1 package SQL::Abstract; # see doc at end of file
10 our @EXPORT_OK = qw(is_plain_value is_literal_value);
20 *SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION = $ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}
26 #======================================================================
28 #======================================================================
30 our $VERSION = '1.86';
32 # This would confuse some packagers
33 $VERSION = eval $VERSION if $VERSION =~ /_/; # numify for warning-free dev releases
37 # special operators (-in, -between). May be extended/overridden by user.
38 # See section WHERE: BUILTIN SPECIAL OPERATORS below for implementation
39 my @BUILTIN_SPECIAL_OPS = (
40 {regex => qr/^ (?: not \s )? between $/ix, handler => sub { die "NOPE" }},
41 {regex => qr/^ is (?: \s+ not )? $/ix, handler => sub { die "NOPE" }},
42 {regex => qr/^ (?: not \s )? in $/ix, handler => sub { die "NOPE" }},
43 {regex => qr/^ ident $/ix, handler => sub { die "NOPE" }},
44 {regex => qr/^ value $/ix, handler => sub { die "NOPE" }},
47 #======================================================================
48 # DEBUGGING AND ERROR REPORTING
49 #======================================================================
52 return unless $_[0]->{debug}; shift; # a little faster
53 my $func = (caller(1))[3];
54 warn "[$func] ", @_, "\n";
58 my($func) = (caller(1))[3];
59 Carp::carp "[$func] Warning: ", @_;
63 my($func) = (caller(1))[3];
64 Carp::croak "[$func] Fatal: ", @_;
67 sub is_literal_value ($) {
68 ref $_[0] eq 'SCALAR' ? [ ${$_[0]} ]
69 : ( ref $_[0] eq 'REF' and ref ${$_[0]} eq 'ARRAY' ) ? [ @${ $_[0] } ]
73 sub is_undef_value ($) {
77 and exists $_[0]->{-value}
78 and not defined $_[0]->{-value}
82 # FIXME XSify - this can be done so much more efficiently
83 sub is_plain_value ($) {
85 ! length ref $_[0] ? \($_[0])
87 ref $_[0] eq 'HASH' and keys %{$_[0]} == 1
89 exists $_[0]->{-value}
90 ) ? \($_[0]->{-value})
92 # reuse @_ for even moar speedz
93 defined ( $_[1] = Scalar::Util::blessed $_[0] )
95 # deliberately not using Devel::OverloadInfo - the checks we are
96 # intersted in are much more limited than the fullblown thing, and
97 # this is a very hot piece of code
99 # simply using ->can('(""') can leave behind stub methods that
100 # break actually using the overload later (see L<perldiag/Stub
101 # found while resolving method "%s" overloading "%s" in package
102 # "%s"> and the source of overload::mycan())
104 # either has stringification which DBI SHOULD prefer out of the box
105 grep { *{ (qq[${_}::(""]) }{CODE} } @{ $_[2] = mro::get_linear_isa( $_[1] ) }
107 # has nummification or boolification, AND fallback is *not* disabled
109 SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION
112 grep { *{"${_}::(0+"}{CODE} } @{$_[2]}
114 grep { *{"${_}::(bool"}{CODE} } @{$_[2]}
118 # no fallback specified at all
119 ! ( ($_[3]) = grep { *{"${_}::()"}{CODE} } @{$_[2]} )
121 # fallback explicitly undef
122 ! defined ${"$_[3]::()"}
135 #======================================================================
137 #======================================================================
141 my $class = ref($self) || $self;
142 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
144 # choose our case by keeping an option around
145 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
147 # default logic for interpreting arrayrefs
148 $opt{logic} = $opt{logic} ? uc $opt{logic} : 'OR';
150 # how to return bind vars
151 $opt{bindtype} ||= 'normal';
153 # default comparison is "=", but can be overridden
156 # try to recognize which are the 'equality' and 'inequality' ops
157 # (temporary quickfix (in 2007), should go through a more seasoned API)
158 $opt{equality_op} = qr/^( \Q$opt{cmp}\E | \= )$/ix;
159 $opt{inequality_op} = qr/^( != | <> )$/ix;
161 $opt{like_op} = qr/^ (is_)?r?like $/xi;
162 $opt{not_like_op} = qr/^ (is_)?not_r?like $/xi;
165 $opt{sqltrue} ||= '1=1';
166 $opt{sqlfalse} ||= '0=1';
169 $opt{special_ops} ||= [];
171 if ($class->isa('DBIx::Class::SQLMaker')) {
172 $opt{warn_once_on_nest} = 1;
173 $opt{disable_old_special_ops} = 1;
177 $opt{unary_ops} ||= [];
179 # rudimentary sanity-check for user supplied bits treated as functions/operators
180 # If a purported function matches this regular expression, an exception is thrown.
181 # Literal SQL is *NOT* subject to this check, only functions (and column names
182 # when quoting is not in effect)
185 # need to guard against ()'s in column names too, but this will break tons of
186 # hacks... ideas anyone?
187 $opt{injection_guard} ||= qr/
193 $opt{expand_unary} = {};
196 not => '_expand_not',
197 bool => '_expand_bool',
198 and => '_expand_op_andor',
199 or => '_expand_op_andor',
200 nest => '_expand_nest',
201 bind => '_expand_bind',
203 not_in => '_expand_in',
204 row => '_expand_row',
205 between => '_expand_between',
206 not_between => '_expand_between',
208 (map +($_ => '_expand_op_is'), ('is', 'is_not')),
209 ident => '_expand_ident',
210 value => '_expand_value',
214 'between' => '_expand_between',
215 'not_between' => '_expand_between',
216 'in' => '_expand_in',
217 'not_in' => '_expand_in',
218 'nest' => '_expand_nest',
219 (map +($_ => '_expand_op_andor'), ('and', 'or')),
220 (map +($_ => '_expand_op_is'), ('is', 'is_not')),
221 'ident' => '_expand_ident',
222 'value' => '_expand_value',
226 (map +($_, "_render_$_"), qw(op func bind ident literal row)),
231 (map +($_ => '_render_op_between'), 'between', 'not_between'),
232 (map +($_ => '_render_op_in'), 'in', 'not_in'),
233 (map +($_ => '_render_unop_postfix'),
234 'is_null', 'is_not_null', 'asc', 'desc',
236 (not => '_render_unop_paren'),
237 (map +($_ => '_render_op_andor'), qw(and or)),
238 ',' => '_render_op_multop',
241 return bless \%opt, $class;
244 sub sqltrue { +{ -literal => [ $_[0]->{sqltrue} ] } }
245 sub sqlfalse { +{ -literal => [ $_[0]->{sqlfalse} ] } }
247 sub _assert_pass_injection_guard {
248 if ($_[1] =~ $_[0]->{injection_guard}) {
249 my $class = ref $_[0];
250 puke "Possible SQL injection attempt '$_[1]'. If this is indeed a part of the "
251 . "desired SQL use literal SQL ( \'...' or \[ '...' ] ) or supply your own "
252 . "{injection_guard} attribute to ${class}->new()"
257 #======================================================================
259 #======================================================================
263 my $table = $self->_table(shift);
264 my $data = shift || return;
267 my ($sql, @bind) = do {
268 if (is_literal_value($data)) {
269 $self->render_expr($data);
271 my ($fields, $values) = (
272 ref($data) eq 'HASH' ?
