->_render_op(undef, [ ',', @$values ]),
1526 my ($self, undef, $rest) = @_;
1527 my ($func, @args) = @$rest;
1528 return $self->join_query_parts('',
1529 $self->_sqlcase($func),
1530 $self->join_query_parts('',
1532 $self->join_query_parts(', ', @args),
1539 my ($self, undef, $bind) = @_;
1540 return [ '?', $self->_bindtype(@$bind) ];
1543 sub _render_literal {
1544 my ($self, undef, $literal) = @_;
1545 $self->_assert_bindval_matches_bindtype(@{$literal}[1..$#$literal]);
1549 sub _render_keyword {
1550 my ($self, undef, $keyword) = @_;
1551 return [ $self->_sqlcase(
1552 ref($keyword) ? $$keyword : join ' ', split '_', $keyword
1557 my ($self, undef, $v) = @_;
1558 my ($op, @args) = @$v;
1559 if (my $r = $self->{render_op}{$op}) {
1560 return $self->$r($op, \@args);
1565 my $op = join(' ', split '_', $op);
1567 my $ss = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}};
1568 if ($ss and @args > 1) {
1569 puke "Special op '${op}' requires first value to be identifier"
1570 unless my ($ident) = map $_->{-ident}, grep ref($_) eq 'HASH', $args[0];
1571 my $k = join(($self->{name_sep}||'.'), @$ident);
1572 local our $Expand_Depth = 1;
1573 return [ $self->${\($ss->{handler})}($k, $op, $args[1]) ];
1575 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1576 return [ $self->${\($us->{handler})}($op, $args[0]) ];
1579 return $self->_render_unop_paren($op, \@args);
1583 return $self->_render_unop_prefix($op, \@args);
1585 return $self->_render_op_multop($op, \@args);
1591 sub _render_op_between {
1592 my ($self, $op, $args) = @_;
1593 my ($left, $low, $high) = @$args;
1596 puke "Single arg to between must be a literal"
1597 unless $low->{-literal};
1600 +($low, { -keyword => 'and' }, $high);
1603 return $self->join_query_parts(' ',
1604 '(', $left, { -keyword => $op }, @rh, ')',
1609 my ($self, $op, $args) = @_;
1610 my ($lhs, @rhs) = @$args;
1612 return $self->join_query_parts(' ',
1614 { -keyword => $op },
1615 $self->join_query_parts(' ',
1617 $self->join_query_parts(', ', @rhs),
1623 sub _render_op_andor {
1624 my ($self, $op, $args) = @_;
1625 return undef unless @$args;
1626 return $self->join_query_parts('', $args->[0]) if @$args == 1;
1627 my $inner = $self->_render_op_multop($op, $args);
1628 return undef unless defined($inner->[0]) and length($inner->[0]);
1629 return $self->join_query_parts(' ',
1634 sub _render_op_multop {
1635 my ($self, $op, $args) = @_;
1637 return undef unless @parts;
1638 return $self->render_aqt($parts[0]) if @parts == 1;
1639 my $join = ($op eq ','
1641 : { -keyword => " ${op} " }
1643 return $self->join_query_parts($join, @parts);
1646 sub _render_values {
1647 my ($self, undef, $values) = @_;
1648 my $inner = $self->join_query_parts(' ',
1649 { -keyword => 'values' },
1650 $self->join_query_parts(', ',
1651 ref($values) eq 'ARRAY' ? @$values : $values
1654 return $self->join_query_parts('',
1655 (our $Render_Top_Level ? $inner : ('(', $inner, ')'))
1659 sub join_query_parts {
1660 my ($self, $join, @parts) = @_;
1661 if (ref($join) eq 'HASH') {
1662 $join = $self->render_aqt($join)->[0];
1666 ? $self->render_aqt($_)
1667 : ((ref($_) eq 'ARRAY') ? $_ : [ $_ ])
1670 $self->{join_sql_parts}->(
1671 $join, grep defined && length, map $_->[0], @final
1673 (map @{$_}[1..$#$_], @final),
1677 sub _render_unop_paren {
1678 my ($self, $op, $v) = @_;
1679 return $self->join_query_parts('',
1680 '(', $self->_render_unop_prefix($op, $v), ')'
1684 sub _render_unop_prefix {
1685 my ($self, $op, $v) = @_;
1686 my $op_sql = $self->{restore_old_unop_handling}
1687 ? $self->_sqlcase($op)
1688 : { -keyword => $op };
1689 return $self->join_query_parts(' ',
1690 ($self->{restore_old_unop_handling}
1691 ? $self->_sqlcase($op)
1692 : { -keyword => \$op }),
1697 sub _render_unop_postfix {
1698 my ($self, $op, $v) = @_;
1699 return $self->join_query_parts(' ',
1700 $v->[0], { -keyword => $op },
1704 # Some databases (SQLite) treat col IN (1, 2) different from
1705 # col IN ( (1, 2) ). Use this to strip all outer parens while
1706 # adding them back in the corresponding method
1707 sub _open_outer_paren {
1708 my ($self, $sql) = @_;
1710 while (my ($inner) = $sql =~ /^ \s* \( (.*) \) \s* $/xs) {
1712 # there are closing parens inside, need the heavy duty machinery
1713 # to reevaluate the extraction starting from $sql (full reevaluation)
1714 if ($inner =~ /\)/) {
1715 require Text::Balanced;
1717 my (undef, $remainder) = do {
1718 # idiotic design - writes to $@ but *DOES NOT* throw exceptions
1720 Text::Balanced::extract_bracketed($sql, '()', qr/\s*/);
1723 # the entire expression needs to be a balanced bracketed thing
1724 # (after an extract no remainder sans trailing space)
1725 last if defined $remainder and $remainder =~ /\S/;
1734 sub _where_field_IN {
1735 my ($self, $k, $op, $vals) = @_;
1736 @{$self->_render_op_in(
1739 $self->expand_expr($k, -ident),
1740 map $self->expand_expr($_, -value),
1741 ref($vals) eq 'ARRAY' ? @$vals : $vals
1746 sub _where_field_BETWEEN {
1747 my ($self, $k, $op, $vals) = @_;
1748 @{$self->_render_op_between(
1750 [ $self->expand_expr($k, -ident), ref($vals) eq 'ARRAY' ? @$vals : $vals ]
1754 #======================================================================
1756 #======================================================================
1758 sub _expand_order_by {
1759 my ($self, $arg) = @_;
1761 return unless defined($arg) and not (ref($arg) eq 'ARRAY' and !@$arg);
1763 return $self->expand_expr({ -list => $arg })
1764 if ref($arg) eq 'HASH' and ($arg->{-op}||[''])->[0] eq ',';
1766 my $expander = sub {
1767 my ($self, $dir, $expr) = @_;
1768 my @to_expand = ref($expr) eq 'ARRAY' ? @$expr : $expr;
1769 foreach my $arg (@to_expand) {
1773 and grep /^-(asc|desc)$/, keys %$arg
1775 puke "ordering direction hash passed to order by must have exactly one key (-asc or -desc)";
1779 defined($dir) ? { -op => [ $dir =~ /^-?(.*)$/ ,=> $_ ] } : $_
1781 map $self->expand_expr($_, -ident),
1782 map ref($_) eq 'ARRAY' ? @$_ : $_, @to_expand;
1783 return undef unless @exp;
1784 return undef if @exp == 1 and not defined($exp[0]);
1785 return +{ -op => [ ',', @exp ] };
1788 local @{$self->{expand}}{qw(asc desc)} = (($expander) x 2);
1790 return $self->$expander(undef, $arg);
1794 my ($self, $arg) = @_;
1796 return '' unless defined(my $expanded = $self->_expand_order_by($arg));
1798 my ($sql, @bind) = @{ $self->render_aqt($expanded) };
1800 return '' unless length($sql);
1802 my $final_sql = $self->_sqlcase(' order by ').$sql;
1804 return $final_sql unless wantarray;
1806 return ($final_sql, @bind);
1809 # _order_by no longer needs to call this so doesn't but DBIC uses it.
