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);
21 *SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION = $ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}
27 #======================================================================
29 #======================================================================
31 our $VERSION = '1.86';
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 not => '_expand_not',
143 bool => '_expand_bool',
144 and => '_expand_op_andor',
145 or => '_expand_op_andor',
146 nest => '_expand_nest',
147 bind => '_expand_bind',
149 not_in => '_expand_in',
150 row => '_expand_row',
151 between => '_expand_between',
152 not_between => '_expand_between',
154 (map +($_ => '_expand_op_is'), ('is', 'is_not')),
155 ident => '_expand_ident',
156 value => '_expand_value',
157 func => '_expand_func',
160 'between' => '_expand_between',
161 'not_between' => '_expand_between',
162 'in' => '_expand_in',
163 'not_in' => '_expand_in',
164 'nest' => '_expand_nest',
165 (map +($_ => '_expand_op_andor'), ('and', 'or')),
166 (map +($_ => '_expand_op_is'), ('is', 'is_not')),
167 'ident' => '_expand_ident',
168 'value' => '_expand_value',
171 (map +($_, "_render_$_"), qw(op func bind ident literal row)),
174 (map +($_ => '_render_op_between'), 'between', 'not_between'),
175 (map +($_ => '_render_op_in'), 'in', 'not_in'),
176 (map +($_ => '_render_unop_postfix'),
177 'is_null', 'is_not_null', 'asc', 'desc',
179 (not => '_render_unop_paren'),
180 (map +($_ => '_render_op_andor'), qw(and or)),
181 ',' => '_render_op_multop',
187 my $class = ref($self) || $self;
188 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
190 # choose our case by keeping an option around
191 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
193 # default logic for interpreting arrayrefs
194 $opt{logic} = $opt{logic} ? uc $opt{logic} : 'OR';
196 # how to return bind vars
197 $opt{bindtype} ||= 'normal';
199 # default comparison is "=", but can be overridden
202 # try to recognize which are the 'equality' and 'inequality' ops
203 # (temporary quickfix (in 2007), should go through a more seasoned API)
204 $opt{equality_op} = qr/^( \Q$opt{cmp}\E | \= )$/ix;
205 $opt{inequality_op} = qr/^( != | <> )$/ix;
207 $opt{like_op} = qr/^ (is_)?r?like $/xi;
208 $opt{not_like_op} = qr/^ (is_)?not_r?like $/xi;
211 $opt{sqltrue} ||= '1=1';
212 $opt{sqlfalse} ||= '0=1';
215 $opt{special_ops} ||= [];
217 if ($class->isa('DBIx::Class::SQLMaker')) {
218 $opt{warn_once_on_nest} = 1;
219 $opt{disable_old_special_ops} = 1;
223 $opt{unary_ops} ||= [];
225 # rudimentary sanity-check for user supplied bits treated as functions/operators
226 # If a purported function matches this regular expression, an exception is thrown.
227 # Literal SQL is *NOT* subject to this check, only functions (and column names
228 # when quoting is not in effect)
231 # need to guard against ()'s in column names too, but this will break tons of
232 # hacks... ideas anyone?
233 $opt{injection_guard} ||= qr/
239 $opt{expand_unary} = {};
241 foreach my $name (sort keys %Defaults) {
242 $opt{$name} = { %{$Defaults{$name}} };
245 if ($opt{lazy_join_sql_parts}) {
246 my $mod = Module::Runtime::use_module('SQL::Abstract::Parts');
247 $opt{join_sql_parts} ||= sub { $mod->new(@_) };
250 $opt{join_sql_parts} ||= sub { join $_[0], @_[1..$#_] };
252 return bless \%opt, $class;
255 sub sqltrue { +{ -literal => [ $_[0]->{sqltrue} ] } }
256 sub sqlfalse { +{ -literal => [ $_[0]->{sqlfalse} ] } }
258 sub _assert_pass_injection_guard {
259 if ($_[1] =~ $_[0]->{injection_guard}) {
260 my $class = ref $_[0];
261 puke "Possible SQL injection attempt '$_[1]'. If this is indeed a part of the "
262 . "desired SQL use literal SQL ( \'...' or \[ '...' ] ) or supply your own "
263 . "{injection_guard} attribute to ${class}->new()"
268 #======================================================================
270 #======================================================================
274 my $table = $self->_table(shift);
275 my $data = shift || return;
280 my ($f_aqt, $v_aqt) = $self->_expand_insert_values($data);
282 my @parts = ([ $self->_sqlcase('insert into').' '.$table ]);
283 push @parts, $self->render_aqt($f_aqt) if $f_aqt;
284 push @parts, [ $self->_sqlcase('values') ], $self->render_aqt($v_aqt);
286 if ($options->{returning}) {
287 push @parts, [ $self->_insert_returning($options) ];
290 my ($sql, @bind) = @{ $self->join_query_parts(' ', @parts) };
291 return wantarray ? ($sql, @bind) : $sql;
294 sub _expand_insert_values {
295 my ($self, $data) = @_;
296 if (is_literal_value($data)) {
297 (undef, $self->expand_expr($data));
299 my ($fields, $values) = (
300 ref($data) eq 'HASH' ?
301 ([ sort keys %$data ], [ @{$data}{sort keys %$data} ])
305 # no names (arrayref) means can't generate bindtype
306 !($fields) && $self->{bindtype} eq 'columns'
307 && belch "can't do 'columns' bindtype when called with arrayref";
311 ? $self->expand_expr({ -row => $fields }, -ident)
316 local our $Cur_Col_Meta = $fields->[$_];
317 $self->_expand_insert_value($values->[$_])
324 # So that subclasses can override INSERT ... RETURNING separately from
325 # UPDATE and DELETE (e.g. DBIx::Class::SQLMaker::Oracle does this)
326 sub _insert_returning { shift->_returning(@_) }
329 my ($self, $options) = @_;
331 my $f = $options->{returning};
333 my ($sql, @bind) = @{ $self->render_aqt(
334 $self->_expand_maybe_list_expr($f, -ident)
336 return ($self->_sqlcase(' returning ').$sql, @bind);
339 sub _expand_insert_value {
342 my $k = our $Cur_Col_Meta;
344 if (ref($v) eq 'ARRAY') {
345 if ($self->{array_datatypes}) {
346 return +{ -bind => [ $k, $v ] };
348 my ($sql, @bind) = @$v;
349 $self->_assert_bindval_matches_bindtype(@bind);
350 return +{ -literal => $v };
352 if (ref($v) eq 'HASH') {
353 if (grep !/^-/, keys %$v) {
354 belch "HASH ref as bind value in insert is not supported";
355 return +{ -bind => [ $k, $v ] };
359 return +{ -bind => [ $k, undef ] };
361 return $self->expand_expr($v);
366 #======================================================================
368 #======================================================================
373 my $table = $self->_table(shift);
374 my $data = shift || return;
378 # first build the 'SET' part of the sql statement
379 puke "Unsupported data type specified to \$sql->update"
380 unless ref $data eq 'HASH';
382 my ($sql, @all_bind) = $self->_update_set_values($data);
383 $sql = $self->_sqlcase('update ') . $table . $self->_sqlcase(' set ')
387 my($where_sql, @where_bind) = $self->where($where);
389 push @all_bind, @where_bind;
392 if ($options->{returning}) {
393 my ($returning_sql, @returning_bind) = $self->_update_returning($options);
394 $sql .= $returning_sql;
395 push @all_bind, @returning_bind;
398 return wantarray ? ($sql, @all_bind) : $sql;
401 sub _update_set_values {
402 my ($self, $data) = @_;
404 return @{ $self->render_aqt(
405 $self->_expand_update_set_values(undef, $data),
409 sub _expand_update_set_values {
410 my ($self, undef, $data) = @_;
411 $self->_expand_maybe_list_expr( [
414 $set = { -bind => $_ } unless defined $set;
415 +{ -op => [ '=', { -ident => $k }, $set ] };
421 ? ($self->{array_datatypes}
422 ? [ $k, +{ -bind => [ $k, $v ] } ]
423 : [ $k, +{ -literal => $v } ])
425 local our $Cur_Col_Meta = $k;
426 [ $k, $self->_expand_expr($v) ]
433 # So that subclasses can override UPDATE ... RETURNING separately from
435 sub _update_returning { shift->_returning(@_) }
439 #======================================================================
441 #======================================================================
446 my $table = $self->_table(shift);
447 my $fields = shift || '*';
451 my ($fields_sql, @bind) = $self->_select_fields($fields);
453 my ($where_sql, @where_bind) = $self->where($where, $order);
454 push @bind, @where_bind;
456 my $sql = join(' ', $self->_sqlcase('select'), $fields_sql,
457 $self->_sqlcase('from'), $table)
460 return wantarray ? ($sql, @bind) : $sql;
464 my ($self, $fields) = @_;
465 return $fields unless ref($fields);
466 return @{ $self->render_aqt(
467 $self->_expand_maybe_list_expr($fields, '-ident')
471 #======================================================================
473 #======================================================================
478 my $table = $self->_table(shift);
482 my($where_sql, @bind) = $self->where($where);
483 my $sql = $self->_sqlcase('delete from ') . $table . $where_sql;
485 if ($options->{returning}) {
486 my ($returning_sql, @returning_bind) = $self->_delete_returning($options);
487 $sql .= $returning_sql;
488 push @bind, @returning_bind;
491 return wantarray ? ($sql, @bind) : $sql;
494 # So that subclasses can override DELETE ... RETURNING separately from
496 sub _delete_returning { shift->_returning(@_) }
500 #======================================================================
502 #======================================================================
506 # Finally, a separate routine just to handle WHERE clauses
508 my ($self, $where, $order) = @_;
510 local $self->{convert_where} = $self->{convert};
513 my ($sql, @bind) = defined($where)
514 ? $self->_recurse_where($where)
516 $sql = (defined $sql and length $sql) ? $self->_sqlcase(' where ') . "( $sql )" : '';