273 ([ sort keys %$data ], [ @{$data}{sort keys %$data} ])
277 # no names (arrayref) means can't generate bindtype
278 !($fields) && $self->{bindtype} eq 'columns'
279 && belch "can't do 'columns' bindtype when called with arrayref";
281 my $fields_sql = ($fields
282 ? ' '.($self->render_expr({ -row => $fields }, '-ident'))[0]
285 my ($values_sql, @bind) = $self->render_aqt(
288 local our $Cur_Col_Meta = $fields->[$_];
289 $self->_expand_insert_value($values->[$_])
293 ($fields_sql.' '.$self->_sqlcase('values').' '.$values_sql, @bind);
297 $sql = (join " ", $self->_sqlcase('insert into'), $table).$sql;
299 if ($options->{returning}) {
300 my ($s, @b) = $self->_insert_returning($options);
305 return wantarray ? ($sql, @bind) : $sql;
308 # So that subclasses can override INSERT ... RETURNING separately from
309 # UPDATE and DELETE (e.g. DBIx::Class::SQLMaker::Oracle does this)
310 sub _insert_returning { shift->_returning(@_) }
313 my ($self, $options) = @_;
315 my $f = $options->{returning};
317 my ($sql, @bind) = $self->render_aqt(
318 $self->_expand_maybe_list_expr($f, -ident)
321 ? $self->_sqlcase(' returning ') . $sql
322 : ($self->_sqlcase(' returning ').$sql, @bind);
325 sub _expand_insert_value {
328 my $k = our $Cur_Col_Meta;
330 if (ref($v) eq 'ARRAY') {
331 if ($self->{array_datatypes}) {
332 return +{ -bind => [ $k, $v ] };
334 my ($sql, @bind) = @$v;
335 $self->_assert_bindval_matches_bindtype(@bind);
336 return +{ -literal => $v };
338 if (ref($v) eq 'HASH') {
339 if (grep !/^-/, keys %$v) {
340 belch "HASH ref as bind value in insert is not supported";
341 return +{ -bind => [ $k, $v ] };
345 return +{ -bind => [ $k, undef ] };
347 return $self->expand_expr($v);
352 #======================================================================
354 #======================================================================
359 my $table = $self->_table(shift);
360 my $data = shift || return;
364 # first build the 'SET' part of the sql statement
365 puke "Unsupported data type specified to \$sql->update"
366 unless ref $data eq 'HASH';
368 my ($sql, @all_bind) = $self->_update_set_values($data);
369 $sql = $self->_sqlcase('update ') . $table . $self->_sqlcase(' set ')
373 my($where_sql, @where_bind) = $self->where($where);
375 push @all_bind, @where_bind;
378 if ($options->{returning}) {
379 my ($returning_sql, @returning_bind) = $self->_update_returning($options);
380 $sql .= $returning_sql;
381 push @all_bind, @returning_bind;
384 return wantarray ? ($sql, @all_bind) : $sql;
387 sub _update_set_values {
388 my ($self, $data) = @_;
390 return $self->render_aqt(
391 $self->_expand_update_set_values($data),
395 sub _expand_update_set_values {
396 my ($self, $data) = @_;
397 $self->_expand_maybe_list_expr( [
400 $set = { -bind => $_ } unless defined $set;
401 +{ -op => [ '=', $self->_expand_ident(-ident => $k), $set ] };
407 ? ($self->{array_datatypes}
408 ? [ $k, +{ -bind => [ $k, $v ] } ]
409 : [ $k, +{ -literal => $v } ])
411 local our $Cur_Col_Meta = $k;
412 [ $k, $self->_expand_expr($v) ]
419 # So that subclasses can override UPDATE ... RETURNING separately from
421 sub _update_returning { shift->_returning(@_) }
425 #======================================================================
427 #======================================================================
432 my $table = $self->_table(shift);
433 my $fields = shift || '*';
437 my ($fields_sql, @bind) = $self->_select_fields($fields);
439 my ($where_sql, @where_bind) = $self->where($where, $order);
440 push @bind, @where_bind;
442 my $sql = join(' ', $self->_sqlcase('select'), $fields_sql,
443 $self->_sqlcase('from'), $table)
446 return wantarray ? ($sql, @bind) : $sql;
450 my ($self, $fields) = @_;
451 return $fields unless ref($fields);
452 return $self->render_aqt(
453 $self->_expand_maybe_list_expr($fields, '-ident')
457 #======================================================================
459 #======================================================================
464 my $table = $self->_table(shift);
468 my($where_sql, @bind) = $self->where($where);
469 my $sql = $self->_sqlcase('delete from ') . $table . $where_sql;
471 if ($options->{returning}) {
472 my ($returning_sql, @returning_bind) = $self->_delete_returning($options);
473 $sql .= $returning_sql;
474 push @bind, @returning_bind;
477 return wantarray ? ($sql, @bind) : $sql;
480 # So that subclasses can override DELETE ... RETURNING separately from
482 sub _delete_returning { shift->_returning(@_) }
486 #======================================================================
488 #======================================================================
492 # Finally, a separate routine just to handle WHERE clauses
494 my ($self, $where, $order) = @_;
496 local $self->{convert_where} = $self->{convert};
499 my ($sql, @bind) = defined($where)
500 ? $self->_recurse_where($where)
502 $sql = (defined $sql and length $sql) ? $self->_sqlcase(' where ') . "( $sql )" : '';
506 my ($order_sql, @order_bind) = $self->_order_by($order);
508 push @bind, @order_bind;
511 return wantarray ? ($sql, @bind) : $sql;
514 { our $Default_Scalar_To = -value }
517 my ($self, $expr, $default_scalar_to) = @_;
518 local our $Default_Scalar_To = $default_scalar_to if $default_scalar_to;
519 $self->_expand_expr($expr);
523 my ($self, $aqt) = @_;
524 my ($k, $v, @rest) = %$aqt;
526 die "Not a node type: $k" unless $k =~ s/^-//;
527 if (my $meth = $self->{render}{$k}) {
528 return $self->$meth($v);
530 die "notreached: $k";
534 my ($self, $expr, $default_scalar_to) = @_;
535 $self->render_aqt($self->expand_expr($expr, $default_scalar_to));
539 my ($self, $raw) = @_;
540 s/^-(?=.)//, s/\s+/_/g for my $op = lc $raw;
545 my ($self, $expr) = @_;
546 our $Expand_Depth ||= 0; local $Expand_Depth = $Expand_Depth + 1;
547 return undef unless defined($expr);
548 if (ref($expr) eq 'HASH') {
549 return undef unless my $kc = keys %$expr;
551 return $self->_expand_op_andor(and => $expr);
553 my ($key, $value) = %$expr;
554 if ($key =~ /^-/ and $key =~ s/ [_\s]? \d+ $//x ) {
555 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
556 . "You probably wanted ...-and => [ $key => COND1, $key => COND2 ... ]";
558 return $self->_expand_hashpair($key, $value);
560 if (ref($expr) eq 'ARRAY') {
561 return $self->_expand_op_andor(lc($self->{logic}), $expr);
563 if (my $literal = is_literal_value($expr)) {
564 return +{ -literal => $literal };
566 if (!ref($expr) or Scalar::Util::blessed($expr)) {
567 return $self->_expand_scalar($expr);
572 sub _expand_hashpair {
573 my ($self, $k, $v) = @_;
574 unless (defined($k) and length($k)) {
575 if (defined($k) and my $literal = is_literal_value($v)) {
576 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
577 return { -literal => $literal };
579 puke "Supplying an empty left hand side argument is not supported";
582 return $self->_expand_hashpair_op($k, $v);
584 return $self->_expand_hashpair_ident($k, $v);
587 sub _expand_hashpair_ident {
588 my ($self, $k, $v) = @_;
590 local our $Cur_Col_Meta = $k;
592 # hash with multiple or no elements is andor
594 if (ref($v) eq 'HASH' and keys %$v != 1) {
595 return $self->_expand_op_andor(and => $v, $k);
598 # undef needs to be re-sent with cmp to achieve IS/IS NOT NULL
600 if (is_undef_value($v)) {
601 return $self->_expand_hashpair_cmp($k => undef);
604 # scalars and objects get expanded as whatever requested or values
606 if (!ref($v) or Scalar::Util::blessed($v)) {
607 return $self->_expand_hashpair_scalar($k, $v);
610 # single key hashref is a hashtriple
612 if (ref($v) eq 'HASH') {
613 return $self->_expand_hashtriple($k, %$v);
616 # arrayref needs re-engineering over the elements
618 if (ref($v) eq 'ARRAY') {
619 return $self->sqlfalse unless @$v;
620 $self->_debug("ARRAY($k) means distribute over elements");
622 $v->[0] =~ /^-(and|or)$/i
623 ? (shift(@{$v = [ @$v ]}), $1)
624 : lc($self->{logic} || 'OR')
626 return $self->_expand_op_andor(
631 if (my $literal = is_literal_value($v)) {
633 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
636 my ($sql, @bind) = @$literal;
637 if ($self->{bindtype} eq 'columns') {
639 $self->_assert_bindval_matches_bindtype($_);
642 return +{ -literal => [ $self->_quote($k).' '.$sql, @bind ] };
648 my ($self, $expr) = @_;
650 return $self->_expand_expr({ (our $Default_Scalar_To) => $expr });
653 sub _expand_hashpair_scalar {
654 my ($self, $k, $v) = @_;
656 return $self->_expand_hashpair_cmp(
657 $k, $self->_expand_scalar($v),
661 sub _expand_hashpair_op {
662 my ($self, $k, $v) = @_;
664 $self->_assert_pass_injection_guard($k =~ /\A-(.*)\Z/s);
666 my $op = $self->_normalize_op($k);
668 if (my $exp = $self->{expand}{$op}) {