1811 sub _order_by_chunks {
1812 my ($self, $arg) = @_;
1814 return () unless defined(my $expanded = $self->_expand_order_by($arg));
1816 my @res = $self->_chunkify_order_by($expanded);
1817 (ref() ? $_->[0] : $_) .= '' for @res;
1821 sub _chunkify_order_by {
1822 my ($self, $expanded) = @_;
1824 return grep length, @{ $self->render_aqt($expanded) }
1825 if $expanded->{-ident} or @{$expanded->{-literal}||[]} == 1;
1828 if (ref() eq 'HASH' and $_->{-op} and $_->{-op}[0] eq ',') {
1829 my ($comma, @list) = @{$_->{-op}};
1830 return map $self->_chunkify_order_by($_), @list;
1832 return $self->render_aqt($_);
1836 #======================================================================
1837 # DATASOURCE (FOR NOW, JUST PLAIN TABLE OR LIST OF TABLES)
1838 #======================================================================
1844 $self->expand_expr({ -list => $from }, -ident)
1849 #======================================================================
1851 #======================================================================
1853 # highly optimized, as it's called way too often
1855 # my ($self, $label) = @_;
1857 return '' unless defined $_[1];
1858 return ${$_[1]} if ref($_[1]) eq 'SCALAR';
1859 puke 'Identifier cannot be hashref' if ref($_[1]) eq 'HASH';
1861 unless ($_[0]->{quote_char}) {
1862 if (ref($_[1]) eq 'ARRAY') {
1863 return join($_[0]->{name_sep}||'.', @{$_[1]});
1865 $_[0]->_assert_pass_injection_guard($_[1]);
1870 my $qref = ref $_[0]->{quote_char};
1872 !$qref ? ($_[0]->{quote_char}, $_[0]->{quote_char})
1873 : ($qref eq 'ARRAY') ? @{$_[0]->{quote_char}}
1874 : puke "Unsupported quote_char format: $_[0]->{quote_char}";
1876 my $esc = $_[0]->{escape_char} || $r;
1878 # parts containing * are naturally unquoted
1880 $_[0]->{name_sep}||'',
1884 : do { (my $n = $_) =~ s/(\Q$esc\E|\Q$r\E)/$esc$1/g; $l . $n . $r }
1886 (ref($_[1]) eq 'ARRAY'
1890 ? split (/\Q$_[0]->{name_sep}\E/, $_[1] )
1898 # Conversion, if applicable
1900 #my ($self, $arg) = @_;
1901 if (my $conv = $_[0]->{convert_where}) {
1902 return @{ $_[0]->join_query_parts('',
1903 $_[0]->_sqlcase($conv),
1912 #my ($self, $col, @vals) = @_;
1913 # called often - tighten code
1914 return $_[0]->{bindtype} eq 'columns'
1915 ? map {[$_[1], $_]} @_[2 .. $#_]
1920 # Dies if any element of @bind is not in [colname => value] format
1921 # if bindtype is 'columns'.
1922 sub _assert_bindval_matches_bindtype {
1923 # my ($self, @bind) = @_;
1925 if ($self->{bindtype} eq 'columns') {
1927 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
1928 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
1934 # Fix SQL case, if so requested
1936 # LDNOTE: if $self->{case} is true, then it contains 'lower', so we
1937 # don't touch the argument ... crooked logic, but let's not change it!
1938 return $_[0]->{case} ? $_[1] : uc($_[1]);
1941 #======================================================================
1942 # DISPATCHING FROM REFKIND
1943 #======================================================================
1946 my ($self, $data) = @_;
1948 return 'UNDEF' unless defined $data;
1950 # blessed objects are treated like scalars
1951 my $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1953 return 'SCALAR' unless $ref;
1956 while ($ref eq 'REF') {
1958 $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1962 return ($ref||'SCALAR') . ('REF' x $n_steps);
1966 my ($self, $data) = @_;
1967 my @try = ($self->_refkind($data));
1968 push @try, 'SCALAR_or_UNDEF' if $try[0] eq 'SCALAR' || $try[0] eq 'UNDEF';
1969 push @try, 'FALLBACK';
1973 sub _METHOD_FOR_refkind {
1974 my ($self, $meth_prefix, $data) = @_;
1977 for (@{$self->_try_refkind($data)}) {
1978 $method = $self->can($meth_prefix."_".$_)
1982 return $method || puke "cannot dispatch on '$meth_prefix' for ".$self->_refkind($data);
1986 sub _SWITCH_refkind {
1987 my ($self, $data, $dispatch_table) = @_;
1990 for (@{$self->_try_refkind($data)}) {
1991 $coderef = $dispatch_table->{$_}
1995 puke "no dispatch entry for ".$self->_refkind($data)
2004 #======================================================================
2005 # VALUES, GENERATE, AUTOLOAD
2006 #======================================================================
2008 # LDNOTE: original code from nwiger, didn't touch code in that section
2009 # I feel the AUTOLOAD stuff should not be the default, it should
2010 # only be activated on explicit demand by user.
2014 my $data = shift || return;
2015 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
2016 unless ref $data eq 'HASH';
2019 foreach my $k (sort keys %$data) {
2020 my $v = $data->{$k};
2021 $self->_SWITCH_refkind($v, {
2023 if ($self->{array_datatypes}) { # array datatype
2024 push @all_bind, $self->_bindtype($k, $v);
2026 else { # literal SQL with bind
2027 my ($sql, @bind) = @$v;
2028 $self->_assert_bindval_matches_bindtype(@bind);
2029 push @all_bind, @bind;
2032 ARRAYREFREF => sub { # literal SQL with bind
2033 my ($sql, @bind) = @${$v};
2034 $self->_assert_bindval_matches_bindtype(@bind);
2035 push @all_bind, @bind;
2037 SCALARREF => sub { # literal SQL without bind
2039 SCALAR_or_UNDEF => sub {
2040 push @all_bind, $self->_bindtype($k, $v);
2051 my(@sql, @sqlq, @sqlv);
2055 if ($ref eq 'HASH') {
2056 for my $k (sort keys %$_) {
2059 my $label = $self->_quote($k);
2060 if ($r eq 'ARRAY') {
2061 # literal SQL with bind
2062 my ($sql, @bind) = @$v;
2063 $self->_assert_bindval_matches_bindtype(@bind);
2064 push @sqlq, "$label = $sql";
2066 } elsif ($r eq 'SCALAR') {
2067 # literal SQL without bind
2068 push @sqlq, "$label = $$v";
2070 push @sqlq, "$label = ?";
2071 push @sqlv, $self->_bindtype($k, $v);
2074 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
2075 } elsif ($ref eq 'ARRAY') {
2076 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
2079 if ($r eq 'ARRAY') { # literal SQL with bind
2080 my ($sql, @bind) = @$v;
2081 $self->_assert_bindval_matches_bindtype(@bind);
2084 } elsif ($r eq 'SCALAR') { # literal SQL without bind
2085 # embedded literal SQL
2092 push @sql, '(' . join(', ', @sqlq) . ')';
2093 } elsif ($ref eq 'SCALAR') {
2097 # strings get case twiddled
2098 push @sql, $self->_sqlcase($_);
2102 my $sql = join ' ', @sql;
2104 # this is pretty tricky
2105 # if ask for an array, return ($stmt, @bind)
2106 # otherwise, s/?/shift @sqlv/ to put it inline
2108 return ($sql, @sqlv);
2110 1 while $sql =~ s/\?/my $d = shift(@sqlv);
2111 ref $d ? $d->[1] : $d/e;
2120 # This allows us to check for a local, then _form, attr
2122 my($name) = $AUTOLOAD =~ /.*::(.+)/;
2123 puke "AUTOLOAD invoked for method name ${name} and allow_autoload option not set" unless $self->{allow_autoload};
2124 return $self->generate($name, @_);
2135 SQL::Abstract - Generate SQL from Perl data structures
2141 my $sql = SQL::Abstract->new;
2143 my($stmt, @bind) = $sql->select($source, \@fields, \%where, $order);
2145 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
2147 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
2149 my($stmt, @bind) = $sql->delete($table, \%where);
2151 # Then, use these in your DBI statements
2152 my $sth = $dbh->prepare($stmt);
2153 $sth->execute(@bind);
2155 # Just generate the WHERE clause
2156 my($stmt, @bind) = $sql->where(\%where, $order);
2158 # Return values in the same order, for hashed queries
2159 # See PERFORMANCE section for more details
2160 my @bind = $sql->values(\%fieldvals);
2164 This module was inspired by the excellent L<DBIx::Abstract>.
2165 However, in using that module I found that what I really wanted
2166 to do was generate SQL, but still retain complete control over my
2167 statement handles and use the DBI interface. So, I set out to
2168 create an abstract SQL generation module.