520 my ($order_sql, @order_bind) = $self->_order_by($order);
522 push @bind, @order_bind;
525 return wantarray ? ($sql, @bind) : $sql;
528 { our $Default_Scalar_To = -value }
531 my ($self, $expr, $default_scalar_to) = @_;
532 local our $Default_Scalar_To = $default_scalar_to if $default_scalar_to;
533 $self->_expand_expr($expr);
537 my ($self, $aqt) = @_;
538 my ($k, $v, @rest) = %$aqt;
540 die "Not a node type: $k" unless $k =~ s/^-//;
541 if (my $meth = $self->{render}{$k}) {
542 return $self->$meth($k, $v);
544 die "notreached: $k";
548 my ($self, $expr, $default_scalar_to) = @_;
549 return @{ $self->render_aqt(
550 $self->expand_expr($expr, $default_scalar_to)
555 my ($self, $raw) = @_;
557 return $op if grep $_->{$op}, @{$self}{qw(is_op expand_op render_op)};
558 s/^-(?=.)//, s/\s+/_/g for $op;
563 my ($self, $expr) = @_;
564 our $Expand_Depth ||= 0; local $Expand_Depth = $Expand_Depth + 1;
565 return undef unless defined($expr);
566 if (ref($expr) eq 'HASH') {
567 return undef unless my $kc = keys %$expr;
569 return $self->_expand_op_andor(and => $expr);
571 my ($key, $value) = %$expr;
572 if ($key =~ /^-/ and $key =~ s/ [_\s]? \d+ $//x ) {
573 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
574 . "You probably wanted ...-and => [ $key => COND1, $key => COND2 ... ]";
576 return $self->_expand_hashpair($key, $value);
578 if (ref($expr) eq 'ARRAY') {
579 return $self->_expand_op_andor(lc($self->{logic}), $expr);
581 if (my $literal = is_literal_value($expr)) {
582 return +{ -literal => $literal };
584 if (!ref($expr) or Scalar::Util::blessed($expr)) {
585 return $self->_expand_scalar($expr);
590 sub _expand_hashpair {
591 my ($self, $k, $v) = @_;
592 unless (defined($k) and length($k)) {
593 if (defined($k) and my $literal = is_literal_value($v)) {
594 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
595 return { -literal => $literal };
597 puke "Supplying an empty left hand side argument is not supported";
600 return $self->_expand_hashpair_op($k, $v);
601 } elsif ($k =~ /^[^\w]/i) {
602 my ($lhs, @rhs) = ref($v) eq 'ARRAY' ? @$v : $v;
603 return $self->_expand_op(
604 -op, [ $k, $self->expand_expr($lhs, -ident), @rhs ]
607 return $self->_expand_hashpair_ident($k, $v);
610 sub _expand_hashpair_ident {
611 my ($self, $k, $v) = @_;
613 local our $Cur_Col_Meta = $k;
615 # hash with multiple or no elements is andor
617 if (ref($v) eq 'HASH' and keys %$v != 1) {
618 return $self->_expand_op_andor(and => $v, $k);
621 # undef needs to be re-sent with cmp to achieve IS/IS NOT NULL
623 if (is_undef_value($v)) {
624 return $self->_expand_hashpair_cmp($k => undef);
627 # scalars and objects get expanded as whatever requested or values
629 if (!ref($v) or Scalar::Util::blessed($v)) {
630 return $self->_expand_hashpair_scalar($k, $v);
633 # single key hashref is a hashtriple
635 if (ref($v) eq 'HASH') {
636 return $self->_expand_hashtriple($k, %$v);
639 # arrayref needs re-engineering over the elements
641 if (ref($v) eq 'ARRAY') {
642 return $self->sqlfalse unless @$v;
643 $self->_debug("ARRAY($k) means distribute over elements");
645 $v->[0] =~ /^-(and|or)$/i
646 ? (shift(@{$v = [ @$v ]}), $1)
647 : lc($self->{logic} || 'OR')
649 return $self->_expand_op_andor(
654 if (my $literal = is_literal_value($v)) {
656 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
659 my ($sql, @bind) = @$literal;
660 if ($self->{bindtype} eq 'columns') {
662 $self->_assert_bindval_matches_bindtype($_);
665 return +{ -literal => [ $self->_quote($k).' '.$sql, @bind ] };
671 my ($self, $expr) = @_;
673 return $self->_expand_expr({ (our $Default_Scalar_To) => $expr });
676 sub _expand_hashpair_scalar {
677 my ($self, $k, $v) = @_;
679 return $self->_expand_hashpair_cmp(
680 $k, $self->_expand_scalar($v),
684 sub _expand_hashpair_op {
685 my ($self, $k, $v) = @_;
687 $self->_assert_pass_injection_guard($k =~ /\A-(.*)\Z/s);
689 my $op = $self->_normalize_op($k);
693 my $op = join(' ', split '_', $op);
695 # the old special op system requires illegality for top-level use
698 (our $Expand_Depth) == 1
700 List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}
702 $self->{disable_old_special_ops}
703 and List::Util::first { $op =~ $_->{regex} } @BUILTIN_SPECIAL_OPS
707 puke "Illegal use of top-level '-$op'"
711 if (my $exp = $self->{expand}{$op}) {
712 return $self->$exp($op, $v);
715 # Ops prefixed with -not_ get converted
717 if (my ($rest) = $op =~/^not_(.*)$/) {
720 $self->_expand_expr({ "-${rest}", $v })
726 # the old unary op system means we should touch nothing and let it work
728 my $op = join(' ', split '_', $op);
730 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
731 return { -op => [ $op, $v ] };
735 # an explicit node type is currently assumed to be expanded (this is almost
736 # certainly wrong and there should be expansion anyway)
738 if ($self->{render}{$op}) {
742 my $type = $self->{unknown_unop_always_func} ? -func : -op;
749 and (keys %$v)[0] =~ /^-/
752 (List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}})
761 ($type eq -func and ref($v) eq 'ARRAY')
762 ? map $self->_expand_expr($_), @$v
763 : $self->_expand_expr($v)
767 sub _expand_hashpair_cmp {
768 my ($self, $k, $v) = @_;
769 $self->_expand_hashtriple($k, $self->{cmp}, $v);
772 sub _expand_hashtriple {
773 my ($self, $k, $vk, $vv) = @_;
775 my $ik = $self->_expand_expr({ -ident => $k });
777 my $op = $self->_normalize_op($vk);
778 $self->_assert_pass_injection_guard($op);
780 if ($op =~ s/ _? \d+ $//x ) {
781 return $self->_expand_expr($k, { $vk, $vv });
783 if (my $x = $self->{expand_op}{$op}) {
784 local our $Cur_Col_Meta = $k;
785 return $self->$x($op, $vv, $k);
789 my $op = join(' ', split '_', $op);
791 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}) {
792 return { -op => [ $op, $ik, $vv ] };
794 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
798 { -op => [ $op, $vv ] }
802 if (ref($vv) eq 'ARRAY') {
804 my $logic = (defined($raw[0]) and $raw[0] =~ /^-(and|or)$/i)
805 ? (shift(@raw), lc $1) : 'or';
806 my @values = map +{ $vk => $_ }, @raw;
808 $op =~ $self->{inequality_op}
809 or $op =~ $self->{not_like_op}
811 if (lc($logic) eq 'or' and @values > 1) {
812 belch "A multi-element arrayref as an argument to the inequality op '${\uc(join ' ', split '_', $op)}' "
813 . 'is technically equivalent to an always-true 1=1 (you probably wanted '
814 . "to say ...{ \$inequality_op => [ -and => \@values ] }... instead)"
819 # try to DWIM on equality operators
820 return ($self->_dwim_op_to_is($op,
821 "Supplying an empty arrayref to '%s' is deprecated",
822 "operator '%s' applied on an empty array (field '$k')"
823 ) ? $self->sqlfalse : $self->sqltrue);
825 return $self->_expand_op_andor($logic => \@values, $k);
827 if (is_undef_value($vv)) {
828 my $is = ($self->_dwim_op_to_is($op,
829 "Supplying an undefined argument to '%s' is deprecated",
830 "unexpected operator '%s' with undef operand",
831 ) ? 'is' : 'is not');
833 return $self->_expand_hashpair($k => { $is, undef });
835 local our $Cur_Col_Meta = $k;
839 $self->_expand_expr($vv)
844 my ($self, $raw, $empty, $fail) = @_;
846 my $op = $self->_normalize_op($raw);
848 if ($op =~ /^not$/i) {
851 if ($op =~ $self->{equality_op}) {
854 if ($op =~ $self->{like_op}) {
855 belch(sprintf $empty, uc(join ' ', split '_', $op));
858 if ($op =~ $self->{inequality_op}) {
861 if ($op =~ $self->{not_like_op}) {
862 belch(sprintf $empty, uc(join ' ', split '_', $op));
865 puke(sprintf $fail, $op);
869 my ($self, undef, $args) = @_;
870 my ($func, @args) = @$args;
871 return +{ -func => [ $func, map $self->expand_expr($_), @args ] };
875 my ($self, undef, $body, $k) = @_;
876 return $self->_expand_hashpair_cmp(
877 $k, { -ident => $body }
879 unless (defined($body) or (ref($body) and ref($body) eq 'ARRAY')) {
880 puke "-ident requires a single plain scalar argument (a quotable identifier) or an arrayref of identifier parts";
882 my @parts = map split(/\Q${\($self->{name_sep}||'.')}\E/, $_),
883 ref($body) ? @$body : $body;
884 return { -ident => $parts[-1] } if $self->{_dequalify_idents};
885 unless ($self->{quote_char}) {
886 $self->_assert_pass_injection_guard($_) for @parts;
888 return +{ -ident => \@parts };
892 return $_[0]->_expand_hashpair_cmp(
893 $_[3], { -value => $_[2] },
895 +{ -bind => [ our $Cur_Col_Meta, $_[2] ] };
899 +{ -op => [ 'not', $_[0]->_expand_expr($_[2]) ] };
903 my ($self, undef, $args) = @_;
904 +{ -row => [ map $self->expand_expr($_), @$args ] };
908 my ($self, undef, $args) = @_;
909 my ($op, @opargs) = @$args;
910 if (my $exp = $self->{expand_op}{$op}) {
911 return $self->$exp($op, \@opargs);
913 +{ -op => [ $op, map $self->expand_expr($_), @opargs ] };
917 my ($self, undef, $v) = @_;
919 return $self->_expand_expr($v);
921 puke "-bool => undef not supported" unless defined($v);
922 return $self->_expand_expr({ -ident => $v });
925 sub _expand_op_andor {
926 my ($self, $logop, $v, $k) = @_;
928 $v = [ map +{ $k, $_ },
930 ? (map +{ $_ => $v->{$_} }, sort keys %$v)
934 if (ref($v) eq 'HASH') {
935 return undef unless keys %$v;
938 map $self->_expand_expr({ $_ => $v->{$_} }),
942 if (ref($v) eq 'ARRAY') {
943 $logop eq 'and' or $logop eq 'or' or puke "unknown logic: $logop";
946 (ref($_) eq 'ARRAY' and @$_)