669 return $self->$exp($op, $v);
672 # Ops prefixed with -not_ get converted
674 if (my ($rest) = $op =~/^not_(.*)$/) {
677 $self->_expand_expr({ "-${rest}", $v })
683 my $op = join(' ', split '_', $op);
685 # the old special op system requires illegality for top-level use
688 (our $Expand_Depth) == 1
690 List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}
692 $self->{disable_old_special_ops}
693 and List::Util::first { $op =~ $_->{regex} } @BUILTIN_SPECIAL_OPS
697 puke "Illegal use of top-level '-$op'"
700 # the old unary op system means we should touch nothing and let it work
702 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
703 return { -op => [ $op, $v ] };
707 # an explicit node type is currently assumed to be expanded (this is almost
708 # certainly wrong and there should be expansion anyway)
710 if ($self->{render}{$op}) {
714 # hashref RHS values get expanded and used as op/func args
719 and (keys %$v)[0] =~ /^-/
721 my ($func) = $k =~ /^-(.*)$/;
723 if (List::Util::first { $func =~ $_->{regex} } @{$self->{special_ops}}) {
724 return +{ -op => [ $func, $self->_expand_expr($v) ] };
727 return +{ -func => [ $func, $self->_expand_expr($v) ] };
730 # scalars and literals get simply expanded
732 if (!ref($v) or is_literal_value($v)) {
733 return +{ -op => [ $op, $self->_expand_expr($v) ] };
739 sub _expand_hashpair_cmp {
740 my ($self, $k, $v) = @_;
741 $self->_expand_hashtriple($k, $self->{cmp}, $v);
744 sub _expand_hashtriple {
745 my ($self, $k, $vk, $vv) = @_;
747 my $ik = $self->_expand_ident(-ident => $k);
749 my $op = $self->_normalize_op($vk);
750 $self->_assert_pass_injection_guard($op);
752 if ($op =~ s/ _? \d+ $//x ) {
753 return $self->_expand_expr($k, { $vk, $vv });
755 if (my $x = $self->{expand_op}{$op}) {
756 local our $Cur_Col_Meta = $k;
757 return $self->$x($op, $vv, $k);
761 my $op = join(' ', split '_', $op);
763 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}) {
764 return { -op => [ $op, $ik, $vv ] };
766 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
770 { -op => [ $op, $vv ] }
774 if (ref($vv) eq 'ARRAY') {
776 my $logic = (defined($raw[0]) and $raw[0] =~ /^-(and|or)$/i)
777 ? (shift(@raw), $1) : 'or';
778 my @values = map +{ $vk => $_ }, @raw;
780 $op =~ $self->{inequality_op}
781 or $op =~ $self->{not_like_op}
783 if (lc($logic) eq 'or' and @values > 1) {
784 belch "A multi-element arrayref as an argument to the inequality op '${\uc(join ' ', split '_', $op)}' "
785 . 'is technically equivalent to an always-true 1=1 (you probably wanted '
786 . "to say ...{ \$inequality_op => [ -and => \@values ] }... instead)"
791 # try to DWIM on equality operators
792 return ($self->_dwim_op_to_is($op,
793 "Supplying an empty arrayref to '%s' is deprecated",
794 "operator '%s' applied on an empty array (field '$k')"
795 ) ? $self->sqlfalse : $self->sqltrue);
797 return $self->_expand_op_andor($logic => \@values, $k);
799 if (is_undef_value($vv)) {
800 my $is = ($self->_dwim_op_to_is($op,
801 "Supplying an undefined argument to '%s' is deprecated",
802 "unexpected operator '%s' with undef operand",
803 ) ? 'is' : 'is not');
805 return $self->_expand_hashpair($k => { $is, undef });
807 local our $Cur_Col_Meta = $k;
811 $self->_expand_expr($vv)
816 my ($self, $raw, $empty, $fail) = @_;
818 my $op = $self->_normalize_op($raw);
820 if ($op =~ /^not$/i) {
823 if ($op =~ $self->{equality_op}) {
826 if ($op =~ $self->{like_op}) {
827 belch(sprintf $empty, uc(join ' ', split '_', $op));
830 if ($op =~ $self->{inequality_op}) {
833 if ($op =~ $self->{not_like_op}) {
834 belch(sprintf $empty, uc(join ' ', split '_', $op));
837 puke(sprintf $fail, $op);
841 my ($self, undef, $body, $k) = @_;
842 return $self->_expand_hashpair_cmp(
843 $k, { -ident => $body }
845 unless (defined($body) or (ref($body) and ref($body) eq 'ARRAY')) {
846 puke "-ident requires a single plain scalar argument (a quotable identifier) or an arrayref of identifier parts";
848 my @parts = map split(/\Q${\($self->{name_sep}||'.')}\E/, $_),
849 ref($body) ? @$body : $body;
850 return { -ident => $parts[-1] } if $self->{_dequalify_idents};
851 unless ($self->{quote_char}) {
852 $self->_assert_pass_injection_guard($_) for @parts;
854 return +{ -ident => \@parts };
858 return $_[0]->_expand_hashpair_cmp(
859 $_[3], { -value => $_[2] },
861 +{ -bind => [ our $Cur_Col_Meta, $_[2] ] };
865 +{ -op => [ 'not', $_[0]->_expand_expr($_[2]) ] };
869 my ($self, undef, $args) = @_;
870 +{ -row => [ map $self->expand_expr($_), @$args ] };
874 my ($self, undef, $args) = @_;
875 my ($op, @opargs) = @$args;
876 if (my $exp = $self->{expand_op}{$op}) {
877 return $self->$exp($op, \@opargs);
879 +{ -op => [ $op, map $self->expand_expr($_), @opargs ] };
883 my ($self, undef, $v) = @_;
885 return $self->_expand_expr($v);
887 puke "-bool => undef not supported" unless defined($v);
888 return $self->_expand_ident(-ident => $v);
891 sub _expand_op_andor {
892 my ($self, $logop, $v, $k) = @_;
894 $v = [ map +{ $k, $_ },
896 ? (map +{ $_ => $v->{$_} }, sort keys %$v)
900 if (ref($v) eq 'HASH') {
901 return undef unless keys %$v;
904 map $self->_expand_expr({ $_ => $v->{$_} }),
908 if (ref($v) eq 'ARRAY') {
909 $logop eq 'and' or $logop eq 'or' or puke "unknown logic: $logop";
912 (ref($_) eq 'ARRAY' and @$_)
913 or (ref($_) eq 'HASH' and %$_)
919 while (my ($el) = splice @expr, 0, 1) {
920 puke "Supplying an empty left hand side argument is not supported in array-pairs"
921 unless defined($el) and length($el);
922 my $elref = ref($el);
924 local our $Expand_Depth = 0;
925 push(@res, grep defined, $self->_expand_expr({ $el, shift(@expr) }));
926 } elsif ($elref eq 'ARRAY') {
927 push(@res, grep defined, $self->_expand_expr($el)) if @$el;
928 } elsif (my $l = is_literal_value($el)) {
929 push @res, { -literal => $l };
930 } elsif ($elref eq 'HASH') {
931 local our $Expand_Depth = 0;
932 push @res, grep defined, $self->_expand_expr($el) if %$el;
938 # return $res[0] if @res == 1;
939 return { -op => [ $logop, @res ] };
945 my ($self, $op, $vv, $k) = @_;
946 ($k, $vv) = @$vv unless defined $k;
947 puke "$op can only take undef as argument"
951 and exists($vv->{-value})
952 and !defined($vv->{-value})
954 return +{ -op => [ $op.'_null', $self->expand_expr($k, -ident) ] };
957 sub _expand_between {
958 my ($self, $op, $vv, $k) = @_;
959 $k = shift @{$vv = [ @$vv ]} unless defined $k;
960 my @rhs = map $self->_expand_expr($_),
961 ref($vv) eq 'ARRAY' ? @$vv : $vv;
963 (@rhs == 1 and ref($rhs[0]) eq 'HASH' and $rhs[0]->{-literal})
965 (@rhs == 2 and defined($rhs[0]) and defined($rhs[1]))
967 puke "Operator '${\uc($op)}' requires either an arrayref with two defined values or expressions, or a single literal scalarref/arrayref-ref";
971 $self->expand_expr(ref($k) ? $k : { -ident => $k }),
977 my ($self, $op, $vv, $k) = @_;
978 $k = shift @{$vv = [ @$vv ]} unless defined $k;
979 if (my $literal = is_literal_value($vv)) {
980 my ($sql, @bind) = @$literal;
981 my $opened_sql = $self->_open_outer_paren($sql);
983 $op, $self->expand_expr($k, -ident),
984 { -literal => [ $opened_sql, @bind ] }
988 'SQL::Abstract before v1.75 used to generate incorrect SQL when the '
989 . "-${\uc($op)} operator was given an undef-containing list: !!!AUDIT YOUR CODE "
990 . 'AND DATA!!! (the upcoming Data::Query-based version of SQL::Abstract '
991 . 'will emit the logically correct SQL instead of raising this exception)'
993 puke("Argument passed to the '${\uc($op)}' operator can not be undefined")
995 my @rhs = map $self->expand_expr($_, -value),
996 map { defined($_) ? $_: puke($undef_err) }
997 (ref($vv) eq 'ARRAY' ? @$vv : $vv);
998 return $self->${\($op =~ /^not/ ? 'sqltrue' : 'sqlfalse')} unless @rhs;
1002 $self->expand_expr($k, -ident),
1008 my ($self, undef, $v) = @_;
1009 # DBIx::Class requires a nest warning to be emitted once but the private
1010 # method it overrode to do so no longer exists
1011 if ($self->{warn_once_on_nest}) {
1012 unless (our $Nest_Warned) {
1014 "-nest in search conditions is deprecated, you most probably wanted:\n"
1015 .q|{..., -and => [ \%cond0, \@cond1, \'cond2', \[ 'cond3', [ col => bind ] ], etc. ], ... }|
1020 return $self->_expand_expr($v);
1024 my ($self, undef, $bind) = @_;
1025 return { -bind => $bind };
1028 sub _recurse_where {
1029 my ($self, $where, $logic) = @_;
1031 # Special case: top level simple string treated as literal
1033 my $where_exp = (ref($where)
1034 ? $self->_expand_expr($where, $logic)
1035 : { -literal => [ $where ] });
1037 # dispatch expanded expression
1039 my ($sql, @bind) = defined($where_exp) ? $self->render_aqt($where_exp) : (undef);
1040 # DBIx::Class used to call _recurse_where in scalar context
1041 # something else might too...