2170 While based on the concepts used by L<DBIx::Abstract>, there are
2171 several important differences, especially when it comes to WHERE
2172 clauses. I have modified the concepts used to make the SQL easier
2173 to generate from Perl data structures and, IMO, more intuitive.
2174 The underlying idea is for this module to do what you mean, based
2175 on the data structures you provide it. The big advantage is that
2176 you don't have to modify your code every time your data changes,
2177 as this module figures it out.
2179 To begin with, an SQL INSERT is as easy as just specifying a hash
2180 of C<key=value> pairs:
2183 name => 'Jimbo Bobson',
2184 phone => '123-456-7890',
2185 address => '42 Sister Lane',
2186 city => 'St. Louis',
2187 state => 'Louisiana',
2190 The SQL can then be generated with this:
2192 my($stmt, @bind) = $sql->insert('people', \%data);
2194 Which would give you something like this:
2196 $stmt = "INSERT INTO people
2197 (address, city, name, phone, state)
2198 VALUES (?, ?, ?, ?, ?)";
2199 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
2200 '123-456-7890', 'Louisiana');
2202 These are then used directly in your DBI code:
2204 my $sth = $dbh->prepare($stmt);
2205 $sth->execute(@bind);
2207 =head2 Inserting and Updating Arrays
2209 If your database has array types (like for example Postgres),
2210 activate the special option C<< array_datatypes => 1 >>
2211 when creating the C<SQL::Abstract> object.
2212 Then you may use an arrayref to insert and update database array types:
2214 my $sql = SQL::Abstract->new(array_datatypes => 1);
2216 planets => [qw/Mercury Venus Earth Mars/]
2219 my($stmt, @bind) = $sql->insert('solar_system', \%data);
2223 $stmt = "INSERT INTO solar_system (planets) VALUES (?)"
2225 @bind = (['Mercury', 'Venus', 'Earth', 'Mars']);
2228 =head2 Inserting and Updating SQL
2230 In order to apply SQL functions to elements of your C<%data> you may
2231 specify a reference to an arrayref for the given hash value. For example,
2232 if you need to execute the Oracle C<to_date> function on a value, you can
2233 say something like this:
2237 date_entered => \[ "to_date(?,'MM/DD/YYYY')", "03/02/2003" ],
2240 The first value in the array is the actual SQL. Any other values are
2241 optional and would be included in the bind values array. This gives
2244 my($stmt, @bind) = $sql->insert('people', \%data);
2246 $stmt = "INSERT INTO people (name, date_entered)
2247 VALUES (?, to_date(?,'MM/DD/YYYY'))";
2248 @bind = ('Bill', '03/02/2003');
2250 An UPDATE is just as easy, all you change is the name of the function:
2252 my($stmt, @bind) = $sql->update('people', \%data);
2254 Notice that your C<%data> isn't touched; the module will generate
2255 the appropriately quirky SQL for you automatically. Usually you'll
2256 want to specify a WHERE clause for your UPDATE, though, which is
2257 where handling C<%where> hashes comes in handy...
2259 =head2 Complex where statements
2261 This module can generate pretty complicated WHERE statements
2262 easily. For example, simple C<key=value> pairs are taken to mean
2263 equality, and if you want to see if a field is within a set
2264 of values, you can use an arrayref. Let's say we wanted to
2265 SELECT some data based on this criteria:
2268 requestor => 'inna',
2269 worker => ['nwiger', 'rcwe', 'sfz'],
2270 status => { '!=', 'completed' }
2273 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
2275 The above would give you something like this:
2277 $stmt = "SELECT * FROM tickets WHERE
2278 ( requestor = ? ) AND ( status != ? )
2279 AND ( worker = ? OR worker = ? OR worker = ? )";
2280 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
2282 Which you could then use in DBI code like so:
2284 my $sth = $dbh->prepare($stmt);
2285 $sth->execute(@bind);
2291 The methods are simple. There's one for every major SQL operation,
2292 and a constructor you use first. The arguments are specified in a
2293 similar order for each method (table, then fields, then a where
2294 clause) to try and simplify things.
2296 =head2 new(option => 'value')
2298 The C<new()> function takes a list of options and values, and returns
2299 a new B<SQL::Abstract> object which can then be used to generate SQL
2300 through the methods below. The options accepted are:
2306 If set to 'lower', then SQL will be generated in all lowercase. By
2307 default SQL is generated in "textbook" case meaning something like:
2309 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
2311 Any setting other than 'lower' is ignored.
2315 This determines what the default comparison operator is. By default
2316 it is C<=>, meaning that a hash like this:
2318 %where = (name => 'nwiger', email => 'nate@wiger.org');
2320 Will generate SQL like this:
2322 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
2324 However, you may want loose comparisons by default, so if you set
2325 C<cmp> to C<like> you would get SQL such as:
2327 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
2329 You can also override the comparison on an individual basis - see
2330 the huge section on L</"WHERE CLAUSES"> at the bottom.
2332 =item sqltrue, sqlfalse
2334 Expressions for inserting boolean values within SQL statements.
2335 By default these are C<1=1> and C<1=0>. They are used
2336 by the special operators C<-in> and C<-not_in> for generating
2337 correct SQL even when the argument is an empty array (see below).
2341 This determines the default logical operator for multiple WHERE
2342 statements in arrays or hashes. If absent, the default logic is "or"
2343 for arrays, and "and" for hashes. This means that a WHERE
2347 event_date => {'>=', '2/13/99'},
2348 event_date => {'<=', '4/24/03'},
2351 will generate SQL like this:
2353 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
2355 This is probably not what you want given this query, though (look
2356 at the dates). To change the "OR" to an "AND", simply specify:
2358 my $sql = SQL::Abstract->new(logic => 'and');
2360 Which will change the above C<WHERE> to:
2362 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
2364 The logic can also be changed locally by inserting
2365 a modifier in front of an arrayref:
2367 @where = (-and => [event_date => {'>=', '2/13/99'},
2368 event_date => {'<=', '4/24/03'} ]);
2370 See the L</"WHERE CLAUSES"> section for explanations.
2374 This will automatically convert comparisons using the specified SQL
2375 function for both column and value. This is mostly used with an argument
2376 of C<upper> or C<lower>, so that the SQL will have the effect of
2377 case-insensitive "searches". For example, this:
2379 $sql = SQL::Abstract->new(convert => 'upper');
2380 %where = (keywords => 'MaKe iT CAse inSeNSItive');
2382 Will turn out the following SQL:
2384 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
2386 The conversion can be C<upper()>, C<lower()>, or any other SQL function
2387 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
2388 not validate this option; it will just pass through what you specify verbatim).
2392 This is a kludge because many databases suck. For example, you can't
2393 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
2394 Instead, you have to use C<bind_param()>:
2396 $sth->bind_param(1, 'reg data');
2397 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
2399 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
2400 which loses track of which field each slot refers to. Fear not.
2402 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
2403 Currently, you can specify either C<normal> (default) or C<columns>. If you
2404 specify C<columns>, you will get an array that looks like this:
2406 my $sql = SQL::Abstract->new(bindtype => 'columns');
2407 my($stmt, @bind) = $sql->insert(...);
2410 [ 'column1', 'value1' ],
2411 [ 'column2', 'value2' ],
2412 [ 'column3', 'value3' ],
2415 You can then iterate through this manually, using DBI's C<bind_param()>.
2417 $sth->prepare($stmt);
2420 my($col, $data) = @$_;
2421 if ($col eq 'details' || $col eq 'comments') {
2422 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
2423 } elsif ($col eq 'image') {
2424 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
2426 $sth->bind_param($i, $data);
2430 $sth->execute; # execute without @bind now
2432 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
2433 Basically, the advantage is still that you don't have to care which fields
2434 are or are not included. You could wrap that above C<for> loop in a simple
2435 sub called C<bind_fields()> or something and reuse it repeatedly. You still
2436 get a layer of abstraction over manual SQL specification.