947 or (ref($_) eq 'HASH' and %$_)
953 while (my ($el) = splice @expr, 0, 1) {
954 puke "Supplying an empty left hand side argument is not supported in array-pairs"
955 unless defined($el) and length($el);
956 my $elref = ref($el);
958 local our $Expand_Depth = 0;
959 push(@res, grep defined, $self->_expand_expr({ $el, shift(@expr) }));
960 } elsif ($elref eq 'ARRAY') {
961 push(@res, grep defined, $self->_expand_expr($el)) if @$el;
962 } elsif (my $l = is_literal_value($el)) {
963 push @res, { -literal => $l };
964 } elsif ($elref eq 'HASH') {
965 local our $Expand_Depth = 0;
966 push @res, grep defined, $self->_expand_expr($el) if %$el;
972 # return $res[0] if @res == 1;
973 return { -op => [ $logop, @res ] };
979 my ($self, $op, $vv, $k) = @_;
980 ($k, $vv) = @$vv unless defined $k;
981 puke "$op can only take undef as argument"
985 and exists($vv->{-value})
986 and !defined($vv->{-value})
988 return +{ -op => [ $op.'_null', $self->expand_expr($k, -ident) ] };
991 sub _expand_between {
992 my ($self, $op, $vv, $k) = @_;
993 $k = shift @{$vv = [ @$vv ]} unless defined $k;
994 my @rhs = map $self->_expand_expr($_),
995 ref($vv) eq 'ARRAY' ? @$vv : $vv;
997 (@rhs == 1 and ref($rhs[0]) eq 'HASH' and $rhs[0]->{-literal})
999 (@rhs == 2 and defined($rhs[0]) and defined($rhs[1]))
1001 puke "Operator '${\uc($op)}' requires either an arrayref with two defined values or expressions, or a single literal scalarref/arrayref-ref";
1005 $self->expand_expr(ref($k) ? $k : { -ident => $k }),
1011 my ($self, $op, $vv, $k) = @_;
1012 $k = shift @{$vv = [ @$vv ]} unless defined $k;
1013 if (my $literal = is_literal_value($vv)) {
1014 my ($sql, @bind) = @$literal;
1015 my $opened_sql = $self->_open_outer_paren($sql);
1017 $op, $self->expand_expr($k, -ident),
1018 { -literal => [ $opened_sql, @bind ] }
1022 'SQL::Abstract before v1.75 used to generate incorrect SQL when the '
1023 . "-${\uc($op)} operator was given an undef-containing list: !!!AUDIT YOUR CODE "
1024 . 'AND DATA!!! (the upcoming Data::Query-based version of SQL::Abstract '
1025 . 'will emit the logically correct SQL instead of raising this exception)'
1027 puke("Argument passed to the '${\uc($op)}' operator can not be undefined")
1029 my @rhs = map $self->expand_expr($_, -value),
1030 map { defined($_) ? $_: puke($undef_err) }
1031 (ref($vv) eq 'ARRAY' ? @$vv : $vv);
1032 return $self->${\($op =~ /^not/ ? 'sqltrue' : 'sqlfalse')} unless @rhs;
1036 $self->expand_expr($k, -ident),
1042 my ($self, undef, $v) = @_;
1043 # DBIx::Class requires a nest warning to be emitted once but the private
1044 # method it overrode to do so no longer exists
1045 if ($self->{warn_once_on_nest}) {
1046 unless (our $Nest_Warned) {
1048 "-nest in search conditions is deprecated, you most probably wanted:\n"
1049 .q|{..., -and => [ \%cond0, \@cond1, \'cond2', \[ 'cond3', [ col => bind ] ], etc. ], ... }|
1054 return $self->_expand_expr($v);
1058 my ($self, undef, $bind) = @_;
1059 return { -bind => $bind };
1062 sub _recurse_where {
1063 my ($self, $where, $logic) = @_;
1065 # Special case: top level simple string treated as literal
1067 my $where_exp = (ref($where)
1068 ? $self->_expand_expr($where, $logic)
1069 : { -literal => [ $where ] });
1071 # dispatch expanded expression
1073 my ($sql, @bind) = defined($where_exp) ? @{ $self->render_aqt($where_exp) || [] } : ();
1074 # DBIx::Class used to call _recurse_where in scalar context
1075 # something else might too...
1077 return ($sql, @bind);
1080 belch "Calling _recurse_where in scalar context is deprecated and will go away before 2.0";
1086 my ($self, undef, $ident) = @_;
1088 return [ $self->_convert($self->_quote($ident)) ];
1092 my ($self, undef, $values) = @_;
1093 return $self->join_query_parts('',
1095 $self->_render_op(undef, [ ',', @$values ]),
1101 my ($self, undef, $rest) = @_;
1102 my ($func, @args) = @$rest;
1103 return $self->join_query_parts('',
1104 $self->_sqlcase($func),
1105 $self->join_query_parts('',
1107 $self->join_query_parts(', ', @args),
1114 my ($self, undef, $bind) = @_;
1115 return [ $self->_convert('?'), $self->_bindtype(@$bind) ];
1118 sub _render_literal {
1119 my ($self, undef, $literal) = @_;
1120 $self->_assert_bindval_matches_bindtype(@{$literal}[1..$#$literal]);
1125 my ($self, undef, $v) = @_;
1126 my ($op, @args) = @$v;
1127 if (my $r = $self->{render_op}{$op}) {
1128 return $self->$r($op, \@args);
1133 my $op = join(' ', split '_', $op);
1135 my $ss = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}};
1136 if ($ss and @args > 1) {
1137 puke "Special op '${op}' requires first value to be identifier"
1138 unless my ($ident) = map $_->{-ident}, grep ref($_) eq 'HASH', $args[0];
1139 my $k = join(($self->{name_sep}||'.'), @$ident);
1140 local our $Expand_Depth = 1;
1141 return [ $self->${\($ss->{handler})}($k, $op, $args[1]) ];
1143 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1144 return [ $self->${\($us->{handler})}($op, $args[0]) ];
1147 return $self->_render_unop_paren($op, \@args);
1151 return $self->_render_unop_prefix($op, \@args);
1153 return $self->_render_op_multop($op, \@args);
1159 sub _render_op_between {
1160 my ($self, $op, $args) = @_;
1161 my ($left, $low, $high) = @$args;
1164 puke "Single arg to between must be a literal"
1165 unless $low->{-literal};
1168 +($low, $self->format_keyword('and'), $high);
1171 return $self->join_query_parts(' ',
1172 '(', $left, $self->format_keyword($op), @rh, ')',
1177 my ($self, $op, $args) = @_;
1178 my ($lhs, @rhs) = @$args;
1180 return $self->join_query_parts(' ',
1182 $self->format_keyword($op),
1183 $self->join_query_parts(' ',
1185 $self->join_query_parts(', ', @rhs),
1191 sub _render_op_andor {
1192 my ($self, $op, $args) = @_;
1193 return undef unless @$args;
1194 return $self->join_query_parts('', $args->[0]) if @$args == 1;
1195 my $inner = $self->_render_op_multop($op, $args);
1196 return undef unless defined($inner->[0]) and length($inner->[0]);
1197 return $self->join_query_parts(' ',
1202 sub _render_op_multop {
1203 my ($self, $op, $args) = @_;
1205 return undef unless @parts;
1206 return $self->render_aqt($parts[0]) if @parts == 1;
1207 my $join = ($op eq ','
1209 : ' '.$self->format_keyword($op).' '
1211 return $self->join_query_parts($join, @parts);
1214 sub join_query_parts {
1215 my ($self, $join, @parts) = @_;
1218 ? $self->render_aqt($_)
1219 : ((ref($_) eq 'ARRAY') ? $_ : [ $_ ])
1222 $self->{join_sql_parts}->($join, grep defined, map $_->[0], @final),
1223 (map @{$_}[1..$#$_], @final),
1227 sub _render_unop_paren {
1228 my ($self, $op, $v) = @_;
1229 return $self->join_query_parts('',
1230 '(', $self->_render_unop_prefix($op, $v), ')'
1234 sub _render_unop_prefix {
1235 my ($self, $op, $v) = @_;
1236 return $self->join_query_parts(' ',
1237 $self->_sqlcase($op), $v->[0]
1241 sub _render_unop_postfix {
1242 my ($self, $op, $v) = @_;
1243 return $self->join_query_parts(' ',
1244 $v->[0], $self->format_keyword($op),
1248 # Some databases (SQLite) treat col IN (1, 2) different from
1249 # col IN ( (1, 2) ). Use this to strip all outer parens while
1250 # adding them back in the corresponding method
1251 sub _open_outer_paren {
1252 my ($self, $sql) = @_;
1254 while (my ($inner) = $sql =~ /^ \s* \( (.*) \) \s* $/xs) {
1256 # there are closing parens inside, need the heavy duty machinery
1257 # to reevaluate the extraction starting from $sql (full reevaluation)
1258 if ($inner =~ /\)/) {
1259 require Text::Balanced;
1261 my (undef, $remainder) = do {
1262 # idiotic design - writes to $@ but *DOES NOT* throw exceptions
1264 Text::Balanced::extract_bracketed($sql, '()', qr/\s*/);
1267 # the entire expression needs to be a balanced bracketed thing
1268 # (after an extract no remainder sans trailing space)
1269 last if defined $remainder and $remainder =~ /\S/;
1279 #======================================================================
1281 #======================================================================
1283 sub _expand_order_by {
1284 my ($self, $arg) = @_;
1286 return unless defined($arg) and not (ref($arg) eq 'ARRAY' and !@$arg);
1288 return $self->_expand_maybe_list_expr($arg)
1289 if ref($arg) eq 'HASH' and ($arg->{-op}||[''])->[0] eq ',';
1291 my $expander = sub {
1292 my ($self, $dir, $expr) = @_;
1293 my @to_expand = ref($expr) eq 'ARRAY' ? @$expr : $expr;
1294 foreach my $arg (@to_expand) {
1298 and grep /^-(asc|desc)$/, keys %$arg
1300 puke "ordering direction hash passed to order by must have exactly one key (-asc or -desc)";
1304 defined($dir) ? { -op => [ $dir =~ /^-?(.*)$/ ,=> $_ ] } : $_
1306 map $self->expand_expr($_, -ident),
1307 map ref($_) eq 'ARRAY' ? @$_ : $_, @to_expand;
1308 return undef unless @exp;
1309 return undef if @exp == 1 and not defined($exp[0]);
1310 return +{ -op => [ ',', @exp ] };
1313 local @{$self->{expand}}{qw(asc desc)} = (($expander) x 2);
1315 return $self->$expander(undef, $arg);
1319 my ($self, $arg) = @_;
1321 return '' unless defined(my $expanded = $self->_expand_order_by($arg));
1323 my ($sql, @bind) = @{ $self->render_aqt($expanded) };
1325 return '' unless length($sql);
1327 my $final_sql = $self->_sqlcase(' order by ').$sql;
1329 return ($final_sql, @bind);
1332 # _order_by no longer needs to call this so doesn't but DBIC uses it.