1043 return ($sql, @bind);
1046 belch "Calling _recurse_where in scalar context is deprecated and will go away before 2.0";
1052 my ($self, $ident) = @_;
1054 return $self->_convert($self->_quote($ident));
1058 my ($self, $values) = @_;
1059 my ($sql, @bind) = $self->_render_op([ ',', @$values ]);
1060 return "($sql)", @bind;
1064 my ($self, $rest) = @_;
1065 my ($func, @args) = @$rest;
1069 push @arg_sql, shift @x;
1071 } map [ $self->render_aqt($_) ], @args;
1072 return ($self->_sqlcase($func).'('.join(', ', @arg_sql).')', @bind);
1076 my ($self, $bind) = @_;
1077 return ($self->_convert('?'), $self->_bindtype(@$bind));
1080 sub _render_literal {
1081 my ($self, $literal) = @_;
1082 $self->_assert_bindval_matches_bindtype(@{$literal}[1..$#$literal]);
1087 my ($self, $v) = @_;
1088 my ($op, @args) = @$v;
1089 if (my $r = $self->{render_op}{$op}) {
1090 return $self->$r($op, \@args);
1095 my $op = join(' ', split '_', $op);
1097 my $ss = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}};
1098 if ($ss and @args > 1) {
1099 puke "Special op '${op}' requires first value to be identifier"
1100 unless my ($ident) = map $_->{-ident}, grep ref($_) eq 'HASH', $args[0];
1101 my $k = join(($self->{name_sep}||'.'), @$ident);
1102 local our $Expand_Depth = 1;
1103 return $self->${\($ss->{handler})}($k, $op, $args[1]);
1105 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1106 return $self->${\($us->{handler})}($op, $args[0]);
1109 return $self->_render_unop_paren($op, \@args);
1113 return $self->_render_unop_prefix($op, \@args);
1115 return $self->_render_op_multop($op, \@args);
1121 sub _render_op_between {
1122 my ($self, $op, $args) = @_;
1123 my ($left, $low, $high) = @$args;
1124 my ($rhsql, @rhbind) = do {
1126 puke "Single arg to between must be a literal"
1127 unless $low->{-literal};
1130 my ($l, $h) = map [ $self->render_aqt($_) ], $low, $high;
1131 (join(' ', $l->[0], $self->_sqlcase('and'), $h->[0]),
1132 @{$l}[1..$#$l], @{$h}[1..$#$h])
1135 my ($lhsql, @lhbind) = $self->render_aqt($left);
1139 $self->_sqlcase(join ' ', split '_', $op),
1147 my ($self, $op, $args) = @_;
1148 my ($lhs, @rhs) = @$args;
1151 my ($sql, @bind) = $self->render_aqt($_);
1152 push @in_bind, @bind;
1155 my ($lhsql, @lbind) = $self->render_aqt($lhs);
1157 $lhsql.' '.$self->_sqlcase(join ' ', split '_', $op).' ( '
1158 .join(', ', @in_sql)
1164 sub _render_op_andor {
1165 my ($self, $op, $args) = @_;
1166 my @parts = grep length($_->[0]), map [ $self->render_aqt($_) ], @$args;
1167 return '' unless @parts;
1168 return @{$parts[0]} if @parts == 1;
1169 my ($sql, @bind) = $self->_join_parts(' '.$self->_sqlcase($op).' ', @parts);
1170 return '( '.$sql.' )', @bind;
1173 sub _render_op_multop {
1174 my ($self, $op, $args) = @_;
1175 my @parts = grep length($_->[0]), map [ $self->render_aqt($_) ], @$args;
1176 return '' unless @parts;
1177 return @{$parts[0]} if @parts == 1;
1178 my $join = ($op eq ','
1180 : ' '.$self->_sqlcase(join ' ', split '_', $op).' '
1182 return $self->_join_parts($join, @parts);
1186 my ($self, $join, @parts) = @_;
1188 join($join, map $_->[0], @parts),
1189 map @{$_}[1..$#$_], @parts
1193 sub _render_unop_paren {
1194 my ($self, $op, $v) = @_;
1195 my ($sql, @bind) = $self->_render_unop_prefix($op, $v);
1196 return "(${sql})", @bind;
1199 sub _render_unop_prefix {
1200 my ($self, $op, $v) = @_;
1201 my ($expr_sql, @bind) = $self->render_aqt($v->[0]);
1203 my $op_sql = $self->_sqlcase($op); # join ' ', split '_', $op);
1204 return ("${op_sql} ${expr_sql}", @bind);
1207 sub _render_unop_postfix {
1208 my ($self, $op, $v) = @_;
1209 my ($expr_sql, @bind) = $self->render_aqt($v->[0]);
1210 my $op_sql = $self->_sqlcase(join ' ', split '_', $op);
1211 return ($expr_sql.' '.$op_sql, @bind);
1214 # Some databases (SQLite) treat col IN (1, 2) different from
1215 # col IN ( (1, 2) ). Use this to strip all outer parens while
1216 # adding them back in the corresponding method
1217 sub _open_outer_paren {
1218 my ($self, $sql) = @_;
1220 while (my ($inner) = $sql =~ /^ \s* \( (.*) \) \s* $/xs) {
1222 # there are closing parens inside, need the heavy duty machinery
1223 # to reevaluate the extraction starting from $sql (full reevaluation)
1224 if ($inner =~ /\)/) {
1225 require Text::Balanced;
1227 my (undef, $remainder) = do {
1228 # idiotic design - writes to $@ but *DOES NOT* throw exceptions
1230 Text::Balanced::extract_bracketed($sql, '()', qr/\s*/);
1233 # the entire expression needs to be a balanced bracketed thing
1234 # (after an extract no remainder sans trailing space)
1235 last if defined $remainder and $remainder =~ /\S/;
1245 #======================================================================
1247 #======================================================================
1249 sub _expand_order_by {
1250 my ($self, $arg) = @_;
1252 return unless defined($arg) and not (ref($arg) eq 'ARRAY' and !@$arg);
1254 my $expander = sub {
1255 my ($self, $dir, $expr) = @_;
1256 my @to_expand = ref($expr) eq 'ARRAY' ? @$expr : $expr;
1257 foreach my $arg (@to_expand) {
1261 and grep /^-(asc|desc)$/, keys %$arg
1263 puke "ordering direction hash passed to order by must have exactly one key (-asc or -desc)";
1267 defined($dir) ? { -op => [ $dir =~ /^-?(.*)$/ ,=> $_ ] } : $_
1269 map $self->expand_expr($_, -ident),
1270 map ref($_) eq 'ARRAY' ? @$_ : $_, @to_expand;
1271 return undef unless @exp;
1272 return undef if @exp == 1 and not defined($exp[0]);
1273 return +{ -op => [ ',', @exp ] };
1276 local @{$self->{expand}}{qw(asc desc)} = (($expander) x 2);
1278 return $self->$expander(undef, $arg);
1282 my ($self, $arg) = @_;
1284 return '' unless defined(my $expanded = $self->_expand_order_by($arg));
1286 my ($sql, @bind) = $self->render_aqt($expanded);
1288 return '' unless length($sql);
1290 my $final_sql = $self->_sqlcase(' order by ').$sql;
1292 return wantarray ? ($final_sql, @bind) : $final_sql;
1295 # _order_by no longer needs to call this so doesn't but DBIC uses it.
1297 sub _order_by_chunks {
1298 my ($self, $arg) = @_;
1300 return () unless defined(my $expanded = $self->_expand_order_by($arg));
1302 return $self->_chunkify_order_by($expanded);
1305 sub _chunkify_order_by {
1306 my ($self, $expanded) = @_;
1308 return grep length, $self->render_aqt($expanded)
1309 if $expanded->{-ident} or @{$expanded->{-literal}||[]} == 1;
1312 if (ref() eq 'HASH' and $_->{-op} and $_->{-op}[0] eq ',') {
1313 my ($comma, @list) = @{$_->{-op}};
1314 return map $self->_chunkify_order_by($_), @list;
1316 return [ $self->render_aqt($_) ];
1320 #======================================================================
1321 # DATASOURCE (FOR NOW, JUST PLAIN TABLE OR LIST OF TABLES)
1322 #======================================================================
1328 $self->_expand_maybe_list_expr($from, -ident)
1333 #======================================================================
1335 #======================================================================
1337 sub _expand_maybe_list_expr {
1338 my ($self, $expr, $default) = @_;
1339 return +{ -op => [ ',',
1340 map $self->expand_expr($_, $default),
1341 ref($expr) eq 'ARRAY' ? @$expr : $expr
1345 # highly optimized, as it's called way too often
1347 # my ($self, $label) = @_;
1349 return '' unless defined $_[1];
1350 return ${$_[1]} if ref($_[1]) eq 'SCALAR';
1351 puke 'Identifier cannot be hashref' if ref($_[1]) eq 'HASH';
1353 unless ($_[0]->{quote_char}) {
1354 if (ref($_[1]) eq 'ARRAY') {
1355 return join($_[0]->{name_sep}||'.', @{$_[1]});
1357 $_[0]->_assert_pass_injection_guard($_[1]);
1362 my $qref = ref $_[0]->{quote_char};
1364 !$qref ? ($_[0]->{quote_char}, $_[0]->{quote_char})
1365 : ($qref eq 'ARRAY') ? @{$_[0]->{quote_char}}
1366 : puke "Unsupported quote_char format: $_[0]->{quote_char}";
1368 my $esc = $_[0]->{escape_char} || $r;
1370 # parts containing * are naturally unquoted
1372 $_[0]->{name_sep}||'',
1376 : do { (my $n = $_) =~ s/(\Q$esc\E|\Q$r\E)/$esc$1/g; $l . $n . $r }
1378 (ref($_[1]) eq 'ARRAY'
1382 ? split (/\Q$_[0]->{name_sep}\E/, $_[1] )
1390 # Conversion, if applicable
1392 #my ($self, $arg) = @_;
1393 if ($_[0]->{convert_where}) {
1394 return $_[0]->_sqlcase($_[0]->{convert_where}) .'(' . $_[1] . ')';
1401 #my ($self, $col, @vals) = @_;
1402 # called often - tighten code
1403 return $_[0]->{bindtype} eq 'columns'
1404 ? map {[$_[1], $_]} @_[2 .. $#_]
1409 # Dies if any element of @bind is not in [colname => value] format
1410 # if bindtype is 'columns'.