2438 Note that if you set L</bindtype> to C<columns>, the C<\[ $sql, @bind ]>
2439 construct (see L</Literal SQL with placeholders and bind values (subqueries)>)
2440 will expect the bind values in this format.
2444 This is the character that a table or column name will be quoted
2445 with. By default this is an empty string, but you could set it to
2446 the character C<`>, to generate SQL like this:
2448 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
2450 Alternatively, you can supply an array ref of two items, the first being the left
2451 hand quote character, and the second the right hand quote character. For
2452 example, you could supply C<['[',']']> for SQL Server 2000 compliant quotes
2453 that generates SQL like this:
2455 SELECT [a_field] FROM [a_table] WHERE [some_field] LIKE '%someval%'
2457 Quoting is useful if you have tables or columns names that are reserved
2458 words in your database's SQL dialect.
2462 This is the character that will be used to escape L</quote_char>s appearing
2463 in an identifier before it has been quoted.
2465 The parameter default in case of a single L</quote_char> character is the quote
2468 When opening-closing-style quoting is used (L</quote_char> is an arrayref)
2469 this parameter defaults to the B<closing (right)> L</quote_char>. Occurrences
2470 of the B<opening (left)> L</quote_char> within the identifier are currently left
2471 untouched. The default for opening-closing-style quotes may change in future
2472 versions, thus you are B<strongly encouraged> to specify the escape character
2477 This is the character that separates a table and column name. It is
2478 necessary to specify this when the C<quote_char> option is selected,
2479 so that tables and column names can be individually quoted like this:
2481 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
2483 =item injection_guard
2485 A regular expression C<qr/.../> that is applied to any C<-function> and unquoted
2486 column name specified in a query structure. This is a safety mechanism to avoid
2487 injection attacks when mishandling user input e.g.:
2489 my %condition_as_column_value_pairs = get_values_from_user();
2490 $sqla->select( ... , \%condition_as_column_value_pairs );
2492 If the expression matches an exception is thrown. Note that literal SQL
2493 supplied via C<\'...'> or C<\['...']> is B<not> checked in any way.
2495 Defaults to checking for C<;> and the C<GO> keyword (TransactSQL)
2497 =item array_datatypes
2499 When this option is true, arrayrefs in INSERT or UPDATE are
2500 interpreted as array datatypes and are passed directly
2502 When this option is false, arrayrefs are interpreted
2503 as literal SQL, just like refs to arrayrefs
2504 (but this behavior is for backwards compatibility; when writing
2505 new queries, use the "reference to arrayref" syntax
2511 Takes a reference to a list of "special operators"
2512 to extend the syntax understood by L<SQL::Abstract>.
2513 See section L</"SPECIAL OPERATORS"> for details.
2517 Takes a reference to a list of "unary operators"
2518 to extend the syntax understood by L<SQL::Abstract>.
2519 See section L</"UNARY OPERATORS"> for details.
2525 =head2 insert($table, \@values || \%fieldvals, \%options)
2527 This is the simplest function. You simply give it a table name
2528 and either an arrayref of values or hashref of field/value pairs.
2529 It returns an SQL INSERT statement and a list of bind values.
2530 See the sections on L</"Inserting and Updating Arrays"> and
2531 L</"Inserting and Updating SQL"> for information on how to insert
2532 with those data types.
2534 The optional C<\%options> hash reference may contain additional
2535 options to generate the insert SQL. Currently supported options
2542 Takes either a scalar of raw SQL fields, or an array reference of
2543 field names, and adds on an SQL C<RETURNING> statement at the end.
2544 This allows you to return data generated by the insert statement
2545 (such as row IDs) without performing another C<SELECT> statement.
2546 Note, however, this is not part of the SQL standard and may not
2547 be supported by all database engines.
2551 =head2 update($table, \%fieldvals, \%where, \%options)
2553 This takes a table, hashref of field/value pairs, and an optional
2554 hashref L<WHERE clause|/WHERE CLAUSES>. It returns an SQL UPDATE function and a list
2556 See the sections on L</"Inserting and Updating Arrays"> and
2557 L</"Inserting and Updating SQL"> for information on how to insert
2558 with those data types.
2560 The optional C<\%options> hash reference may contain additional
2561 options to generate the update SQL. Currently supported options
2568 See the C<returning> option to
2569 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2573 =head2 select($source, $fields, $where, $order)
2575 This returns a SQL SELECT statement and associated list of bind values, as
2576 specified by the arguments:
2582 Specification of the 'FROM' part of the statement.
2583 The argument can be either a plain scalar (interpreted as a table
2584 name, will be quoted), or an arrayref (interpreted as a list
2585 of table names, joined by commas, quoted), or a scalarref
2586 (literal SQL, not quoted).
2590 Specification of the list of fields to retrieve from
2592 The argument can be either an arrayref (interpreted as a list
2593 of field names, will be joined by commas and quoted), or a
2594 plain scalar (literal SQL, not quoted).
2595 Please observe that this API is not as flexible as that of
2596 the first argument C<$source>, for backwards compatibility reasons.
2600 Optional argument to specify the WHERE part of the query.
2601 The argument is most often a hashref, but can also be
2602 an arrayref or plain scalar --
2603 see section L<WHERE clause|/"WHERE CLAUSES"> for details.
2607 Optional argument to specify the ORDER BY part of the query.
2608 The argument can be a scalar, a hashref or an arrayref
2609 -- see section L<ORDER BY clause|/"ORDER BY CLAUSES">
2615 =head2 delete($table, \%where, \%options)
2617 This takes a table name and optional hashref L<WHERE clause|/WHERE CLAUSES>.
2618 It returns an SQL DELETE statement and list of bind values.
2620 The optional C<\%options> hash reference may contain additional
2621 options to generate the delete SQL. Currently supported options
2628 See the C<returning> option to
2629 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2633 =head2 where(\%where, $order)
2635 This is used to generate just the WHERE clause. For example,
2636 if you have an arbitrary data structure and know what the
2637 rest of your SQL is going to look like, but want an easy way
2638 to produce a WHERE clause, use this. It returns an SQL WHERE
2639 clause and list of bind values.
2642 =head2 values(\%data)
2644 This just returns the values from the hash C<%data>, in the same
2645 order that would be returned from any of the other above queries.
2646 Using this allows you to markedly speed up your queries if you
2647 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
2649 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
2651 Warning: This is an experimental method and subject to change.
2653 This returns arbitrarily generated SQL. It's a really basic shortcut.
2654 It will return two different things, depending on return context:
2656 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
2657 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
2659 These would return the following:
2661 # First calling form
2662 $stmt = "CREATE TABLE test (?, ?)";
2663 @bind = (field1, field2);
2665 # Second calling form
2666 $stmt_and_val = "CREATE TABLE test (field1, field2)";
2668 Depending on what you're trying to do, it's up to you to choose the correct
2669 format. In this example, the second form is what you would want.
2673 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
2677 ALTER SESSION SET nls_date_format = 'MM/YY'
2679 You get the idea. Strings get their case twiddled, but everything
2680 else remains verbatim.
2682 =head1 EXPORTABLE FUNCTIONS
2684 =head2 is_plain_value
2686 Determines if the supplied argument is a plain value as understood by this
2691 =item * The value is C<undef>
2693 =item * The value is a non-reference
2695 =item * The value is an object with stringification overloading
2697 =item * The value is of the form C<< { -value => $anything } >>
2701 On failure returns C<undef>, on success returns a B<scalar> reference
2702 to the original supplied argument.
2708 The stringification overloading detection is rather advanced: it takes
2709 into consideration not only the presence of a C<""> overload, but if that
2710 fails also checks for enabled
2711 L<autogenerated versions of C<"">|overload/Magic Autogeneration>, based
2712 on either C<0+> or C<bool>.
2714 Unfortunately testing in the field indicates that this
2715 detection B<< may tickle a latent bug in perl versions before 5.018 >>,
2716 but only when very large numbers of stringifying objects are involved.
2717 At the time of writing ( Sep 2014 ) there is no clear explanation of
2718 the direct cause, nor is there a manageably small test case that reliably
2719 reproduces the problem.