1334 sub _order_by_chunks {
1335 my ($self, $arg) = @_;
1337 return () unless defined(my $expanded = $self->_expand_order_by($arg));
1339 my @res = $self->_chunkify_order_by($expanded);
1340 (ref() ? $_->[0] : $_) .= '' for @res;
1344 sub _chunkify_order_by {
1345 my ($self, $expanded) = @_;
1347 return grep length, @{ $self->render_aqt($expanded) }
1348 if $expanded->{-ident} or @{$expanded->{-literal}||[]} == 1;
1351 if (ref() eq 'HASH' and $_->{-op} and $_->{-op}[0] eq ',') {
1352 my ($comma, @list) = @{$_->{-op}};
1353 return map $self->_chunkify_order_by($_), @list;
1355 return $self->render_aqt($_);
1359 #======================================================================
1360 # DATASOURCE (FOR NOW, JUST PLAIN TABLE OR LIST OF TABLES)
1361 #======================================================================
1367 $self->_expand_maybe_list_expr($from, -ident)
1372 #======================================================================
1374 #======================================================================
1376 sub _expand_maybe_list_expr {
1377 my ($self, $expr, $default) = @_;
1379 ',', map $self->expand_expr($_, $default),
1380 @{$expr->{-op}}[1..$#{$expr->{-op}}]
1381 ] } if ref($expr) eq 'HASH' and ($expr->{-op}||[''])->[0] eq ',';
1382 return +{ -op => [ ',',
1383 map $self->expand_expr($_, $default),
1384 ref($expr) eq 'ARRAY' ? @$expr : $expr
1388 # highly optimized, as it's called way too often
1390 # my ($self, $label) = @_;
1392 return '' unless defined $_[1];
1393 return ${$_[1]} if ref($_[1]) eq 'SCALAR';
1394 puke 'Identifier cannot be hashref' if ref($_[1]) eq 'HASH';
1396 unless ($_[0]->{quote_char}) {
1397 if (ref($_[1]) eq 'ARRAY') {
1398 return join($_[0]->{name_sep}||'.', @{$_[1]});
1400 $_[0]->_assert_pass_injection_guard($_[1]);
1405 my $qref = ref $_[0]->{quote_char};
1407 !$qref ? ($_[0]->{quote_char}, $_[0]->{quote_char})
1408 : ($qref eq 'ARRAY') ? @{$_[0]->{quote_char}}
1409 : puke "Unsupported quote_char format: $_[0]->{quote_char}";
1411 my $esc = $_[0]->{escape_char} || $r;
1413 # parts containing * are naturally unquoted
1415 $_[0]->{name_sep}||'',
1419 : do { (my $n = $_) =~ s/(\Q$esc\E|\Q$r\E)/$esc$1/g; $l . $n . $r }
1421 (ref($_[1]) eq 'ARRAY'
1425 ? split (/\Q$_[0]->{name_sep}\E/, $_[1] )
1433 # Conversion, if applicable
1435 #my ($self, $arg) = @_;
1436 if (my $conv = $_[0]->{convert_where}) {
1437 return @{ $_[0]->join_query_parts('',
1438 $_[0]->format_keyword($conv),
1447 #my ($self, $col, @vals) = @_;
1448 # called often - tighten code
1449 return $_[0]->{bindtype} eq 'columns'
1450 ? map {[$_[1], $_]} @_[2 .. $#_]
1455 # Dies if any element of @bind is not in [colname => value] format
1456 # if bindtype is 'columns'.
1457 sub _assert_bindval_matches_bindtype {
1458 # my ($self, @bind) = @_;
1460 if ($self->{bindtype} eq 'columns') {
1462 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
1463 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
1469 sub _join_sql_clauses {
1470 my ($self, $logic, $clauses_aref, $bind_aref) = @_;
1472 if (@$clauses_aref > 1) {
1473 my $join = " " . $self->_sqlcase($logic) . " ";
1474 my $sql = '( ' . join($join, @$clauses_aref) . ' )';
1475 return ($sql, @$bind_aref);
1477 elsif (@$clauses_aref) {
1478 return ($clauses_aref->[0], @$bind_aref); # no parentheses
1481 return (); # if no SQL, ignore @$bind_aref
1486 # Fix SQL case, if so requested
1488 # LDNOTE: if $self->{case} is true, then it contains 'lower', so we
1489 # don't touch the argument ... crooked logic, but let's not change it!
1490 return $_[0]->{case} ? $_[1] : uc($_[1]);
1493 sub format_keyword { $_[0]->_sqlcase(join ' ', split '_', $_[1]) }
1495 #======================================================================
1496 # DISPATCHING FROM REFKIND
1497 #======================================================================
1500 my ($self, $data) = @_;
1502 return 'UNDEF' unless defined $data;
1504 # blessed objects are treated like scalars
1505 my $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1507 return 'SCALAR' unless $ref;
1510 while ($ref eq 'REF') {
1512 $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1516 return ($ref||'SCALAR') . ('REF' x $n_steps);
1520 my ($self, $data) = @_;
1521 my @try = ($self->_refkind($data));
1522 push @try, 'SCALAR_or_UNDEF' if $try[0] eq 'SCALAR' || $try[0] eq 'UNDEF';
1523 push @try, 'FALLBACK';
1527 sub _METHOD_FOR_refkind {
1528 my ($self, $meth_prefix, $data) = @_;
1531 for (@{$self->_try_refkind($data)}) {
1532 $method = $self->can($meth_prefix."_".$_)
1536 return $method || puke "cannot dispatch on '$meth_prefix' for ".$self->_refkind($data);
1540 sub _SWITCH_refkind {
1541 my ($self, $data, $dispatch_table) = @_;
1544 for (@{$self->_try_refkind($data)}) {
1545 $coderef = $dispatch_table->{$_}
1549 puke "no dispatch entry for ".$self->_refkind($data)
1558 #======================================================================
1559 # VALUES, GENERATE, AUTOLOAD
1560 #======================================================================
1562 # LDNOTE: original code from nwiger, didn't touch code in that section
1563 # I feel the AUTOLOAD stuff should not be the default, it should
1564 # only be activated on explicit demand by user.
1568 my $data = shift || return;
1569 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
1570 unless ref $data eq 'HASH';
1573 foreach my $k (sort keys %$data) {
1574 my $v = $data->{$k};
1575 $self->_SWITCH_refkind($v, {
1577 if ($self->{array_datatypes}) { # array datatype
1578 push @all_bind, $self->_bindtype($k, $v);
1580 else { # literal SQL with bind
1581 my ($sql, @bind) = @$v;
1582 $self->_assert_bindval_matches_bindtype(@bind);
1583 push @all_bind, @bind;
1586 ARRAYREFREF => sub { # literal SQL with bind
1587 my ($sql, @bind) = @${$v};
1588 $self->_assert_bindval_matches_bindtype(@bind);
1589 push @all_bind, @bind;
1591 SCALARREF => sub { # literal SQL without bind
1593 SCALAR_or_UNDEF => sub {
1594 push @all_bind, $self->_bindtype($k, $v);
1605 my(@sql, @sqlq, @sqlv);
1609 if ($ref eq 'HASH') {
1610 for my $k (sort keys %$_) {
1613 my $label = $self->_quote($k);
1614 if ($r eq 'ARRAY') {
1615 # literal SQL with bind
1616 my ($sql, @bind) = @$v;
1617 $self->_assert_bindval_matches_bindtype(@bind);
1618 push @sqlq, "$label = $sql";
1620 } elsif ($r eq 'SCALAR') {
1621 # literal SQL without bind
1622 push @sqlq, "$label = $$v";
1624 push @sqlq, "$label = ?";
1625 push @sqlv, $self->_bindtype($k, $v);
1628 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
1629 } elsif ($ref eq 'ARRAY') {
1630 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
1633 if ($r eq 'ARRAY') { # literal SQL with bind
1634 my ($sql, @bind) = @$v;
1635 $self->_assert_bindval_matches_bindtype(@bind);
1638 } elsif ($r eq 'SCALAR') { # literal SQL without bind
1639 # embedded literal SQL
1646 push @sql, '(' . join(', ', @sqlq) . ')';
1647 } elsif ($ref eq 'SCALAR') {
1651 # strings get case twiddled
1652 push @sql, $self->_sqlcase($_);
1656 my $sql = join ' ', @sql;
1658 # this is pretty tricky
1659 # if ask for an array, return ($stmt, @bind)
1660 # otherwise, s/?/shift @sqlv/ to put it inline
1662 return ($sql, @sqlv);
1664 1 while $sql =~ s/\?/my $d = shift(@sqlv);
1665 ref $d ? $d->[1] : $d/e;
1674 # This allows us to check for a local, then _form, attr
1676 my($name) = $AUTOLOAD =~ /.*::(.+)/;
1677 puke "AUTOLOAD invoked for method name ${name} and allow_autoload option not set" unless $self->{allow_autoload};
1678 return $self->generate($name, @_);
1689 SQL::Abstract - Generate SQL from Perl data structures
1695 my $sql = SQL::Abstract->new;
1697 my($stmt, @bind) = $sql->select($source, \@fields, \%where, $order);
1699 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
1701 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
1703 my($stmt, @bind) = $sql->delete($table, \%where);
1705 # Then, use these in your DBI statements
1706 my $sth = $dbh->prepare($stmt);
1707 $sth->execute(@bind);
1709 # Just generate the WHERE clause
1710 my($stmt, @bind) = $sql->where(\%where, $order);
1712 # Return values in the same order, for hashed queries
1713 # See PERFORMANCE section for more details
1714 my @bind = $sql->values(\%fieldvals);
1718 This module was inspired by the excellent L<DBIx::Abstract>.
1719 However, in using that module I found that what I really wanted
1720 to do was generate SQL, but still retain complete control over my
1721 statement handles and use the DBI interface. So, I set out to
1722 create an abstract SQL generation module.
1724 While based on the concepts used by L<DBIx::Abstract>, there are
1725 several important differences, especially when it comes to WHERE
1726 clauses. I have modified the concepts used to make the SQL easier
1727 to generate from Perl data structures and, IMO, more intuitive.
1728 The underlying idea is for this module to do what you mean, based
1729 on the data structures you provide it. The big advantage is that
1730 you don't have to modify your code every time your data changes,
1731 as this module figures it out.
1733 To begin with, an SQL INSERT is as easy as just specifying a hash
1734 of C<key=value> pairs:
1737 name => 'Jimbo Bobson',
1738 phone => '123-456-7890',
1739 address => '42 Sister Lane',
1740 city => 'St. Louis',
1741 state => 'Louisiana',
1744 The SQL can then be generated with this:
1746 my($stmt, @bind) = $sql->insert('people', \%data);
1748 Which would give you something like this:
1750 $stmt = "INSERT INTO people
1751 (address, city, name, phone, state)
1752 VALUES (?, ?, ?, ?, ?)";
1753 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
1754 '123-456-7890', 'Louisiana');
1756 These are then used directly in your DBI code:
1758 my $sth = $dbh->prepare($stmt);
1759 $sth->execute(@bind);
1761 =head2 Inserting and Updating Arrays
1763 If your database has array types (like for example Postgres),
1764 activate the special option C<< array_datatypes => 1 >>
1765 when creating the C<SQL::Abstract> object.