1411 sub _assert_bindval_matches_bindtype {
1412 # my ($self, @bind) = @_;
1414 if ($self->{bindtype} eq 'columns') {
1416 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
1417 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
1423 sub _join_sql_clauses {
1424 my ($self, $logic, $clauses_aref, $bind_aref) = @_;
1426 if (@$clauses_aref > 1) {
1427 my $join = " " . $self->_sqlcase($logic) . " ";
1428 my $sql = '( ' . join($join, @$clauses_aref) . ' )';
1429 return ($sql, @$bind_aref);
1431 elsif (@$clauses_aref) {
1432 return ($clauses_aref->[0], @$bind_aref); # no parentheses
1435 return (); # if no SQL, ignore @$bind_aref
1440 # Fix SQL case, if so requested
1442 # LDNOTE: if $self->{case} is true, then it contains 'lower', so we
1443 # don't touch the argument ... crooked logic, but let's not change it!
1444 return $_[0]->{case} ? $_[1] : uc($_[1]);
1448 #======================================================================
1449 # DISPATCHING FROM REFKIND
1450 #======================================================================
1453 my ($self, $data) = @_;
1455 return 'UNDEF' unless defined $data;
1457 # blessed objects are treated like scalars
1458 my $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1460 return 'SCALAR' unless $ref;
1463 while ($ref eq 'REF') {
1465 $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1469 return ($ref||'SCALAR') . ('REF' x $n_steps);
1473 my ($self, $data) = @_;
1474 my @try = ($self->_refkind($data));
1475 push @try, 'SCALAR_or_UNDEF' if $try[0] eq 'SCALAR' || $try[0] eq 'UNDEF';
1476 push @try, 'FALLBACK';
1480 sub _METHOD_FOR_refkind {
1481 my ($self, $meth_prefix, $data) = @_;
1484 for (@{$self->_try_refkind($data)}) {
1485 $method = $self->can($meth_prefix."_".$_)
1489 return $method || puke "cannot dispatch on '$meth_prefix' for ".$self->_refkind($data);
1493 sub _SWITCH_refkind {
1494 my ($self, $data, $dispatch_table) = @_;
1497 for (@{$self->_try_refkind($data)}) {
1498 $coderef = $dispatch_table->{$_}
1502 puke "no dispatch entry for ".$self->_refkind($data)
1511 #======================================================================
1512 # VALUES, GENERATE, AUTOLOAD
1513 #======================================================================
1515 # LDNOTE: original code from nwiger, didn't touch code in that section
1516 # I feel the AUTOLOAD stuff should not be the default, it should
1517 # only be activated on explicit demand by user.
1521 my $data = shift || return;
1522 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
1523 unless ref $data eq 'HASH';
1526 foreach my $k (sort keys %$data) {
1527 my $v = $data->{$k};
1528 $self->_SWITCH_refkind($v, {
1530 if ($self->{array_datatypes}) { # array datatype
1531 push @all_bind, $self->_bindtype($k, $v);
1533 else { # literal SQL with bind
1534 my ($sql, @bind) = @$v;
1535 $self->_assert_bindval_matches_bindtype(@bind);
1536 push @all_bind, @bind;
1539 ARRAYREFREF => sub { # literal SQL with bind
1540 my ($sql, @bind) = @${$v};
1541 $self->_assert_bindval_matches_bindtype(@bind);
1542 push @all_bind, @bind;
1544 SCALARREF => sub { # literal SQL without bind
1546 SCALAR_or_UNDEF => sub {
1547 push @all_bind, $self->_bindtype($k, $v);
1558 my(@sql, @sqlq, @sqlv);
1562 if ($ref eq 'HASH') {
1563 for my $k (sort keys %$_) {
1566 my $label = $self->_quote($k);
1567 if ($r eq 'ARRAY') {
1568 # literal SQL with bind
1569 my ($sql, @bind) = @$v;
1570 $self->_assert_bindval_matches_bindtype(@bind);
1571 push @sqlq, "$label = $sql";
1573 } elsif ($r eq 'SCALAR') {
1574 # literal SQL without bind
1575 push @sqlq, "$label = $$v";
1577 push @sqlq, "$label = ?";
1578 push @sqlv, $self->_bindtype($k, $v);
1581 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
1582 } elsif ($ref eq 'ARRAY') {
1583 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
1586 if ($r eq 'ARRAY') { # literal SQL with bind
1587 my ($sql, @bind) = @$v;
1588 $self->_assert_bindval_matches_bindtype(@bind);
1591 } elsif ($r eq 'SCALAR') { # literal SQL without bind
1592 # embedded literal SQL
1599 push @sql, '(' . join(', ', @sqlq) . ')';
1600 } elsif ($ref eq 'SCALAR') {
1604 # strings get case twiddled
1605 push @sql, $self->_sqlcase($_);
1609 my $sql = join ' ', @sql;
1611 # this is pretty tricky
1612 # if ask for an array, return ($stmt, @bind)
1613 # otherwise, s/?/shift @sqlv/ to put it inline
1615 return ($sql, @sqlv);
1617 1 while $sql =~ s/\?/my $d = shift(@sqlv);
1618 ref $d ? $d->[1] : $d/e;
1627 # This allows us to check for a local, then _form, attr
1629 my($name) = $AUTOLOAD =~ /.*::(.+)/;
1630 return $self->generate($name, @_);
1641 SQL::Abstract - Generate SQL from Perl data structures
1647 my $sql = SQL::Abstract->new;
1649 my($stmt, @bind) = $sql->select($source, \@fields, \%where, $order);
1651 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
1653 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
1655 my($stmt, @bind) = $sql->delete($table, \%where);
1657 # Then, use these in your DBI statements
1658 my $sth = $dbh->prepare($stmt);
1659 $sth->execute(@bind);
1661 # Just generate the WHERE clause
1662 my($stmt, @bind) = $sql->where(\%where, $order);
1664 # Return values in the same order, for hashed queries
1665 # See PERFORMANCE section for more details
1666 my @bind = $sql->values(\%fieldvals);
1670 This module was inspired by the excellent L<DBIx::Abstract>.
1671 However, in using that module I found that what I really wanted
1672 to do was generate SQL, but still retain complete control over my
1673 statement handles and use the DBI interface. So, I set out to
1674 create an abstract SQL generation module.
1676 While based on the concepts used by L<DBIx::Abstract>, there are
1677 several important differences, especially when it comes to WHERE
1678 clauses. I have modified the concepts used to make the SQL easier
1679 to generate from Perl data structures and, IMO, more intuitive.
1680 The underlying idea is for this module to do what you mean, based
1681 on the data structures you provide it. The big advantage is that
1682 you don't have to modify your code every time your data changes,
1683 as this module figures it out.
1685 To begin with, an SQL INSERT is as easy as just specifying a hash
1686 of C<key=value> pairs:
1689 name => 'Jimbo Bobson',
1690 phone => '123-456-7890',
1691 address => '42 Sister Lane',
1692 city => 'St. Louis',
1693 state => 'Louisiana',
1696 The SQL can then be generated with this:
1698 my($stmt, @bind) = $sql->insert('people', \%data);
1700 Which would give you something like this:
1702 $stmt = "INSERT INTO people
1703 (address, city, name, phone, state)
1704 VALUES (?, ?, ?, ?, ?)";
1705 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
1706 '123-456-7890', 'Louisiana');
1708 These are then used directly in your DBI code:
1710 my $sth = $dbh->prepare($stmt);
1711 $sth->execute(@bind);
1713 =head2 Inserting and Updating Arrays
1715 If your database has array types (like for example Postgres),
1716 activate the special option C<< array_datatypes => 1 >>
1717 when creating the C<SQL::Abstract> object.
1718 Then you may use an arrayref to insert and update database array types:
1720 my $sql = SQL::Abstract->new(array_datatypes => 1);
1722 planets => [qw/Mercury Venus Earth Mars/]
1725 my($stmt, @bind) = $sql->insert('solar_system', \%data);
1729 $stmt = "INSERT INTO solar_system (planets) VALUES (?)"
1731 @bind = (['Mercury', 'Venus', 'Earth', 'Mars']);
1734 =head2 Inserting and Updating SQL
1736 In order to apply SQL functions to elements of your C<%data> you may
1737 specify a reference to an arrayref for the given hash value. For example,
1738 if you need to execute the Oracle C<to_date> function on a value, you can
1739 say something like this:
1743 date_entered => \[ "to_date(?,'MM/DD/YYYY')", "03/02/2003" ],
1746 The first value in the array is the actual SQL. Any other values are
1747 optional and would be included in the bind values array. This gives
1750 my($stmt, @bind) = $sql->insert('people', \%data);
1752 $stmt = "INSERT INTO people (name, date_entered)
1753 VALUES (?, to_date(?,'MM/DD/YYYY'))";
1754 @bind = ('Bill', '03/02/2003');
1756 An UPDATE is just as easy, all you change is the name of the function:
1758 my($stmt, @bind) = $sql->update('people', \%data);
1760 Notice that your C<%data> isn't touched; the module will generate
1761 the appropriately quirky SQL for you automatically. Usually you'll
1762 want to specify a WHERE clause for your UPDATE, though, which is
1763 where handling C<%where> hashes comes in handy...