2721 If you encounter any of the following exceptions in B<random places within
2722 your application stack> - this module may be to blame:
2724 Operation "ne": no method found,
2725 left argument in overloaded package <something>,
2726 right argument in overloaded package <something>
2730 Stub found while resolving method "???" overloading """" in package <something>
2732 If you fall victim to the above - please attempt to reduce the problem
2733 to something that could be sent to the L<SQL::Abstract developers
2734 |DBIx::Class/GETTING HELP/SUPPORT>
2735 (either publicly or privately). As a workaround in the meantime you can
2736 set C<$ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}> to a true
2737 value, which will most likely eliminate your problem (at the expense of
2738 not being able to properly detect exotic forms of stringification).
2740 This notice and environment variable will be removed in a future version,
2741 as soon as the underlying problem is found and a reliable workaround is
2746 =head2 is_literal_value
2748 Determines if the supplied argument is a literal value as understood by this
2753 =item * C<\$sql_string>
2755 =item * C<\[ $sql_string, @bind_values ]>
2759 On failure returns C<undef>, on success returns an B<array> reference
2760 containing the unpacked version of the supplied literal SQL and bind values.
2762 =head2 is_undef_value
2764 Tests for undef, whether expanded or not.
2766 =head1 WHERE CLAUSES
2770 This module uses a variation on the idea from L<DBIx::Abstract>. It
2771 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
2772 module is that things in arrays are OR'ed, and things in hashes
2775 The easiest way to explain is to show lots of examples. After
2776 each C<%where> hash shown, it is assumed you used:
2778 my($stmt, @bind) = $sql->where(\%where);
2780 However, note that the C<%where> hash can be used directly in any
2781 of the other functions as well, as described above.
2783 =head2 Key-value pairs
2785 So, let's get started. To begin, a simple hash:
2789 status => 'completed'
2792 Is converted to SQL C<key = val> statements:
2794 $stmt = "WHERE user = ? AND status = ?";
2795 @bind = ('nwiger', 'completed');
2797 One common thing I end up doing is having a list of values that
2798 a field can be in. To do this, simply specify a list inside of
2803 status => ['assigned', 'in-progress', 'pending'];
2806 This simple code will create the following:
2808 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
2809 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
2811 A field associated to an empty arrayref will be considered a
2812 logical false and will generate 0=1.
2814 =head2 Tests for NULL values
2816 If the value part is C<undef> then this is converted to SQL <IS NULL>
2825 $stmt = "WHERE user = ? AND status IS NULL";
2828 To test if a column IS NOT NULL:
2832 status => { '!=', undef },
2835 =head2 Specific comparison operators
2837 If you want to specify a different type of operator for your comparison,
2838 you can use a hashref for a given column:
2842 status => { '!=', 'completed' }
2845 Which would generate:
2847 $stmt = "WHERE user = ? AND status != ?";
2848 @bind = ('nwiger', 'completed');
2850 To test against multiple values, just enclose the values in an arrayref:
2852 status => { '=', ['assigned', 'in-progress', 'pending'] };
2854 Which would give you:
2856 "WHERE status = ? OR status = ? OR status = ?"
2859 The hashref can also contain multiple pairs, in which case it is expanded
2860 into an C<AND> of its elements:
2864 status => { '!=', 'completed', -not_like => 'pending%' }
2867 # Or more dynamically, like from a form
2868 $where{user} = 'nwiger';
2869 $where{status}{'!='} = 'completed';
2870 $where{status}{'-not_like'} = 'pending%';
2872 # Both generate this
2873 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
2874 @bind = ('nwiger', 'completed', 'pending%');
2877 To get an OR instead, you can combine it with the arrayref idea:
2881 priority => [ { '=', 2 }, { '>', 5 } ]
2884 Which would generate:
2886 $stmt = "WHERE ( priority = ? OR priority > ? ) AND user = ?";
2887 @bind = ('2', '5', 'nwiger');
2889 If you want to include literal SQL (with or without bind values), just use a
2890 scalar reference or reference to an arrayref as the value:
2893 date_entered => { '>' => \["to_date(?, 'MM/DD/YYYY')", "11/26/2008"] },
2894 date_expires => { '<' => \"now()" }
2897 Which would generate:
2899 $stmt = "WHERE date_entered > to_date(?, 'MM/DD/YYYY') AND date_expires < now()";
2900 @bind = ('11/26/2008');
2903 =head2 Logic and nesting operators
2905 In the example above,
2906 there is a subtle trap if you want to say something like
2907 this (notice the C<AND>):
2909 WHERE priority != ? AND priority != ?
2911 Because, in Perl you I<can't> do this:
2913 priority => { '!=' => 2, '!=' => 1 }
2915 As the second C<!=> key will obliterate the first. The solution
2916 is to use the special C<-modifier> form inside an arrayref:
2918 priority => [ -and => {'!=', 2},
2922 Normally, these would be joined by C<OR>, but the modifier tells it
2923 to use C<AND> instead. (Hint: You can use this in conjunction with the
2924 C<logic> option to C<new()> in order to change the way your queries
2925 work by default.) B<Important:> Note that the C<-modifier> goes
2926 B<INSIDE> the arrayref, as an extra first element. This will
2927 B<NOT> do what you think it might:
2929 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
2931 Here is a quick list of equivalencies, since there is some overlap:
2934 status => {'!=', 'completed', 'not like', 'pending%' }
2935 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
2938 status => {'=', ['assigned', 'in-progress']}
2939 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
2940 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
2944 =head2 Special operators: IN, BETWEEN, etc.
2946 You can also use the hashref format to compare a list of fields using the
2947 C<IN> comparison operator, by specifying the list as an arrayref:
2950 status => 'completed',
2951 reportid => { -in => [567, 2335, 2] }
2954 Which would generate:
2956 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
2957 @bind = ('completed', '567', '2335', '2');
2959 The reverse operator C<-not_in> generates SQL C<NOT IN> and is used in
2962 If the argument to C<-in> is an empty array, 'sqlfalse' is generated
2963 (by default: C<1=0>). Similarly, C<< -not_in => [] >> generates
2964 'sqltrue' (by default: C<1=1>).
2966 In addition to the array you can supply a chunk of literal sql or
2967 literal sql with bind:
2970 customer => { -in => \[
2971 'SELECT cust_id FROM cust WHERE balance > ?',
2974 status => { -in => \'SELECT status_codes FROM states' },
2980 customer IN ( SELECT cust_id FROM cust WHERE balance > ? )
2981 AND status IN ( SELECT status_codes FROM states )
2985 Finally, if the argument to C<-in> is not a reference, it will be
2986 treated as a single-element array.
2988 Another pair of operators is C<-between> and C<-not_between>,
2989 used with an arrayref of two values:
2993 completion_date => {
2994 -not_between => ['2002-10-01', '2003-02-06']
3000 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
3002 Just like with C<-in> all plausible combinations of literal SQL
3006 start0 => { -between => [ 1, 2 ] },
3007 start1 => { -between => \["? AND ?", 1, 2] },
3008 start2 => { -between => \"lower(x) AND upper(y)" },
3009 start3 => { -between => [
3011 \["upper(?)", 'stuff' ],
3018 ( start0 BETWEEN ? AND ? )
3019 AND ( start1 BETWEEN ? AND ? )
3020 AND ( start2 BETWEEN lower(x) AND upper(y) )
3021 AND ( start3 BETWEEN lower(x) AND upper(?) )
3023 @bind = (1, 2, 1, 2, 'stuff');
3026 These are the two builtin "special operators"; but the
3027 list can be expanded: see section L</"SPECIAL OPERATORS"> below.