1766 Then you may use an arrayref to insert and update database array types:
1768 my $sql = SQL::Abstract->new(array_datatypes => 1);
1770 planets => [qw/Mercury Venus Earth Mars/]
1773 my($stmt, @bind) = $sql->insert('solar_system', \%data);
1777 $stmt = "INSERT INTO solar_system (planets) VALUES (?)"
1779 @bind = (['Mercury', 'Venus', 'Earth', 'Mars']);
1782 =head2 Inserting and Updating SQL
1784 In order to apply SQL functions to elements of your C<%data> you may
1785 specify a reference to an arrayref for the given hash value. For example,
1786 if you need to execute the Oracle C<to_date> function on a value, you can
1787 say something like this:
1791 date_entered => \[ "to_date(?,'MM/DD/YYYY')", "03/02/2003" ],
1794 The first value in the array is the actual SQL. Any other values are
1795 optional and would be included in the bind values array. This gives
1798 my($stmt, @bind) = $sql->insert('people', \%data);
1800 $stmt = "INSERT INTO people (name, date_entered)
1801 VALUES (?, to_date(?,'MM/DD/YYYY'))";
1802 @bind = ('Bill', '03/02/2003');
1804 An UPDATE is just as easy, all you change is the name of the function:
1806 my($stmt, @bind) = $sql->update('people', \%data);
1808 Notice that your C<%data> isn't touched; the module will generate
1809 the appropriately quirky SQL for you automatically. Usually you'll
1810 want to specify a WHERE clause for your UPDATE, though, which is
1811 where handling C<%where> hashes comes in handy...
1813 =head2 Complex where statements
1815 This module can generate pretty complicated WHERE statements
1816 easily. For example, simple C<key=value> pairs are taken to mean
1817 equality, and if you want to see if a field is within a set
1818 of values, you can use an arrayref. Let's say we wanted to
1819 SELECT some data based on this criteria:
1822 requestor => 'inna',
1823 worker => ['nwiger', 'rcwe', 'sfz'],
1824 status => { '!=', 'completed' }
1827 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
1829 The above would give you something like this:
1831 $stmt = "SELECT * FROM tickets WHERE
1832 ( requestor = ? ) AND ( status != ? )
1833 AND ( worker = ? OR worker = ? OR worker = ? )";
1834 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
1836 Which you could then use in DBI code like so:
1838 my $sth = $dbh->prepare($stmt);
1839 $sth->execute(@bind);
1845 The methods are simple. There's one for every major SQL operation,
1846 and a constructor you use first. The arguments are specified in a
1847 similar order for each method (table, then fields, then a where
1848 clause) to try and simplify things.
1850 =head2 new(option => 'value')
1852 The C<new()> function takes a list of options and values, and returns
1853 a new B<SQL::Abstract> object which can then be used to generate SQL
1854 through the methods below. The options accepted are:
1860 If set to 'lower', then SQL will be generated in all lowercase. By
1861 default SQL is generated in "textbook" case meaning something like:
1863 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
1865 Any setting other than 'lower' is ignored.
1869 This determines what the default comparison operator is. By default
1870 it is C<=>, meaning that a hash like this:
1872 %where = (name => 'nwiger', email => 'nate@wiger.org');
1874 Will generate SQL like this:
1876 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
1878 However, you may want loose comparisons by default, so if you set
1879 C<cmp> to C<like> you would get SQL such as:
1881 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
1883 You can also override the comparison on an individual basis - see
1884 the huge section on L</"WHERE CLAUSES"> at the bottom.
1886 =item sqltrue, sqlfalse
1888 Expressions for inserting boolean values within SQL statements.
1889 By default these are C<1=1> and C<1=0>. They are used
1890 by the special operators C<-in> and C<-not_in> for generating
1891 correct SQL even when the argument is an empty array (see below).
1895 This determines the default logical operator for multiple WHERE
1896 statements in arrays or hashes. If absent, the default logic is "or"
1897 for arrays, and "and" for hashes. This means that a WHERE
1901 event_date => {'>=', '2/13/99'},
1902 event_date => {'<=', '4/24/03'},
1905 will generate SQL like this:
1907 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
1909 This is probably not what you want given this query, though (look
1910 at the dates). To change the "OR" to an "AND", simply specify:
1912 my $sql = SQL::Abstract->new(logic => 'and');
1914 Which will change the above C<WHERE> to:
1916 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
1918 The logic can also be changed locally by inserting
1919 a modifier in front of an arrayref:
1921 @where = (-and => [event_date => {'>=', '2/13/99'},
1922 event_date => {'<=', '4/24/03'} ]);
1924 See the L</"WHERE CLAUSES"> section for explanations.
1928 This will automatically convert comparisons using the specified SQL
1929 function for both column and value. This is mostly used with an argument
1930 of C<upper> or C<lower>, so that the SQL will have the effect of
1931 case-insensitive "searches". For example, this:
1933 $sql = SQL::Abstract->new(convert => 'upper');
1934 %where = (keywords => 'MaKe iT CAse inSeNSItive');
1936 Will turn out the following SQL:
1938 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
1940 The conversion can be C<upper()>, C<lower()>, or any other SQL function
1941 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
1942 not validate this option; it will just pass through what you specify verbatim).
1946 This is a kludge because many databases suck. For example, you can't
1947 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
1948 Instead, you have to use C<bind_param()>:
1950 $sth->bind_param(1, 'reg data');
1951 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
1953 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
1954 which loses track of which field each slot refers to. Fear not.
1956 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
1957 Currently, you can specify either C<normal> (default) or C<columns>. If you
1958 specify C<columns>, you will get an array that looks like this:
1960 my $sql = SQL::Abstract->new(bindtype => 'columns');
1961 my($stmt, @bind) = $sql->insert(...);
1964 [ 'column1', 'value1' ],
1965 [ 'column2', 'value2' ],
1966 [ 'column3', 'value3' ],
1969 You can then iterate through this manually, using DBI's C<bind_param()>.
1971 $sth->prepare($stmt);
1974 my($col, $data) = @$_;
1975 if ($col eq 'details' || $col eq 'comments') {
1976 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
1977 } elsif ($col eq 'image') {
1978 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
1980 $sth->bind_param($i, $data);
1984 $sth->execute; # execute without @bind now
1986 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
1987 Basically, the advantage is still that you don't have to care which fields
1988 are or are not included. You could wrap that above C<for> loop in a simple
1989 sub called C<bind_fields()> or something and reuse it repeatedly. You still
1990 get a layer of abstraction over manual SQL specification.
1992 Note that if you set L</bindtype> to C<columns>, the C<\[ $sql, @bind ]>
1993 construct (see L</Literal SQL with placeholders and bind values (subqueries)>)
1994 will expect the bind values in this format.
1998 This is the character that a table or column name will be quoted
1999 with. By default this is an empty string, but you could set it to
2000 the character C<`>, to generate SQL like this:
2002 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
2004 Alternatively, you can supply an array ref of two items, the first being the left
2005 hand quote character, and the second the right hand quote character. For
2006 example, you could supply C<['[',']']> for SQL Server 2000 compliant quotes
2007 that generates SQL like this:
2009 SELECT [a_field] FROM [a_table] WHERE [some_field] LIKE '%someval%'
2011 Quoting is useful if you have tables or columns names that are reserved
2012 words in your database's SQL dialect.
2016 This is the character that will be used to escape L</quote_char>s appearing
2017 in an identifier before it has been quoted.
2019 The parameter default in case of a single L</quote_char> character is the quote
2022 When opening-closing-style quoting is used (L</quote_char> is an arrayref)
2023 this parameter defaults to the B<closing (right)> L</quote_char>. Occurrences
2024 of the B<opening (left)> L</quote_char> within the identifier are currently left
2025 untouched. The default for opening-closing-style quotes may change in future
2026 versions, thus you are B<strongly encouraged> to specify the escape character
2031 This is the character that separates a table and column name. It is
2032 necessary to specify this when the C<quote_char> option is selected,
2033 so that tables and column names can be individually quoted like this:
2035 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
2037 =item injection_guard
2039 A regular expression C<qr/.../> that is applied to any C<-function> and unquoted
2040 column name specified in a query structure. This is a safety mechanism to avoid
2041 injection attacks when mishandling user input e.g.:
2043 my %condition_as_column_value_pairs = get_values_from_user();
2044 $sqla->select( ... , \%condition_as_column_value_pairs );
2046 If the expression matches an exception is thrown. Note that literal SQL
2047 supplied via C<\'...'> or C<\['...']> is B<not> checked in any way.
2049 Defaults to checking for C<;> and the C<GO> keyword (TransactSQL)
2051 =item array_datatypes
2053 When this option is true, arrayrefs in INSERT or UPDATE are
2054 interpreted as array datatypes and are passed directly
2056 When this option is false, arrayrefs are interpreted
2057 as literal SQL, just like refs to arrayrefs
2058 (but this behavior is for backwards compatibility; when writing
2059 new queries, use the "reference to arrayref" syntax
2065 Takes a reference to a list of "special operators"
2066 to extend the syntax understood by L<SQL::Abstract>.
2067 See section L</"SPECIAL OPERATORS"> for details.
2071 Takes a reference to a list of "unary operators"
2072 to extend the syntax understood by L<SQL::Abstract>.
2073 See section L</"UNARY OPERATORS"> for details.
2079 =head2 insert($table, \@values || \%fieldvals, \%options)
2081 This is the simplest function. You simply give it a table name
2082 and either an arrayref of values or hashref of field/value pairs.
2083 It returns an SQL INSERT statement and a list of bind values.
2084 See the sections on L</"Inserting and Updating Arrays"> and
2085 L</"Inserting and Updating SQL"> for information on how to insert
2086 with those data types.
2088 The optional C<\%options> hash reference may contain additional
2089 options to generate the insert SQL. Currently supported options
2096 Takes either a scalar of raw SQL fields, or an array reference of
2097 field names, and adds on an SQL C<RETURNING> statement at the end.
2098 This allows you to return data generated by the insert statement
2099 (such as row IDs) without performing another C<SELECT> statement.
2100 Note, however, this is not part of the SQL standard and may not
2101 be supported by all database engines.
2105 =head2 update($table, \%fieldvals, \%where, \%options)
2107 This takes a table, hashref of field/value pairs, and an optional
2108 hashref L<WHERE clause|/WHERE CLAUSES>. It returns an SQL UPDATE function and a list
2110 See the sections on L</"Inserting and Updating Arrays"> and
2111 L</"Inserting and Updating SQL"> for information on how to insert
2112 with those data types.