1765 =head2 Complex where statements
1767 This module can generate pretty complicated WHERE statements
1768 easily. For example, simple C<key=value> pairs are taken to mean
1769 equality, and if you want to see if a field is within a set
1770 of values, you can use an arrayref. Let's say we wanted to
1771 SELECT some data based on this criteria:
1774 requestor => 'inna',
1775 worker => ['nwiger', 'rcwe', 'sfz'],
1776 status => { '!=', 'completed' }
1779 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
1781 The above would give you something like this:
1783 $stmt = "SELECT * FROM tickets WHERE
1784 ( requestor = ? ) AND ( status != ? )
1785 AND ( worker = ? OR worker = ? OR worker = ? )";
1786 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
1788 Which you could then use in DBI code like so:
1790 my $sth = $dbh->prepare($stmt);
1791 $sth->execute(@bind);
1797 The methods are simple. There's one for every major SQL operation,
1798 and a constructor you use first. The arguments are specified in a
1799 similar order for each method (table, then fields, then a where
1800 clause) to try and simplify things.
1802 =head2 new(option => 'value')
1804 The C<new()> function takes a list of options and values, and returns
1805 a new B<SQL::Abstract> object which can then be used to generate SQL
1806 through the methods below. The options accepted are:
1812 If set to 'lower', then SQL will be generated in all lowercase. By
1813 default SQL is generated in "textbook" case meaning something like:
1815 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
1817 Any setting other than 'lower' is ignored.
1821 This determines what the default comparison operator is. By default
1822 it is C<=>, meaning that a hash like this:
1824 %where = (name => 'nwiger', email => 'nate@wiger.org');
1826 Will generate SQL like this:
1828 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
1830 However, you may want loose comparisons by default, so if you set
1831 C<cmp> to C<like> you would get SQL such as:
1833 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
1835 You can also override the comparison on an individual basis - see
1836 the huge section on L</"WHERE CLAUSES"> at the bottom.
1838 =item sqltrue, sqlfalse
1840 Expressions for inserting boolean values within SQL statements.
1841 By default these are C<1=1> and C<1=0>. They are used
1842 by the special operators C<-in> and C<-not_in> for generating
1843 correct SQL even when the argument is an empty array (see below).
1847 This determines the default logical operator for multiple WHERE
1848 statements in arrays or hashes. If absent, the default logic is "or"
1849 for arrays, and "and" for hashes. This means that a WHERE
1853 event_date => {'>=', '2/13/99'},
1854 event_date => {'<=', '4/24/03'},
1857 will generate SQL like this:
1859 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
1861 This is probably not what you want given this query, though (look
1862 at the dates). To change the "OR" to an "AND", simply specify:
1864 my $sql = SQL::Abstract->new(logic => 'and');
1866 Which will change the above C<WHERE> to:
1868 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
1870 The logic can also be changed locally by inserting
1871 a modifier in front of an arrayref:
1873 @where = (-and => [event_date => {'>=', '2/13/99'},
1874 event_date => {'<=', '4/24/03'} ]);
1876 See the L</"WHERE CLAUSES"> section for explanations.
1880 This will automatically convert comparisons using the specified SQL
1881 function for both column and value. This is mostly used with an argument
1882 of C<upper> or C<lower>, so that the SQL will have the effect of
1883 case-insensitive "searches". For example, this:
1885 $sql = SQL::Abstract->new(convert => 'upper');
1886 %where = (keywords => 'MaKe iT CAse inSeNSItive');
1888 Will turn out the following SQL:
1890 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
1892 The conversion can be C<upper()>, C<lower()>, or any other SQL function
1893 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
1894 not validate this option; it will just pass through what you specify verbatim).
1898 This is a kludge because many databases suck. For example, you can't
1899 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
1900 Instead, you have to use C<bind_param()>:
1902 $sth->bind_param(1, 'reg data');
1903 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
1905 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
1906 which loses track of which field each slot refers to. Fear not.
1908 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
1909 Currently, you can specify either C<normal> (default) or C<columns>. If you
1910 specify C<columns>, you will get an array that looks like this:
1912 my $sql = SQL::Abstract->new(bindtype => 'columns');
1913 my($stmt, @bind) = $sql->insert(...);
1916 [ 'column1', 'value1' ],
1917 [ 'column2', 'value2' ],
1918 [ 'column3', 'value3' ],
1921 You can then iterate through this manually, using DBI's C<bind_param()>.
1923 $sth->prepare($stmt);
1926 my($col, $data) = @$_;
1927 if ($col eq 'details' || $col eq 'comments') {
1928 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
1929 } elsif ($col eq 'image') {
1930 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
1932 $sth->bind_param($i, $data);
1936 $sth->execute; # execute without @bind now
1938 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
1939 Basically, the advantage is still that you don't have to care which fields
1940 are or are not included. You could wrap that above C<for> loop in a simple
1941 sub called C<bind_fields()> or something and reuse it repeatedly. You still
1942 get a layer of abstraction over manual SQL specification.
1944 Note that if you set L</bindtype> to C<columns>, the C<\[ $sql, @bind ]>
1945 construct (see L</Literal SQL with placeholders and bind values (subqueries)>)
1946 will expect the bind values in this format.
1950 This is the character that a table or column name will be quoted
1951 with. By default this is an empty string, but you could set it to
1952 the character C<`>, to generate SQL like this:
1954 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
1956 Alternatively, you can supply an array ref of two items, the first being the left
1957 hand quote character, and the second the right hand quote character. For
1958 example, you could supply C<['[',']']> for SQL Server 2000 compliant quotes
1959 that generates SQL like this:
1961 SELECT [a_field] FROM [a_table] WHERE [some_field] LIKE '%someval%'
1963 Quoting is useful if you have tables or columns names that are reserved
1964 words in your database's SQL dialect.
1968 This is the character that will be used to escape L</quote_char>s appearing
1969 in an identifier before it has been quoted.
1971 The parameter default in case of a single L</quote_char> character is the quote
1974 When opening-closing-style quoting is used (L</quote_char> is an arrayref)
1975 this parameter defaults to the B<closing (right)> L</quote_char>. Occurrences
1976 of the B<opening (left)> L</quote_char> within the identifier are currently left
1977 untouched. The default for opening-closing-style quotes may change in future
1978 versions, thus you are B<strongly encouraged> to specify the escape character
1983 This is the character that separates a table and column name. It is
1984 necessary to specify this when the C<quote_char> option is selected,
1985 so that tables and column names can be individually quoted like this:
1987 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
1989 =item injection_guard
1991 A regular expression C<qr/.../> that is applied to any C<-function> and unquoted
1992 column name specified in a query structure. This is a safety mechanism to avoid
1993 injection attacks when mishandling user input e.g.:
1995 my %condition_as_column_value_pairs = get_values_from_user();
1996 $sqla->select( ... , \%condition_as_column_value_pairs );
1998 If the expression matches an exception is thrown. Note that literal SQL
1999 supplied via C<\'...'> or C<\['...']> is B<not> checked in any way.
2001 Defaults to checking for C<;> and the C<GO> keyword (TransactSQL)
2003 =item array_datatypes
2005 When this option is true, arrayrefs in INSERT or UPDATE are
2006 interpreted as array datatypes and are passed directly
2008 When this option is false, arrayrefs are interpreted
2009 as literal SQL, just like refs to arrayrefs
2010 (but this behavior is for backwards compatibility; when writing
2011 new queries, use the "reference to arrayref" syntax
2017 Takes a reference to a list of "special operators"
2018 to extend the syntax understood by L<SQL::Abstract>.
2019 See section L</"SPECIAL OPERATORS"> for details.
2023 Takes a reference to a list of "unary operators"
2024 to extend the syntax understood by L<SQL::Abstract>.
2025 See section L</"UNARY OPERATORS"> for details.
2031 =head2 insert($table, \@values || \%fieldvals, \%options)
2033 This is the simplest function. You simply give it a table name
2034 and either an arrayref of values or hashref of field/value pairs.
2035 It returns an SQL INSERT statement and a list of bind values.
2036 See the sections on L</"Inserting and Updating Arrays"> and
2037 L</"Inserting and Updating SQL"> for information on how to insert
2038 with those data types.
2040 The optional C<\%options> hash reference may contain additional
2041 options to generate the insert SQL. Currently supported options
2048 Takes either a scalar of raw SQL fields, or an array reference of
2049 field names, and adds on an SQL C<RETURNING> statement at the end.
2050 This allows you to return data generated by the insert statement
2051 (such as row IDs) without performing another C<SELECT> statement.
2052 Note, however, this is not part of the SQL standard and may not
2053 be supported by all database engines.
2057 =head2 update($table, \%fieldvals, \%where, \%options)
2059 This takes a table, hashref of field/value pairs, and an optional
2060 hashref L<WHERE clause|/WHERE CLAUSES>. It returns an SQL UPDATE function and a list
2062 See the sections on L</"Inserting and Updating Arrays"> and
2063 L</"Inserting and Updating SQL"> for information on how to insert
2064 with those data types.
2066 The optional C<\%options> hash reference may contain additional
2067 options to generate the update SQL. Currently supported options
2074 See the C<returning> option to
2075 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2079 =head2 select($source, $fields, $where, $order)
2081 This returns a SQL SELECT statement and associated list of bind values, as
2082 specified by the arguments:
2088 Specification of the 'FROM' part of the statement.
2089 The argument can be either a plain scalar (interpreted as a table
2090 name, will be quoted), or an arrayref (interpreted as a list
2091 of table names, joined by commas, quoted), or a scalarref
2092 (literal SQL, not quoted).
2096 Specification of the list of fields to retrieve from
2098 The argument can be either an arrayref (interpreted as a list
2099 of field names, will be joined by commas and quoted), or a
2100 plain scalar (literal SQL, not quoted).
2101 Please observe that this API is not as flexible as that of
2102 the first argument C<$source>, for backwards compatibility reasons.