3029 =head2 Unary operators: bool
3031 If you wish to test against boolean columns or functions within your
3032 database you can use the C<-bool> and C<-not_bool> operators. For
3033 example to test the column C<is_user> being true and the column
3034 C<is_enabled> being false you would use:-
3038 -not_bool => 'is_enabled',
3043 WHERE is_user AND NOT is_enabled
3045 If a more complex combination is required, testing more conditions,
3046 then you should use the and/or operators:-
3051 -not_bool => { two=> { -rlike => 'bar' } },
3052 -not_bool => { three => [ { '=', 2 }, { '>', 5 } ] },
3063 (NOT ( three = ? OR three > ? ))
3066 =head2 Nested conditions, -and/-or prefixes
3068 So far, we've seen how multiple conditions are joined with a top-level
3069 C<AND>. We can change this by putting the different conditions we want in
3070 hashes and then putting those hashes in an array. For example:
3075 status => { -like => ['pending%', 'dispatched'] },
3079 status => 'unassigned',
3083 This data structure would create the following:
3085 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
3086 OR ( user = ? AND status = ? ) )";
3087 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
3090 Clauses in hashrefs or arrayrefs can be prefixed with an C<-and> or C<-or>
3091 to change the logic inside:
3097 -and => [ workhrs => {'>', 20}, geo => 'ASIA' ],
3098 -or => { workhrs => {'<', 50}, geo => 'EURO' },
3105 $stmt = "WHERE ( user = ?
3106 AND ( ( workhrs > ? AND geo = ? )
3107 OR ( workhrs < ? OR geo = ? ) ) )";
3108 @bind = ('nwiger', '20', 'ASIA', '50', 'EURO');
3110 =head3 Algebraic inconsistency, for historical reasons
3112 C<Important note>: when connecting several conditions, the C<-and->|C<-or>
3113 operator goes C<outside> of the nested structure; whereas when connecting
3114 several constraints on one column, the C<-and> operator goes
3115 C<inside> the arrayref. Here is an example combining both features:
3118 -and => [a => 1, b => 2],
3119 -or => [c => 3, d => 4],
3120 e => [-and => {-like => 'foo%'}, {-like => '%bar'} ]
3125 WHERE ( ( ( a = ? AND b = ? )
3126 OR ( c = ? OR d = ? )
3127 OR ( e LIKE ? AND e LIKE ? ) ) )
3129 This difference in syntax is unfortunate but must be preserved for
3130 historical reasons. So be careful: the two examples below would
3131 seem algebraically equivalent, but they are not
3134 { -like => 'foo%' },
3135 { -like => '%bar' },
3137 # yields: WHERE ( ( col LIKE ? AND col LIKE ? ) )
3140 { col => { -like => 'foo%' } },
3141 { col => { -like => '%bar' } },
3143 # yields: WHERE ( ( col LIKE ? OR col LIKE ? ) )
3146 =head2 Literal SQL and value type operators
3148 The basic premise of SQL::Abstract is that in WHERE specifications the "left
3149 side" is a column name and the "right side" is a value (normally rendered as
3150 a placeholder). This holds true for both hashrefs and arrayref pairs as you
3151 see in the L</WHERE CLAUSES> examples above. Sometimes it is necessary to
3152 alter this behavior. There are several ways of doing so.
3156 This is a virtual operator that signals the string to its right side is an
3157 identifier (a column name) and not a value. For example to compare two
3158 columns you would write:
3161 priority => { '<', 2 },
3162 requestor => { -ident => 'submitter' },
3167 $stmt = "WHERE priority < ? AND requestor = submitter";
3170 If you are maintaining legacy code you may see a different construct as
3171 described in L</Deprecated usage of Literal SQL>, please use C<-ident> in new
3176 This is a virtual operator that signals that the construct to its right side
3177 is a value to be passed to DBI. This is for example necessary when you want
3178 to write a where clause against an array (for RDBMS that support such
3179 datatypes). For example:
3182 array => { -value => [1, 2, 3] }
3187 $stmt = 'WHERE array = ?';
3188 @bind = ([1, 2, 3]);
3190 Note that if you were to simply say:
3196 the result would probably not be what you wanted:
3198 $stmt = 'WHERE array = ? OR array = ? OR array = ?';
3203 Finally, sometimes only literal SQL will do. To include a random snippet
3204 of SQL verbatim, you specify it as a scalar reference. Consider this only
3205 as a last resort. Usually there is a better way. For example:
3208 priority => { '<', 2 },
3209 requestor => { -in => \'(SELECT name FROM hitmen)' },
3214 $stmt = "WHERE priority < ? AND requestor IN (SELECT name FROM hitmen)"
3217 Note that in this example, you only get one bind parameter back, since
3218 the verbatim SQL is passed as part of the statement.
3222 Never use untrusted input as a literal SQL argument - this is a massive
3223 security risk (there is no way to check literal snippets for SQL
3224 injections and other nastyness). If you need to deal with untrusted input
3225 use literal SQL with placeholders as described next.
3227 =head3 Literal SQL with placeholders and bind values (subqueries)
3229 If the literal SQL to be inserted has placeholders and bind values,
3230 use a reference to an arrayref (yes this is a double reference --
3231 not so common, but perfectly legal Perl). For example, to find a date
3232 in Postgres you can use something like this:
3235 date_column => \[ "= date '2008-09-30' - ?::integer", 10 ]
3240 $stmt = "WHERE ( date_column = date '2008-09-30' - ?::integer )"
3243 Note that you must pass the bind values in the same format as they are returned
3244 by L<where|/where(\%where, $order)>. This means that if you set L</bindtype>
3245 to C<columns>, you must provide the bind values in the
3246 C<< [ column_meta => value ] >> format, where C<column_meta> is an opaque
3247 scalar value; most commonly the column name, but you can use any scalar value
3248 (including references and blessed references), L<SQL::Abstract> will simply
3249 pass it through intact. So if C<bindtype> is set to C<columns> the above
3250 example will look like:
3253 date_column => \[ "= date '2008-09-30' - ?::integer", [ {} => 10 ] ]
3256 Literal SQL is especially useful for nesting parenthesized clauses in the
3257 main SQL query. Here is a first example:
3259 my ($sub_stmt, @sub_bind) = ("SELECT c1 FROM t1 WHERE c2 < ? AND c3 LIKE ?",
3263 bar => \["IN ($sub_stmt)" => @sub_bind],
3268 $stmt = "WHERE (foo = ? AND bar IN (SELECT c1 FROM t1
3269 WHERE c2 < ? AND c3 LIKE ?))";
3270 @bind = (1234, 100, "foo%");
3272 Other subquery operators, like for example C<"E<gt> ALL"> or C<"NOT IN">,
3273 are expressed in the same way. Of course the C<$sub_stmt> and
3274 its associated bind values can be generated through a former call
3277 my ($sub_stmt, @sub_bind)
3278 = $sql->select("t1", "c1", {c2 => {"<" => 100},
3279 c3 => {-like => "foo%"}});
3282 bar => \["> ALL ($sub_stmt)" => @sub_bind],
3285 In the examples above, the subquery was used as an operator on a column;
3286 but the same principle also applies for a clause within the main C<%where>
3287 hash, like an EXISTS subquery:
3289 my ($sub_stmt, @sub_bind)
3290 = $sql->select("t1", "*", {c1 => 1, c2 => \"> t0.c0"});
3291 my %where = ( -and => [
3293 \["EXISTS ($sub_stmt)" => @sub_bind],
3298 $stmt = "WHERE (foo = ? AND EXISTS (SELECT * FROM t1
3299 WHERE c1 = ? AND c2 > t0.c0))";
3303 Observe that the condition on C<c2> in the subquery refers to
3304 column C<t0.c0> of the main query: this is I<not> a bind
3305 value, so we have to express it through a scalar ref.
3306 Writing C<< c2 => {">" => "t0.c0"} >> would have generated
3307 C<< c2 > ? >> with bind value C<"t0.c0"> ... not exactly
3308 what we wanted here.
3310 Finally, here is an example where a subquery is used
3311 for expressing unary negation:
3313 my ($sub_stmt, @sub_bind)
3314 = $sql->where({age => [{"<" => 10}, {">" => 20}]});
3315 $sub_stmt =~ s/^ where //i; # don't want "WHERE" in the subclause
3317 lname => {like => '%son%'},
3318 \["NOT ($sub_stmt)" => @sub_bind],
3323 $stmt = "lname LIKE ? AND NOT ( age < ? OR age > ? )"
3324 @bind = ('%son%', 10, 20)
3326 =head3 Deprecated usage of Literal SQL
3328 Below are some examples of archaic use of literal SQL. It is shown only as
3329 reference for those who deal with legacy code. Each example has a much
3330 better, cleaner and safer alternative that users should opt for in new code.