2114 The optional C<\%options> hash reference may contain additional
2115 options to generate the update SQL. Currently supported options
2122 See the C<returning> option to
2123 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2127 =head2 select($source, $fields, $where, $order)
2129 This returns a SQL SELECT statement and associated list of bind values, as
2130 specified by the arguments:
2136 Specification of the 'FROM' part of the statement.
2137 The argument can be either a plain scalar (interpreted as a table
2138 name, will be quoted), or an arrayref (interpreted as a list
2139 of table names, joined by commas, quoted), or a scalarref
2140 (literal SQL, not quoted).
2144 Specification of the list of fields to retrieve from
2146 The argument can be either an arrayref (interpreted as a list
2147 of field names, will be joined by commas and quoted), or a
2148 plain scalar (literal SQL, not quoted).
2149 Please observe that this API is not as flexible as that of
2150 the first argument C<$source>, for backwards compatibility reasons.
2154 Optional argument to specify the WHERE part of the query.
2155 The argument is most often a hashref, but can also be
2156 an arrayref or plain scalar --
2157 see section L<WHERE clause|/"WHERE CLAUSES"> for details.
2161 Optional argument to specify the ORDER BY part of the query.
2162 The argument can be a scalar, a hashref or an arrayref
2163 -- see section L<ORDER BY clause|/"ORDER BY CLAUSES">
2169 =head2 delete($table, \%where, \%options)
2171 This takes a table name and optional hashref L<WHERE clause|/WHERE CLAUSES>.
2172 It returns an SQL DELETE statement and list of bind values.
2174 The optional C<\%options> hash reference may contain additional
2175 options to generate the delete SQL. Currently supported options
2182 See the C<returning> option to
2183 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2187 =head2 where(\%where, $order)
2189 This is used to generate just the WHERE clause. For example,
2190 if you have an arbitrary data structure and know what the
2191 rest of your SQL is going to look like, but want an easy way
2192 to produce a WHERE clause, use this. It returns an SQL WHERE
2193 clause and list of bind values.
2196 =head2 values(\%data)
2198 This just returns the values from the hash C<%data>, in the same
2199 order that would be returned from any of the other above queries.
2200 Using this allows you to markedly speed up your queries if you
2201 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
2203 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
2205 Warning: This is an experimental method and subject to change.
2207 This returns arbitrarily generated SQL. It's a really basic shortcut.
2208 It will return two different things, depending on return context:
2210 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
2211 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
2213 These would return the following:
2215 # First calling form
2216 $stmt = "CREATE TABLE test (?, ?)";
2217 @bind = (field1, field2);
2219 # Second calling form
2220 $stmt_and_val = "CREATE TABLE test (field1, field2)";
2222 Depending on what you're trying to do, it's up to you to choose the correct
2223 format. In this example, the second form is what you would want.
2227 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
2231 ALTER SESSION SET nls_date_format = 'MM/YY'
2233 You get the idea. Strings get their case twiddled, but everything
2234 else remains verbatim.
2236 =head1 EXPORTABLE FUNCTIONS
2238 =head2 is_plain_value
2240 Determines if the supplied argument is a plain value as understood by this
2245 =item * The value is C<undef>
2247 =item * The value is a non-reference
2249 =item * The value is an object with stringification overloading
2251 =item * The value is of the form C<< { -value => $anything } >>
2255 On failure returns C<undef>, on success returns a B<scalar> reference
2256 to the original supplied argument.
2262 The stringification overloading detection is rather advanced: it takes
2263 into consideration not only the presence of a C<""> overload, but if that
2264 fails also checks for enabled
2265 L<autogenerated versions of C<"">|overload/Magic Autogeneration>, based
2266 on either C<0+> or C<bool>.
2268 Unfortunately testing in the field indicates that this
2269 detection B<< may tickle a latent bug in perl versions before 5.018 >>,
2270 but only when very large numbers of stringifying objects are involved.
2271 At the time of writing ( Sep 2014 ) there is no clear explanation of
2272 the direct cause, nor is there a manageably small test case that reliably
2273 reproduces the problem.
2275 If you encounter any of the following exceptions in B<random places within
2276 your application stack> - this module may be to blame:
2278 Operation "ne": no method found,
2279 left argument in overloaded package <something>,
2280 right argument in overloaded package <something>
2284 Stub found while resolving method "???" overloading """" in package <something>
2286 If you fall victim to the above - please attempt to reduce the problem
2287 to something that could be sent to the L<SQL::Abstract developers
2288 |DBIx::Class/GETTING HELP/SUPPORT>
2289 (either publicly or privately). As a workaround in the meantime you can
2290 set C<$ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}> to a true
2291 value, which will most likely eliminate your problem (at the expense of
2292 not being able to properly detect exotic forms of stringification).
2294 This notice and environment variable will be removed in a future version,
2295 as soon as the underlying problem is found and a reliable workaround is
2300 =head2 is_literal_value
2302 Determines if the supplied argument is a literal value as understood by this
2307 =item * C<\$sql_string>
2309 =item * C<\[ $sql_string, @bind_values ]>
2313 On failure returns C<undef>, on success returns an B<array> reference
2314 containing the unpacked version of the supplied literal SQL and bind values.
2316 =head1 WHERE CLAUSES
2320 This module uses a variation on the idea from L<DBIx::Abstract>. It
2321 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
2322 module is that things in arrays are OR'ed, and things in hashes
2325 The easiest way to explain is to show lots of examples. After
2326 each C<%where> hash shown, it is assumed you used:
2328 my($stmt, @bind) = $sql->where(\%where);
2330 However, note that the C<%where> hash can be used directly in any
2331 of the other functions as well, as described above.
2333 =head2 Key-value pairs
2335 So, let's get started. To begin, a simple hash:
2339 status => 'completed'
2342 Is converted to SQL C<key = val> statements:
2344 $stmt = "WHERE user = ? AND status = ?";
2345 @bind = ('nwiger', 'completed');
2347 One common thing I end up doing is having a list of values that
2348 a field can be in. To do this, simply specify a list inside of
2353 status => ['assigned', 'in-progress', 'pending'];
2356 This simple code will create the following:
2358 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
2359 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
2361 A field associated to an empty arrayref will be considered a
2362 logical false and will generate 0=1.
2364 =head2 Tests for NULL values
2366 If the value part is C<undef> then this is converted to SQL <IS NULL>
2375 $stmt = "WHERE user = ? AND status IS NULL";
2378 To test if a column IS NOT NULL:
2382 status => { '!=', undef },
2385 =head2 Specific comparison operators
2387 If you want to specify a different type of operator for your comparison,
2388 you can use a hashref for a given column:
2392 status => { '!=', 'completed' }
2395 Which would generate:
2397 $stmt = "WHERE user = ? AND status != ?";
2398 @bind = ('nwiger', 'completed');
2400 To test against multiple values, just enclose the values in an arrayref:
2402 status => { '=', ['assigned', 'in-progress', 'pending'] };
2404 Which would give you:
2406 "WHERE status = ? OR status = ? OR status = ?"
2409 The hashref can also contain multiple pairs, in which case it is expanded
2410 into an C<AND> of its elements:
2414 status => { '!=', 'completed', -not_like => 'pending%' }
2417 # Or more dynamically, like from a form
2418 $where{user} = 'nwiger';
2419 $where{status}{'!='} = 'completed';
2420 $where{status}{'-not_like'} = 'pending%';
2422 # Both generate this
2423 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
2424 @bind = ('nwiger', 'completed', 'pending%');
2427 To get an OR instead, you can combine it with the arrayref idea:
2431 priority => [ { '=', 2 }, { '>', 5 } ]
2434 Which would generate:
2436 $stmt = "WHERE ( priority = ? OR priority > ? ) AND user = ?";
2437 @bind = ('2', '5', 'nwiger');
2439 If you want to include literal SQL (with or without bind values), just use a
2440 scalar reference or reference to an arrayref as the value:
2443 date_entered => { '>' => \["to_date(?, 'MM/DD/YYYY')", "11/26/2008"] },
2444 date_expires => { '<' => \"now()" }
2447 Which would generate:
2449 $stmt = "WHERE date_entered > to_date(?, 'MM/DD/YYYY') AND date_expires < now()";
2450 @bind = ('11/26/2008');
2453 =head2 Logic and nesting operators
2455 In the example above,
2456 there is a subtle trap if you want to say something like
2457 this (notice the C<AND>):
2459 WHERE priority != ? AND priority != ?
2461 Because, in Perl you I<can't> do this:
2463 priority => { '!=' => 2, '!=' => 1 }
2465 As the second C<!=> key will obliterate the first. The solution
2466 is to use the special C<-modifier> form inside an arrayref:
2468 priority => [ -and => {'!=', 2},
2472 Normally, these would be joined by C<OR>, but the modifier tells it
2473 to use C<AND> instead. (Hint: You can use this in conjunction with the
2474 C<logic> option to C<new()> in order to change the way your queries
2475 work by default.) B<Important:> Note that the C<-modifier> goes
2476 B<INSIDE> the arrayref, as an extra first element. This will
2477 B<NOT> do what you think it might:
2479 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
2481 Here is a quick list of equivalencies, since there is some overlap:
2484 status => {'!=', 'completed', 'not like', 'pending%' }
2485 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
2488 status => {'=', ['assigned', 'in-progress']}
2489 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
2490 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
2494 =head2 Special operators: IN, BETWEEN, etc.
2496 You can also use the hashref format to compare a list of fields using the
2497 C<IN> comparison operator, by specifying the list as an arrayref:
2500 status => 'completed',
2501 reportid => { -in => [567, 2335, 2] }
2504 Which would generate:
2506 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
2507 @bind = ('completed', '567', '2335', '2');
2509 The reverse operator C<-not_in> generates SQL C<NOT IN> and is used in
2512 If the argument to C<-in> is an empty array, 'sqlfalse' is generated
2513 (by default: C<1=0>). Similarly, C<< -not_in => [] >> generates
2514 'sqltrue' (by default: C<1=1>).
2516 In addition to the array you can supply a chunk of literal sql or
2517 literal sql with bind:
2520 customer => { -in => \[
2521 'SELECT cust_id FROM cust WHERE balance > ?',
2524 status => { -in => \'SELECT status_codes FROM states' },
2530 customer IN ( SELECT cust_id FROM cust WHERE balance > ? )
2531 AND status IN ( SELECT status_codes FROM states )
2535 Finally, if the argument to C<-in> is not a reference, it will be
2536 treated as a single-element array.