2106 Optional argument to specify the WHERE part of the query.
2107 The argument is most often a hashref, but can also be
2108 an arrayref or plain scalar --
2109 see section L<WHERE clause|/"WHERE CLAUSES"> for details.
2113 Optional argument to specify the ORDER BY part of the query.
2114 The argument can be a scalar, a hashref or an arrayref
2115 -- see section L<ORDER BY clause|/"ORDER BY CLAUSES">
2121 =head2 delete($table, \%where, \%options)
2123 This takes a table name and optional hashref L<WHERE clause|/WHERE CLAUSES>.
2124 It returns an SQL DELETE statement and list of bind values.
2126 The optional C<\%options> hash reference may contain additional
2127 options to generate the delete SQL. Currently supported options
2134 See the C<returning> option to
2135 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2139 =head2 where(\%where, $order)
2141 This is used to generate just the WHERE clause. For example,
2142 if you have an arbitrary data structure and know what the
2143 rest of your SQL is going to look like, but want an easy way
2144 to produce a WHERE clause, use this. It returns an SQL WHERE
2145 clause and list of bind values.
2148 =head2 values(\%data)
2150 This just returns the values from the hash C<%data>, in the same
2151 order that would be returned from any of the other above queries.
2152 Using this allows you to markedly speed up your queries if you
2153 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
2155 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
2157 Warning: This is an experimental method and subject to change.
2159 This returns arbitrarily generated SQL. It's a really basic shortcut.
2160 It will return two different things, depending on return context:
2162 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
2163 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
2165 These would return the following:
2167 # First calling form
2168 $stmt = "CREATE TABLE test (?, ?)";
2169 @bind = (field1, field2);
2171 # Second calling form
2172 $stmt_and_val = "CREATE TABLE test (field1, field2)";
2174 Depending on what you're trying to do, it's up to you to choose the correct
2175 format. In this example, the second form is what you would want.
2179 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
2183 ALTER SESSION SET nls_date_format = 'MM/YY'
2185 You get the idea. Strings get their case twiddled, but everything
2186 else remains verbatim.
2188 =head1 EXPORTABLE FUNCTIONS
2190 =head2 is_plain_value
2192 Determines if the supplied argument is a plain value as understood by this
2197 =item * The value is C<undef>
2199 =item * The value is a non-reference
2201 =item * The value is an object with stringification overloading
2203 =item * The value is of the form C<< { -value => $anything } >>
2207 On failure returns C<undef>, on success returns a B<scalar> reference
2208 to the original supplied argument.
2214 The stringification overloading detection is rather advanced: it takes
2215 into consideration not only the presence of a C<""> overload, but if that
2216 fails also checks for enabled
2217 L<autogenerated versions of C<"">|overload/Magic Autogeneration>, based
2218 on either C<0+> or C<bool>.
2220 Unfortunately testing in the field indicates that this
2221 detection B<< may tickle a latent bug in perl versions before 5.018 >>,
2222 but only when very large numbers of stringifying objects are involved.
2223 At the time of writing ( Sep 2014 ) there is no clear explanation of
2224 the direct cause, nor is there a manageably small test case that reliably
2225 reproduces the problem.
2227 If you encounter any of the following exceptions in B<random places within
2228 your application stack> - this module may be to blame:
2230 Operation "ne": no method found,
2231 left argument in overloaded package <something>,
2232 right argument in overloaded package <something>
2236 Stub found while resolving method "???" overloading """" in package <something>
2238 If you fall victim to the above - please attempt to reduce the problem
2239 to something that could be sent to the L<SQL::Abstract developers
2240 |DBIx::Class/GETTING HELP/SUPPORT>
2241 (either publicly or privately). As a workaround in the meantime you can
2242 set C<$ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}> to a true
2243 value, which will most likely eliminate your problem (at the expense of
2244 not being able to properly detect exotic forms of stringification).
2246 This notice and environment variable will be removed in a future version,
2247 as soon as the underlying problem is found and a reliable workaround is
2252 =head2 is_literal_value
2254 Determines if the supplied argument is a literal value as understood by this
2259 =item * C<\$sql_string>
2261 =item * C<\[ $sql_string, @bind_values ]>
2265 On failure returns C<undef>, on success returns an B<array> reference
2266 containing the unpacked version of the supplied literal SQL and bind values.
2268 =head1 WHERE CLAUSES
2272 This module uses a variation on the idea from L<DBIx::Abstract>. It
2273 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
2274 module is that things in arrays are OR'ed, and things in hashes
2277 The easiest way to explain is to show lots of examples. After
2278 each C<%where> hash shown, it is assumed you used:
2280 my($stmt, @bind) = $sql->where(\%where);
2282 However, note that the C<%where> hash can be used directly in any
2283 of the other functions as well, as described above.
2285 =head2 Key-value pairs
2287 So, let's get started. To begin, a simple hash:
2291 status => 'completed'
2294 Is converted to SQL C<key = val> statements:
2296 $stmt = "WHERE user = ? AND status = ?";
2297 @bind = ('nwiger', 'completed');
2299 One common thing I end up doing is having a list of values that
2300 a field can be in. To do this, simply specify a list inside of
2305 status => ['assigned', 'in-progress', 'pending'];
2308 This simple code will create the following:
2310 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
2311 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
2313 A field associated to an empty arrayref will be considered a
2314 logical false and will generate 0=1.
2316 =head2 Tests for NULL values
2318 If the value part is C<undef> then this is converted to SQL <IS NULL>
2327 $stmt = "WHERE user = ? AND status IS NULL";
2330 To test if a column IS NOT NULL:
2334 status => { '!=', undef },
2337 =head2 Specific comparison operators
2339 If you want to specify a different type of operator for your comparison,
2340 you can use a hashref for a given column:
2344 status => { '!=', 'completed' }
2347 Which would generate:
2349 $stmt = "WHERE user = ? AND status != ?";
2350 @bind = ('nwiger', 'completed');
2352 To test against multiple values, just enclose the values in an arrayref:
2354 status => { '=', ['assigned', 'in-progress', 'pending'] };
2356 Which would give you:
2358 "WHERE status = ? OR status = ? OR status = ?"
2361 The hashref can also contain multiple pairs, in which case it is expanded
2362 into an C<AND> of its elements:
2366 status => { '!=', 'completed', -not_like => 'pending%' }
2369 # Or more dynamically, like from a form
2370 $where{user} = 'nwiger';
2371 $where{status}{'!='} = 'completed';
2372 $where{status}{'-not_like'} = 'pending%';
2374 # Both generate this
2375 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
2376 @bind = ('nwiger', 'completed', 'pending%');
2379 To get an OR instead, you can combine it with the arrayref idea:
2383 priority => [ { '=', 2 }, { '>', 5 } ]
2386 Which would generate:
2388 $stmt = "WHERE ( priority = ? OR priority > ? ) AND user = ?";
2389 @bind = ('2', '5', 'nwiger');
2391 If you want to include literal SQL (with or without bind values), just use a
2392 scalar reference or reference to an arrayref as the value:
2395 date_entered => { '>' => \["to_date(?, 'MM/DD/YYYY')", "11/26/2008"] },
2396 date_expires => { '<' => \"now()" }
2399 Which would generate:
2401 $stmt = "WHERE date_entered > to_date(?, 'MM/DD/YYYY') AND date_expires < now()";
2402 @bind = ('11/26/2008');
2405 =head2 Logic and nesting operators
2407 In the example above,
2408 there is a subtle trap if you want to say something like
2409 this (notice the C<AND>):
2411 WHERE priority != ? AND priority != ?
2413 Because, in Perl you I<can't> do this:
2415 priority => { '!=' => 2, '!=' => 1 }
2417 As the second C<!=> key will obliterate the first. The solution
2418 is to use the special C<-modifier> form inside an arrayref:
2420 priority => [ -and => {'!=', 2},
2424 Normally, these would be joined by C<OR>, but the modifier tells it
2425 to use C<AND> instead. (Hint: You can use this in conjunction with the
2426 C<logic> option to C<new()> in order to change the way your queries
2427 work by default.) B<Important:> Note that the C<-modifier> goes
2428 B<INSIDE> the arrayref, as an extra first element. This will
2429 B<NOT> do what you think it might:
2431 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
2433 Here is a quick list of equivalencies, since there is some overlap:
2436 status => {'!=', 'completed', 'not like', 'pending%' }
2437 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
2440 status => {'=', ['assigned', 'in-progress']}
2441 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
2442 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
2446 =head2 Special operators: IN, BETWEEN, etc.
2448 You can also use the hashref format to compare a list of fields using the
2449 C<IN> comparison operator, by specifying the list as an arrayref:
2452 status => 'completed',
2453 reportid => { -in => [567, 2335, 2] }
2456 Which would generate:
2458 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
2459 @bind = ('completed', '567', '2335', '2');
2461 The reverse operator C<-not_in> generates SQL C<NOT IN> and is used in
2464 If the argument to C<-in> is an empty array, 'sqlfalse' is generated
2465 (by default: C<1=0>). Similarly, C<< -not_in => [] >> generates
2466 'sqltrue' (by default: C<1=1>).
2468 In addition to the array you can supply a chunk of literal sql or
2469 literal sql with bind:
2472 customer => { -in => \[
2473 'SELECT cust_id FROM cust WHERE balance > ?',
2476 status => { -in => \'SELECT status_codes FROM states' },
2482 customer IN ( SELECT cust_id FROM cust WHERE balance > ? )
2483 AND status IN ( SELECT status_codes FROM states )
2487 Finally, if the argument to C<-in> is not a reference, it will be
2488 treated as a single-element array.