3336 my %where = ( requestor => \'IS NOT NULL' )
3338 $stmt = "WHERE requestor IS NOT NULL"
3340 This used to be the way of generating NULL comparisons, before the handling
3341 of C<undef> got formalized. For new code please use the superior syntax as
3342 described in L</Tests for NULL values>.
3346 my %where = ( requestor => \'= submitter' )
3348 $stmt = "WHERE requestor = submitter"
3350 This used to be the only way to compare columns. Use the superior L</-ident>
3351 method for all new code. For example an identifier declared in such a way
3352 will be properly quoted if L</quote_char> is properly set, while the legacy
3353 form will remain as supplied.
3357 my %where = ( is_ready => \"", completed => { '>', '2012-12-21' } )
3359 $stmt = "WHERE completed > ? AND is_ready"
3360 @bind = ('2012-12-21')
3362 Using an empty string literal used to be the only way to express a boolean.
3363 For all new code please use the much more readable
3364 L<-bool|/Unary operators: bool> operator.
3370 These pages could go on for a while, since the nesting of the data
3371 structures this module can handle are pretty much unlimited (the
3372 module implements the C<WHERE> expansion as a recursive function
3373 internally). Your best bet is to "play around" with the module a
3374 little to see how the data structures behave, and choose the best
3375 format for your data based on that.
3377 And of course, all the values above will probably be replaced with
3378 variables gotten from forms or the command line. After all, if you
3379 knew everything ahead of time, you wouldn't have to worry about
3380 dynamically-generating SQL and could just hardwire it into your
3383 =head1 ORDER BY CLAUSES
3385 Some functions take an order by clause. This can either be a scalar (just a
3386 column name), a hashref of C<< { -desc => 'col' } >> or C<< { -asc => 'col' }
3387 >>, a scalarref, an arrayref-ref, or an arrayref of any of the previous
3390 Given | Will Generate
3391 ---------------------------------------------------------------
3393 'colA' | ORDER BY colA
3395 [qw/colA colB/] | ORDER BY colA, colB
3397 {-asc => 'colA'} | ORDER BY colA ASC
3399 {-desc => 'colB'} | ORDER BY colB DESC
3401 ['colA', {-asc => 'colB'}] | ORDER BY colA, colB ASC
3403 { -asc => [qw/colA colB/] } | ORDER BY colA ASC, colB ASC
3405 \'colA DESC' | ORDER BY colA DESC
3407 \[ 'FUNC(colA, ?)', $x ] | ORDER BY FUNC(colA, ?)
3408 | /* ...with $x bound to ? */
3411 { -asc => 'colA' }, | colA ASC,
3412 { -desc => [qw/colB/] }, | colB DESC,
3413 { -asc => [qw/colC colD/] },| colC ASC, colD ASC,
3414 \'colE DESC', | colE DESC,
3415 \[ 'FUNC(colF, ?)', $x ], | FUNC(colF, ?)
3416 ] | /* ...with $x bound to ? */
3417 ===============================================================
3421 =head1 OLD EXTENSION SYSTEM
3423 =head2 SPECIAL OPERATORS
3425 my $sqlmaker = SQL::Abstract->new(special_ops => [
3429 my ($self, $field, $op, $arg) = @_;
3435 handler => 'method_name',
3439 A "special operator" is a SQL syntactic clause that can be
3440 applied to a field, instead of a usual binary operator.
3443 WHERE field IN (?, ?, ?)
3444 WHERE field BETWEEN ? AND ?
3445 WHERE MATCH(field) AGAINST (?, ?)
3447 Special operators IN and BETWEEN are fairly standard and therefore
3448 are builtin within C<SQL::Abstract> (as the overridable methods
3449 C<_where_field_IN> and C<_where_field_BETWEEN>). For other operators,
3450 like the MATCH .. AGAINST example above which is specific to MySQL,
3451 you can write your own operator handlers - supply a C<special_ops>
3452 argument to the C<new> method. That argument takes an arrayref of
3453 operator definitions; each operator definition is a hashref with two
3460 the regular expression to match the operator
3464 Either a coderef or a plain scalar method name. In both cases
3465 the expected return is C<< ($sql, @bind) >>.
3467 When supplied with a method name, it is simply called on the
3468 L<SQL::Abstract> object as:
3470 $self->$method_name($field, $op, $arg)
3474 $field is the LHS of the operator
3475 $op is the part that matched the handler regex
3478 When supplied with a coderef, it is called as:
3480 $coderef->($self, $field, $op, $arg)
3485 For example, here is an implementation
3486 of the MATCH .. AGAINST syntax for MySQL
3488 my $sqlmaker = SQL::Abstract->new(special_ops => [
3490 # special op for MySql MATCH (field) AGAINST(word1, word2, ...)
3491 {regex => qr/^match$/i,
3493 my ($self, $field, $op, $arg) = @_;
3494 $arg = [$arg] if not ref $arg;
3495 my $label = $self->_quote($field);
3496 my ($placeholder) = $self->_convert('?');
3497 my $placeholders = join ", ", (($placeholder) x @$arg);
3498 my $sql = $self->_sqlcase('match') . " ($label) "
3499 . $self->_sqlcase('against') . " ($placeholders) ";
3500 my @bind = $self->_bindtype($field, @$arg);
3501 return ($sql, @bind);
3508 =head2 UNARY OPERATORS
3510 my $sqlmaker = SQL::Abstract->new(unary_ops => [
3514 my ($self, $op, $arg) = @_;
3520 handler => 'method_name',
3524 A "unary operator" is a SQL syntactic clause that can be
3525 applied to a field - the operator goes before the field
3527 You can write your own operator handlers - supply a C<unary_ops>
3528 argument to the C<new> method. That argument takes an arrayref of
3529 operator definitions; each operator definition is a hashref with two
3536 the regular expression to match the operator
3540 Either a coderef or a plain scalar method name. In both cases
3541 the expected return is C<< $sql >>.
3543 When supplied with a method name, it is simply called on the
3544 L<SQL::Abstract> object as:
3546 $self->$method_name($op, $arg)
3550 $op is the part that matched the handler regex
3551 $arg is the RHS or argument of the operator
3553 When supplied with a coderef, it is called as:
3555 $coderef->($self, $op, $arg)
3560 =head1 NEW METHODS (EXPERIMENTAL)
3562 See L<SQL::Abstract::Reference> for the C<expr> versus C<aqt> concept and
3563 an explanation of what the below extensions are extending.
3567 $sqla->plugin('+Foo');
3569 Enables plugin SQL::Abstract::Plugin::Foo.
3573 my ($sql, @bind) = $sqla->render_expr($expr);
3575 =head2 render_statement
3577 Use this if you may be rendering a top level statement so e.g. a SELECT
3578 query doesn't get wrapped in parens
3580 my ($sql, @bind) = $sqla->render_statement($expr);
3584 Expression expansion with optional default for scalars.
3586 my $aqt = $self->expand_expr($expr);
3587 my $aqt = $self->expand_expr($expr, -ident);
3591 Top level means avoid parens on statement AQT.
3593 my $res = $self->render_aqt($aqt, $top_level);
3594 my ($sql, @bind) = @$res;
3596 =head2 join_query_parts
3598 Similar to join() but will render hashrefs as nodes for both join and parts,
3599 and treats arrayref as a nested C<[ $join, @parts ]> structure.
3601 my $part = $self->join_query_parts($join, @parts);
3603 =head1 NEW EXTENSION SYSTEM
3607 my $sqla2 = $sqla->clone;
3609 Performs a semi-shallow copy such that extension methods won't leak state
3610 but excessive depth is avoided.