2538 Another pair of operators is C<-between> and C<-not_between>,
2539 used with an arrayref of two values:
2543 completion_date => {
2544 -not_between => ['2002-10-01', '2003-02-06']
2550 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
2552 Just like with C<-in> all plausible combinations of literal SQL
2556 start0 => { -between => [ 1, 2 ] },
2557 start1 => { -between => \["? AND ?", 1, 2] },
2558 start2 => { -between => \"lower(x) AND upper(y)" },
2559 start3 => { -between => [
2561 \["upper(?)", 'stuff' ],
2568 ( start0 BETWEEN ? AND ? )
2569 AND ( start1 BETWEEN ? AND ? )
2570 AND ( start2 BETWEEN lower(x) AND upper(y) )
2571 AND ( start3 BETWEEN lower(x) AND upper(?) )
2573 @bind = (1, 2, 1, 2, 'stuff');
2576 These are the two builtin "special operators"; but the
2577 list can be expanded: see section L</"SPECIAL OPERATORS"> below.
2579 =head2 Unary operators: bool
2581 If you wish to test against boolean columns or functions within your
2582 database you can use the C<-bool> and C<-not_bool> operators. For
2583 example to test the column C<is_user> being true and the column
2584 C<is_enabled> being false you would use:-
2588 -not_bool => 'is_enabled',
2593 WHERE is_user AND NOT is_enabled
2595 If a more complex combination is required, testing more conditions,
2596 then you should use the and/or operators:-
2601 -not_bool => { two=> { -rlike => 'bar' } },
2602 -not_bool => { three => [ { '=', 2 }, { '>', 5 } ] },
2613 (NOT ( three = ? OR three > ? ))
2616 =head2 Nested conditions, -and/-or prefixes
2618 So far, we've seen how multiple conditions are joined with a top-level
2619 C<AND>. We can change this by putting the different conditions we want in
2620 hashes and then putting those hashes in an array. For example:
2625 status => { -like => ['pending%', 'dispatched'] },
2629 status => 'unassigned',
2633 This data structure would create the following:
2635 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
2636 OR ( user = ? AND status = ? ) )";
2637 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
2640 Clauses in hashrefs or arrayrefs can be prefixed with an C<-and> or C<-or>
2641 to change the logic inside:
2647 -and => [ workhrs => {'>', 20}, geo => 'ASIA' ],
2648 -or => { workhrs => {'<', 50}, geo => 'EURO' },
2655 $stmt = "WHERE ( user = ?
2656 AND ( ( workhrs > ? AND geo = ? )
2657 OR ( workhrs < ? OR geo = ? ) ) )";
2658 @bind = ('nwiger', '20', 'ASIA', '50', 'EURO');
2660 =head3 Algebraic inconsistency, for historical reasons
2662 C<Important note>: when connecting several conditions, the C<-and->|C<-or>
2663 operator goes C<outside> of the nested structure; whereas when connecting
2664 several constraints on one column, the C<-and> operator goes
2665 C<inside> the arrayref. Here is an example combining both features:
2668 -and => [a => 1, b => 2],
2669 -or => [c => 3, d => 4],
2670 e => [-and => {-like => 'foo%'}, {-like => '%bar'} ]
2675 WHERE ( ( ( a = ? AND b = ? )
2676 OR ( c = ? OR d = ? )
2677 OR ( e LIKE ? AND e LIKE ? ) ) )
2679 This difference in syntax is unfortunate but must be preserved for
2680 historical reasons. So be careful: the two examples below would
2681 seem algebraically equivalent, but they are not
2684 { -like => 'foo%' },
2685 { -like => '%bar' },
2687 # yields: WHERE ( ( col LIKE ? AND col LIKE ? ) )
2690 { col => { -like => 'foo%' } },
2691 { col => { -like => '%bar' } },
2693 # yields: WHERE ( ( col LIKE ? OR col LIKE ? ) )
2696 =head2 Literal SQL and value type operators
2698 The basic premise of SQL::Abstract is that in WHERE specifications the "left
2699 side" is a column name and the "right side" is a value (normally rendered as
2700 a placeholder). This holds true for both hashrefs and arrayref pairs as you
2701 see in the L</WHERE CLAUSES> examples above. Sometimes it is necessary to
2702 alter this behavior. There are several ways of doing so.
2706 This is a virtual operator that signals the string to its right side is an
2707 identifier (a column name) and not a value. For example to compare two
2708 columns you would write:
2711 priority => { '<', 2 },
2712 requestor => { -ident => 'submitter' },
2717 $stmt = "WHERE priority < ? AND requestor = submitter";
2720 If you are maintaining legacy code you may see a different construct as
2721 described in L</Deprecated usage of Literal SQL>, please use C<-ident> in new
2726 This is a virtual operator that signals that the construct to its right side
2727 is a value to be passed to DBI. This is for example necessary when you want
2728 to write a where clause against an array (for RDBMS that support such
2729 datatypes). For example:
2732 array => { -value => [1, 2, 3] }
2737 $stmt = 'WHERE array = ?';
2738 @bind = ([1, 2, 3]);
2740 Note that if you were to simply say:
2746 the result would probably not be what you wanted:
2748 $stmt = 'WHERE array = ? OR array = ? OR array = ?';
2753 Finally, sometimes only literal SQL will do. To include a random snippet
2754 of SQL verbatim, you specify it as a scalar reference. Consider this only
2755 as a last resort. Usually there is a better way. For example:
2758 priority => { '<', 2 },
2759 requestor => { -in => \'(SELECT name FROM hitmen)' },
2764 $stmt = "WHERE priority < ? AND requestor IN (SELECT name FROM hitmen)"
2767 Note that in this example, you only get one bind parameter back, since
2768 the verbatim SQL is passed as part of the statement.
2772 Never use untrusted input as a literal SQL argument - this is a massive
2773 security risk (there is no way to check literal snippets for SQL
2774 injections and other nastyness). If you need to deal with untrusted input
2775 use literal SQL with placeholders as described next.
2777 =head3 Literal SQL with placeholders and bind values (subqueries)
2779 If the literal SQL to be inserted has placeholders and bind values,
2780 use a reference to an arrayref (yes this is a double reference --
2781 not so common, but perfectly legal Perl). For example, to find a date
2782 in Postgres you can use something like this:
2785 date_column => \[ "= date '2008-09-30' - ?::integer", 10 ]
2790 $stmt = "WHERE ( date_column = date '2008-09-30' - ?::integer )"
2793 Note that you must pass the bind values in the same format as they are returned
2794 by L<where|/where(\%where, $order)>. This means that if you set L</bindtype>
2795 to C<columns>, you must provide the bind values in the
2796 C<< [ column_meta => value ] >> format, where C<column_meta> is an opaque
2797 scalar value; most commonly the column name, but you can use any scalar value
2798 (including references and blessed references), L<SQL::Abstract> will simply
2799 pass it through intact. So if C<bindtype> is set to C<columns> the above
2800 example will look like:
2803 date_column => \[ "= date '2008-09-30' - ?::integer", [ {} => 10 ] ]
2806 Literal SQL is especially useful for nesting parenthesized clauses in the
2807 main SQL query. Here is a first example:
2809 my ($sub_stmt, @sub_bind) = ("SELECT c1 FROM t1 WHERE c2 < ? AND c3 LIKE ?",
2813 bar => \["IN ($sub_stmt)" => @sub_bind],
2818 $stmt = "WHERE (foo = ? AND bar IN (SELECT c1 FROM t1
2819 WHERE c2 < ? AND c3 LIKE ?))";
2820 @bind = (1234, 100, "foo%");
2822 Other subquery operators, like for example C<"E<gt> ALL"> or C<"NOT IN">,
2823 are expressed in the same way. Of course the C<$sub_stmt> and
2824 its associated bind values can be generated through a former call
2827 my ($sub_stmt, @sub_bind)
2828 = $sql->select("t1", "c1", {c2 => {"<" => 100},
2829 c3 => {-like => "foo%"}});
2832 bar => \["> ALL ($sub_stmt)" => @sub_bind],
2835 In the examples above, the subquery was used as an operator on a column;
2836 but the same principle also applies for a clause within the main C<%where>
2837 hash, like an EXISTS subquery:
2839 my ($sub_stmt, @sub_bind)
2840 = $sql->select("t1", "*", {c1 => 1, c2 => \"> t0.c0"});
2841 my %where = ( -and => [
2843 \["EXISTS ($sub_stmt)" => @sub_bind],
2848 $stmt = "WHERE (foo = ? AND EXISTS (SELECT * FROM t1
2849 WHERE c1 = ? AND c2 > t0.c0))";
2853 Observe that the condition on C<c2> in the subquery refers to
2854 column C<t0.c0> of the main query: this is I<not> a bind
2855 value, so we have to express it through a scalar ref.
2856 Writing C<< c2 => {">" => "t0.c0"} >> would have generated
2857 C<< c2 > ? >> with bind value C<"t0.c0"> ... not exactly
2858 what we wanted here.
2860 Finally, here is an example where a subquery is used
2861 for expressing unary negation:
2863 my ($sub_stmt, @sub_bind)
2864 = $sql->where({age => [{"<" => 10}, {">" => 20}]});
2865 $sub_stmt =~ s/^ where //i; # don't want "WHERE" in the subclause
2867 lname => {like => '%son%'},
2868 \["NOT ($sub_stmt)" => @sub_bind],
2873 $stmt = "lname LIKE ? AND NOT ( age < ? OR age > ? )"
2874 @bind = ('%son%', 10, 20)
2876 =head3 Deprecated usage of Literal SQL
2878 Below are some examples of archaic use of literal SQL. It is shown only as
2879 reference for those who deal with legacy code. Each example has a much
2880 better, cleaner and safer alternative that users should opt for in new code.
2886 my %where = ( requestor => \'IS NOT NULL' )
2888 $stmt = "WHERE requestor IS NOT NULL"
2890 This used to be the way of generating NULL comparisons, before the handling
2891 of C<undef> got formalized. For new code please use the superior syntax as
2892 described in L</Tests for NULL values>.
2896 my %where = ( requestor => \'= submitter' )
2898 $stmt = "WHERE requestor = submitter"
2900 This used to be the only way to compare columns. Use the superior L</-ident>
2901 method for all new code. For example an identifier declared in such a way
2902 will be properly quoted if L</quote_char> is properly set, while the legacy
2903 form will remain as supplied.
2907 my %where = ( is_ready => \"", completed => { '>', '2012-12-21' } )
2909 $stmt = "WHERE completed > ? AND is_ready"
2910 @bind = ('2012-12-21')
2912 Using an empty string literal used to be the only way to express a boolean.
2913 For all new code please use the much more readable
2914 L<-bool|/Unary operators: bool> operator.
2920 These pages could go on for a while, since the nesting of the data
2921 structures this module can handle are pretty much unlimited (the
2922 module implements the C<WHERE> expansion as a recursive function
2923 internally). Your best bet is to "play around" with the module a
2924 little to see how the data structures behave, and choose the best
2925 format for your data based on that.
2927 And of course, all the values above will probably be replaced with
2928 variables gotten from forms or the command line. After all, if you
2929 knew everything ahead of time, you wouldn't have to worry about
2930 dynamically-generating SQL and could just hardwire it into your
2933 =head1 ORDER BY CLAUSES
2935 Some functions take an order by clause. This can either be a scalar (just a
2936 column name), a hashref of C<< { -desc => 'col' } >> or C<< { -asc => 'col' }
2937 >>, a scalarref, an arrayref-ref, or an arrayref of any of the previous
2940 Given | Will Generate
2941 ---------------------------------------------------------------
2943 'colA' | ORDER BY colA
2945 [qw/colA colB/] | ORDER BY colA, colB
2947 {-asc => 'colA'} | ORDER BY colA ASC
2949 {-desc => 'colB'} | ORDER BY colB DESC
2951 ['colA', {-asc => 'colB'}] | ORDER BY colA, colB ASC
2953 { -asc => [qw/colA colB/] } | ORDER BY colA ASC, colB ASC
2955 \'colA DESC' | ORDER BY colA DESC
2957 \[ 'FUNC(colA, ?)', $x ] | ORDER BY FUNC(colA, ?)
2958 | /* ...with $x bound to ? */
2961 { -asc => 'colA' }, | colA ASC,
2962 { -desc => [qw/colB/] }, | colB DESC,
2963 { -asc => [qw/colC colD/] },| colC ASC, colD ASC,
2964 \'colE DESC', | colE DESC,
2965 \[ 'FUNC(colF, ?)', $x ], | FUNC(colF, ?)
2966 ] | /* ...with $x bound to ? */
2967 ===============================================================
2971 =head1 SPECIAL OPERATORS
2973 my $sqlmaker = SQL::Abstract->new(special_ops => [
2977 my ($self, $field, $op, $arg) = @_;
2983 handler => 'method_name',
2987 A "special operator" is a SQL syntactic clause that can be
2988 applied to a field, instead of a usual binary operator.
2991 WHERE field IN (?, ?, ?)
2992 WHERE field BETWEEN ? AND ?
2993 WHERE MATCH(field) AGAINST (?, ?)
2995 Special operators IN and BETWEEN are fairly standard and therefore
2996 are builtin within C<SQL::Abstract> (as the overridable methods
2997 C<_where_field_IN> and C<_where_field_BETWEEN>). For other operators,
2998 like the MATCH .. AGAINST example above which is specific to MySQL,
2999 you can write your own operator handlers - supply a C<special_ops>
3000 argument to the C<new> method. That argument takes an arrayref of
3001 operator definitions; each operator definition is a hashref with two
3008 the regular expression to match the operator
3012 Either a coderef or a plain scalar method name. In both cases
3013 the expected return is C<< ($sql, @bind) >>.
3015 When supplied with a method name, it is simply called on the
3016 L<SQL::Abstract> object as:
3018 $self->$method_name($field, $op, $arg)
3022 $field is the LHS of the operator
3023 $op is the part that matched the handler regex
3026 When supplied with a coderef, it is called as:
3028 $coderef->($self, $field, $op, $arg)
3033 For example, here is an implementation
3034 of the MATCH .. AGAINST syntax for MySQL
3036 my $sqlmaker = SQL::Abstract->new(special_ops => [
3038 # special op for MySql MATCH (field) AGAINST(word1, word2, ...)
3039 {regex => qr/^match$/i,
3041 my ($self, $field, $op, $arg) = @_;
3042 $arg = [$arg] if not ref $arg;
3043 my $label = $self->_quote($field);
3044 my ($placeholder) = $self->_convert('?');
3045 my $placeholders = join ", ", (($placeholder) x @$arg);
3046 my $sql = $self->_sqlcase('match') . " ($label) "
3047 . $self->_sqlcase('against') . " ($placeholders) ";
3048 my @bind = $self->_bindtype($field, @$arg);
3049 return ($sql, @bind);
3056 =head1 UNARY OPERATORS
3058 my $sqlmaker = SQL::Abstract->new(unary_ops => [
3062 my ($self, $op, $arg) = @_;
3068 handler => 'method_name',
3072 A "unary operator" is a SQL syntactic clause that can be
3073 applied to a field - the operator goes before the field
3075 You can write your own operator handlers - supply a C<unary_ops>
3076 argument to the C<new> method. That argument takes an arrayref of
3077 operator definitions; each operator definition is a hashref with two
3084 the regular expression to match the operator
3088 Either a coderef or a plain scalar method name. In both cases
3089 the expected return is C<< $sql >>.
3091 When supplied with a method name, it is simply called on the
3092 L<SQL::Abstract> object as:
3094 $self->$method_name($op, $arg)
3098 $op is the part that matched the handler regex
3099 $arg is the RHS or argument of the operator
3101 When supplied with a coderef, it is called as:
3103 $coderef->($self, $op, $arg)
3111 Thanks to some benchmarking by Mark Stosberg, it turns out that
3112 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
3113 I must admit this wasn't an intentional design issue, but it's a
3114 byproduct of the fact that you get to control your C<DBI> handles
3117 To maximize performance, use a code snippet like the following:
3119 # prepare a statement handle using the first row
3120 # and then reuse it for the rest of the rows
3122 for my $href (@array_of_hashrefs) {
3123 $stmt ||= $sql->insert('table', $href);
3124 $sth ||= $dbh->prepare($stmt);
3125 $sth->execute($sql->values($href));
3128 The reason this works is because the keys in your C<$href> are sorted
3129 internally by B<SQL::Abstract>. Thus, as long as your data retains
3130 the same structure, you only have to generate the SQL the first time
3131 around. On subsequent queries, simply use the C<values> function provided
3132 by this module to return your values in the correct order.
3134 However this depends on the values having the same type - if, for
3135 example, the values of a where clause may either have values
3136 (resulting in sql of the form C<column = ?> with a single bind
3137 value), or alternatively the values might be C<undef> (resulting in
3138 sql of the form C<column IS NULL> with no bind value) then the
3139 caching technique suggested will not work.
3143 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
3144 really like this part (I do, at least). Building up a complex query
3145 can be as simple as the following:
3152 use CGI::FormBuilder;
3155 my $form = CGI::FormBuilder->new(...);
3156 my $sql = SQL::Abstract->new;
3158 if ($form->submitted) {
3159 my $field = $form->field;
3160 my $id = delete $field->{id};
3161 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
3164 Of course, you would still have to connect using C<DBI> to run the
3165 query, but the point is that if you make your form look like your
3166 table, the actual query script can be extremely simplistic.
3168 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
3169 a fast interface to returning and formatting data. I frequently
3170 use these three modules together to write complex database query
3171 apps in under 50 lines.
3173 =head1 HOW TO CONTRIBUTE
3175 Contributions are always welcome, in all usable forms (we especially
3176 welcome documentation improvements). The delivery methods include git-
3177 or unified-diff formatted patches, GitHub pull requests, or plain bug
3178 reports either via RT or the Mailing list. Contributors are generally
3179 granted full access to the official repository after their first several
3180 patches pass successful review.
3182 This project is maintained in a git repository. The code and related tools are
3183 accessible at the following locations:
3187 =item * Official repo: L<git://git.shadowcat.co.uk/dbsrgits/SQL-Abstract.git>
3189 =item * Official gitweb: L<http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=dbsrgits/SQL-Abstract.git>
3191 =item * GitHub mirror: L<https://github.com/dbsrgits/sql-abstract>
3193 =item * Authorized committers: L<ssh://dbsrgits@git.shadowcat.co.uk/SQL-Abstract.git>
3199 Version 1.50 was a major internal refactoring of C<SQL::Abstract>.
3200 Great care has been taken to preserve the I<published> behavior
3201 documented in previous versions in the 1.* family; however,
3202 some features that were previously undocumented, or behaved
3203 differently from the documentation, had to be changed in order
3204 to clarify the semantics. Hence, client code that was relying
3205 on some dark areas of C<SQL::Abstract> v1.*
3206 B<might behave differently> in v1.50.
3208 The main changes are:
3214 support for literal SQL through the C<< \ [ $sql, @bind ] >> syntax.
3218 support for the { operator => \"..." } construct (to embed literal SQL)
3222 support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values)
3226 optional support for L<array datatypes|/"Inserting and Updating Arrays">
3230 defensive programming: check arguments
3234 fixed bug with global logic, which was previously implemented
3235 through global variables yielding side-effects. Prior versions would
3236 interpret C<< [ {cond1, cond2}, [cond3, cond4] ] >>
3237 as C<< "(cond1 AND cond2) OR (cond3 AND cond4)" >>.
3238 Now this is interpreted
3239 as C<< "(cond1 AND cond2) OR (cond3 OR cond4)" >>.
3244 fixed semantics of _bindtype on array args
3248 dropped the C<_anoncopy> of the %where tree. No longer necessary,
3249 we just avoid shifting arrays within that tree.
3253 dropped the C<_modlogic> function
3257 =head1 ACKNOWLEDGEMENTS
3259 There are a number of individuals that have really helped out with
3260 this module. Unfortunately, most of them submitted bugs via CPAN
3261 so I have no idea who they are! But the people I do know are:
3263 Ash Berlin (order_by hash term support)
3264 Matt Trout (DBIx::Class support)
3265 Mark Stosberg (benchmarking)
3266 Chas Owens (initial "IN" operator support)
3267 Philip Collins (per-field SQL functions)
3268 Eric Kolve (hashref "AND" support)
3269 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
3270 Dan Kubb (support for "quote_char" and "name_sep")
3271 Guillermo Roditi (patch to cleanup "IN" and "BETWEEN", fix and tests for _order_by)
3272 Laurent Dami (internal refactoring, extensible list of special operators, literal SQL)
3273 Norbert Buchmuller (support for literal SQL in hashpair, misc. fixes & tests)
3274 Peter Rabbitson (rewrite of SQLA::Test, misc. fixes & tests)
3275 Oliver Charles (support for "RETURNING" after "INSERT")
3281 L<DBIx::Class>, L<DBIx::Abstract>, L<CGI::FormBuilder>, L<HTML::QuickTable>.
3285 Copyright (c) 2001-2007 Nathan Wiger <nwiger@cpan.org>. All Rights Reserved.
3287 This module is actively maintained by Matt Trout <mst@shadowcatsystems.co.uk>
3289 For support, your best bet is to try the C<DBIx::Class> users mailing list.
3290 While not an official support venue, C<DBIx::Class> makes heavy use of
3291 C<SQL::Abstract>, and as such list members there are very familiar with
3292 how to create queries.
3296 This module is free software; you may copy this under the same
3297 terms as perl itself (either the GNU General Public License or
3298 the Artistic License)