3620 =head2 clause_expander
3622 =head2 clause_expanders
3624 $sqla->expander('name' => sub { ... });
3625 $sqla->expanders('name1' => sub { ... }, 'name2' => sub { ... });
3627 =head2 expander_list
3629 =head2 op_expander_list
3631 =head2 clause_expander_list
3633 my @names = $sqla->expander_list;
3635 =head2 wrap_expander
3637 =head2 wrap_expanders
3639 =head2 wrap_op_expander
3641 =head2 wrap_op_expanders
3643 =head2 wrap_clause_expander
3645 =head2 wrap_clause_expanders
3647 $sqla->wrap_expander('name' => sub { my ($orig) = @_; sub { ... } });
3648 $sqla->wrap_expanders(
3649 'name1' => sub { my ($orig1) = @_; sub { ... } },
3650 'name2' => sub { my ($orig2) = @_; sub { ... } },
3661 =head2 clause_renderer
3663 =head2 clause_renderers
3665 $sqla->renderer('name' => sub { ... });
3666 $sqla->renderers('name1' => sub { ... }, 'name2' => sub { ... });
3668 =head2 renderer_list
3670 =head2 op_renderer_list
3672 =head2 clause_renderer_list
3674 my @names = $sqla->renderer_list;
3676 =head2 wrap_renderer
3678 =head2 wrap_renderers
3680 =head2 wrap_op_renderer
3682 =head2 wrap_op_renderers
3684 =head2 wrap_clause_renderer
3686 =head2 wrap_clause_renderers
3688 $sqla->wrap_renderer('name' => sub { my ($orig) = @_; sub { ... } });
3689 $sqla->wrap_renderers(
3690 'name1' => sub { my ($orig1) = @_; sub { ... } },
3691 'name2' => sub { my ($orig2) = @_; sub { ... } },
3696 my @clauses = $sqla->clauses_of('select');
3697 $sqla->clauses_of(select => \@new_clauses);
3698 $sqla->clauses_of(select => sub {
3699 my (undef, @old_clauses) = @_;
3701 return @new_clauses;
3704 =head2 statement_list
3706 my @list = $sqla->statement_list;
3708 =head2 make_unop_expander
3710 my $exp = $sqla->make_unop_expander(sub { ... });
3712 If the op is found as a binop, assumes it wants a default comparison, so
3713 the inner expander sub can reliably operate as
3715 sub { my ($self, $name, $body) = @_; ... }
3717 =head2 make_binop_expander
3719 my $exp = $sqla->make_binop_expander(sub { ... });
3721 If the op is found as a unop, assumes the value will be an arrayref with the
3722 LHS as the first entry, and converts that to an ident node if it's a simple
3723 scalar. So the inner expander sub looks like
3726 my ($self, $name, $body, $k) = @_;
3727 { -blah => [ map $self->expand_expr($_), $k, $body ] }
3730 =head2 unop_expander
3732 =head2 unop_expanders
3734 =head2 binop_expander
3736 =head2 binop_expanders
3738 The above methods operate exactly like the op_ versions but wrap the coderef
3739 using the appropriate make_ method first.
3743 Thanks to some benchmarking by Mark Stosberg, it turns out that
3744 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
3745 I must admit this wasn't an intentional design issue, but it's a
3746 byproduct of the fact that you get to control your C<DBI> handles
3749 To maximize performance, use a code snippet like the following:
3751 # prepare a statement handle using the first row
3752 # and then reuse it for the rest of the rows
3754 for my $href (@array_of_hashrefs) {
3755 $stmt ||= $sql->insert('table', $href);
3756 $sth ||= $dbh->prepare($stmt);
3757 $sth->execute($sql->values($href));
3760 The reason this works is because the keys in your C<$href> are sorted
3761 internally by B<SQL::Abstract>. Thus, as long as your data retains
3762 the same structure, you only have to generate the SQL the first time
3763 around. On subsequent queries, simply use the C<values> function provided
3764 by this module to return your values in the correct order.
3766 However this depends on the values having the same type - if, for
3767 example, the values of a where clause may either have values
3768 (resulting in sql of the form C<column = ?> with a single bind
3769 value), or alternatively the values might be C<undef> (resulting in
3770 sql of the form C<column IS NULL> with no bind value) then the
3771 caching technique suggested will not work.
3775 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
3776 really like this part (I do, at least). Building up a complex query
3777 can be as simple as the following:
3784 use CGI::FormBuilder;
3787 my $form = CGI::FormBuilder->new(...);
3788 my $sql = SQL::Abstract->new;
3790 if ($form->submitted) {
3791 my $field = $form->field;
3792 my $id = delete $field->{id};
3793 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
3796 Of course, you would still have to connect using C<DBI> to run the
3797 query, but the point is that if you make your form look like your
3798 table, the actual query script can be extremely simplistic.
3800 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
3801 a fast interface to returning and formatting data. I frequently
3802 use these three modules together to write complex database query
3803 apps in under 50 lines.
3805 =head1 HOW TO CONTRIBUTE
3807 Contributions are always welcome, in all usable forms (we especially
3808 welcome documentation improvements). The delivery methods include git-
3809 or unified-diff formatted patches, GitHub pull requests, or plain bug
3810 reports either via RT or the Mailing list. Contributors are generally
3811 granted full access to the official repository after their first several
3812 patches pass successful review.
3814 This project is maintained in a git repository. The code and related tools are
3815 accessible at the following locations:
3819 =item * Official repo: L<git://git.shadowcat.co.uk/dbsrgits/SQL-Abstract.git>
3821 =item * Official gitweb: L<http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=dbsrgits/SQL-Abstract.git>
3823 =item * GitHub mirror: L<https://github.com/dbsrgits/sql-abstract>
3825 =item * Authorized committers: L<ssh://dbsrgits@git.shadowcat.co.uk/SQL-Abstract.git>
3831 Version 1.50 was a major internal refactoring of C<SQL::Abstract>.
3832 Great care has been taken to preserve the I<published> behavior
3833 documented in previous versions in the 1.* family; however,
3834 some features that were previously undocumented, or behaved
3835 differently from the documentation, had to be changed in order
3836 to clarify the semantics. Hence, client code that was relying
3837 on some dark areas of C<SQL::Abstract> v1.*
3838 B<might behave differently> in v1.50.
3840 The main changes are:
3846 support for literal SQL through the C<< \ [ $sql, @bind ] >> syntax.
3850 support for the { operator => \"..." } construct (to embed literal SQL)
3854 support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values)
3858 optional support for L<array datatypes|/"Inserting and Updating Arrays">
3862 defensive programming: check arguments
3866 fixed bug with global logic, which was previously implemented
3867 through global variables yielding side-effects. Prior versions would
3868 interpret C<< [ {cond1, cond2}, [cond3, cond4] ] >>
3869 as C<< "(cond1 AND cond2) OR (cond3 AND cond4)" >>.
3870 Now this is interpreted
3871 as C<< "(cond1 AND cond2) OR (cond3 OR cond4)" >>.
3876 fixed semantics of _bindtype on array args
3880 dropped the C<_anoncopy> of the %where tree. No longer necessary,
3881 we just avoid shifting arrays within that tree.
3885 dropped the C<_modlogic> function
3889 =head1 ACKNOWLEDGEMENTS
3891 There are a number of individuals that have really helped out with
3892 this module. Unfortunately, most of them submitted bugs via CPAN
3893 so I have no idea who they are! But the people I do know are:
3895 Ash Berlin (order_by hash term support)
3896 Matt Trout (DBIx::Class support)
3897 Mark Stosberg (benchmarking)
3898 Chas Owens (initial "IN" operator support)
3899 Philip Collins (per-field SQL functions)
3900 Eric Kolve (hashref "AND" support)
3901 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
3902 Dan Kubb (support for "quote_char" and "name_sep")
3903 Guillermo Roditi (patch to cleanup "IN" and "BETWEEN", fix and tests for _order_by)
3904 Laurent Dami (internal refactoring, extensible list of special operators, literal SQL)
3905 Norbert Buchmuller (support for literal SQL in hashpair, misc. fixes & tests)
3906 Peter Rabbitson (rewrite of SQLA::Test, misc. fixes & tests)
3907 Oliver Charles (support for "RETURNING" after "INSERT")
3913 L<DBIx::Class>, L<DBIx::Abstract>, L<CGI::FormBuilder>, L<HTML::QuickTable>.
3917 Copyright (c) 2001-2007 Nathan Wiger <nwiger@cpan.org>. All Rights Reserved.
3919 This module is actively maintained by Matt Trout <mst@shadowcatsystems.co.uk>
3921 For support, your best bet is to try the C<DBIx::Class> users mailing list.
3922 While not an official support venue, C<DBIx::Class> makes heavy use of
3923 C<SQL::Abstract>, and as such list members there are very familiar with
3924 how to create queries.
3928 This module is free software; you may copy this under the same
3929 terms as perl itself (either the GNU General Public License or
3930 the Artistic License)