2490 Another pair of operators is C<-between> and C<-not_between>,
2491 used with an arrayref of two values:
2495 completion_date => {
2496 -not_between => ['2002-10-01', '2003-02-06']
2502 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
2504 Just like with C<-in> all plausible combinations of literal SQL
2508 start0 => { -between => [ 1, 2 ] },
2509 start1 => { -between => \["? AND ?", 1, 2] },
2510 start2 => { -between => \"lower(x) AND upper(y)" },
2511 start3 => { -between => [
2513 \["upper(?)", 'stuff' ],
2520 ( start0 BETWEEN ? AND ? )
2521 AND ( start1 BETWEEN ? AND ? )
2522 AND ( start2 BETWEEN lower(x) AND upper(y) )
2523 AND ( start3 BETWEEN lower(x) AND upper(?) )
2525 @bind = (1, 2, 1, 2, 'stuff');
2528 These are the two builtin "special operators"; but the
2529 list can be expanded: see section L</"SPECIAL OPERATORS"> below.
2531 =head2 Unary operators: bool
2533 If you wish to test against boolean columns or functions within your
2534 database you can use the C<-bool> and C<-not_bool> operators. For
2535 example to test the column C<is_user> being true and the column
2536 C<is_enabled> being false you would use:-
2540 -not_bool => 'is_enabled',
2545 WHERE is_user AND NOT is_enabled
2547 If a more complex combination is required, testing more conditions,
2548 then you should use the and/or operators:-
2553 -not_bool => { two=> { -rlike => 'bar' } },
2554 -not_bool => { three => [ { '=', 2 }, { '>', 5 } ] },
2565 (NOT ( three = ? OR three > ? ))
2568 =head2 Nested conditions, -and/-or prefixes
2570 So far, we've seen how multiple conditions are joined with a top-level
2571 C<AND>. We can change this by putting the different conditions we want in
2572 hashes and then putting those hashes in an array. For example:
2577 status => { -like => ['pending%', 'dispatched'] },
2581 status => 'unassigned',
2585 This data structure would create the following:
2587 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
2588 OR ( user = ? AND status = ? ) )";
2589 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
2592 Clauses in hashrefs or arrayrefs can be prefixed with an C<-and> or C<-or>
2593 to change the logic inside:
2599 -and => [ workhrs => {'>', 20}, geo => 'ASIA' ],
2600 -or => { workhrs => {'<', 50}, geo => 'EURO' },
2607 $stmt = "WHERE ( user = ?
2608 AND ( ( workhrs > ? AND geo = ? )
2609 OR ( workhrs < ? OR geo = ? ) ) )";
2610 @bind = ('nwiger', '20', 'ASIA', '50', 'EURO');
2612 =head3 Algebraic inconsistency, for historical reasons
2614 C<Important note>: when connecting several conditions, the C<-and->|C<-or>
2615 operator goes C<outside> of the nested structure; whereas when connecting
2616 several constraints on one column, the C<-and> operator goes
2617 C<inside> the arrayref. Here is an example combining both features:
2620 -and => [a => 1, b => 2],
2621 -or => [c => 3, d => 4],
2622 e => [-and => {-like => 'foo%'}, {-like => '%bar'} ]
2627 WHERE ( ( ( a = ? AND b = ? )
2628 OR ( c = ? OR d = ? )
2629 OR ( e LIKE ? AND e LIKE ? ) ) )
2631 This difference in syntax is unfortunate but must be preserved for
2632 historical reasons. So be careful: the two examples below would
2633 seem algebraically equivalent, but they are not
2636 { -like => 'foo%' },
2637 { -like => '%bar' },
2639 # yields: WHERE ( ( col LIKE ? AND col LIKE ? ) )
2642 { col => { -like => 'foo%' } },
2643 { col => { -like => '%bar' } },
2645 # yields: WHERE ( ( col LIKE ? OR col LIKE ? ) )
2648 =head2 Literal SQL and value type operators
2650 The basic premise of SQL::Abstract is that in WHERE specifications the "left
2651 side" is a column name and the "right side" is a value (normally rendered as
2652 a placeholder). This holds true for both hashrefs and arrayref pairs as you
2653 see in the L</WHERE CLAUSES> examples above. Sometimes it is necessary to
2654 alter this behavior. There are several ways of doing so.
2658 This is a virtual operator that signals the string to its right side is an
2659 identifier (a column name) and not a value. For example to compare two
2660 columns you would write:
2663 priority => { '<', 2 },
2664 requestor => { -ident => 'submitter' },
2669 $stmt = "WHERE priority < ? AND requestor = submitter";
2672 If you are maintaining legacy code you may see a different construct as
2673 described in L</Deprecated usage of Literal SQL>, please use C<-ident> in new
2678 This is a virtual operator that signals that the construct to its right side
2679 is a value to be passed to DBI. This is for example necessary when you want
2680 to write a where clause against an array (for RDBMS that support such
2681 datatypes). For example:
2684 array => { -value => [1, 2, 3] }
2689 $stmt = 'WHERE array = ?';
2690 @bind = ([1, 2, 3]);
2692 Note that if you were to simply say:
2698 the result would probably not be what you wanted:
2700 $stmt = 'WHERE array = ? OR array = ? OR array = ?';
2705 Finally, sometimes only literal SQL will do. To include a random snippet
2706 of SQL verbatim, you specify it as a scalar reference. Consider this only
2707 as a last resort. Usually there is a better way. For example:
2710 priority => { '<', 2 },
2711 requestor => { -in => \'(SELECT name FROM hitmen)' },
2716 $stmt = "WHERE priority < ? AND requestor IN (SELECT name FROM hitmen)"
2719 Note that in this example, you only get one bind parameter back, since
2720 the verbatim SQL is passed as part of the statement.
2724 Never use untrusted input as a literal SQL argument - this is a massive
2725 security risk (there is no way to check literal snippets for SQL
2726 injections and other nastyness). If you need to deal with untrusted input
2727 use literal SQL with placeholders as described next.
2729 =head3 Literal SQL with placeholders and bind values (subqueries)
2731 If the literal SQL to be inserted has placeholders and bind values,
2732 use a reference to an arrayref (yes this is a double reference --
2733 not so common, but perfectly legal Perl). For example, to find a date
2734 in Postgres you can use something like this:
2737 date_column => \[ "= date '2008-09-30' - ?::integer", 10 ]
2742 $stmt = "WHERE ( date_column = date '2008-09-30' - ?::integer )"
2745 Note that you must pass the bind values in the same format as they are returned
2746 by L<where|/where(\%where, $order)>. This means that if you set L</bindtype>
2747 to C<columns>, you must provide the bind values in the
2748 C<< [ column_meta => value ] >> format, where C<column_meta> is an opaque
2749 scalar value; most commonly the column name, but you can use any scalar value
2750 (including references and blessed references), L<SQL::Abstract> will simply
2751 pass it through intact. So if C<bindtype> is set to C<columns> the above
2752 example will look like:
2755 date_column => \[ "= date '2008-09-30' - ?::integer", [ {} => 10 ] ]
2758 Literal SQL is especially useful for nesting parenthesized clauses in the
2759 main SQL query. Here is a first example:
2761 my ($sub_stmt, @sub_bind) = ("SELECT c1 FROM t1 WHERE c2 < ? AND c3 LIKE ?",
2765 bar => \["IN ($sub_stmt)" => @sub_bind],
2770 $stmt = "WHERE (foo = ? AND bar IN (SELECT c1 FROM t1
2771 WHERE c2 < ? AND c3 LIKE ?))";
2772 @bind = (1234, 100, "foo%");
2774 Other subquery operators, like for example C<"E<gt> ALL"> or C<"NOT IN">,
2775 are expressed in the same way. Of course the C<$sub_stmt> and
2776 its associated bind values can be generated through a former call
2779 my ($sub_stmt, @sub_bind)
2780 = $sql->select("t1", "c1", {c2 => {"<" => 100},
2781 c3 => {-like => "foo%"}});
2784 bar => \["> ALL ($sub_stmt)" => @sub_bind],
2787 In the examples above, the subquery was used as an operator on a column;
2788 but the same principle also applies for a clause within the main C<%where>
2789 hash, like an EXISTS subquery:
2791 my ($sub_stmt, @sub_bind)
2792 = $sql->select("t1", "*", {c1 => 1, c2 => \"> t0.c0"});
2793 my %where = ( -and => [
2795 \["EXISTS ($sub_stmt)" => @sub_bind],
2800 $stmt = "WHERE (foo = ? AND EXISTS (SELECT * FROM t1
2801 WHERE c1 = ? AND c2 > t0.c0))";
2805 Observe that the condition on C<c2> in the subquery refers to
2806 column C<t0.c0> of the main query: this is I<not> a bind
2807 value, so we have to express it through a scalar ref.
2808 Writing C<< c2 => {">" => "t0.c0"} >> would have generated
2809 C<< c2 > ? >> with bind value C<"t0.c0"> ... not exactly
2810 what we wanted here.
2812 Finally, here is an example where a subquery is used
2813 for expressing unary negation:
2815 my ($sub_stmt, @sub_bind)
2816 = $sql->where({age => [{"<" => 10}, {">" => 20}]});
2817 $sub_stmt =~ s/^ where //i; # don't want "WHERE" in the subclause
2819 lname => {like => '%son%'},
2820 \["NOT ($sub_stmt)" => @sub_bind],
2825 $stmt = "lname LIKE ? AND NOT ( age < ? OR age > ? )"
2826 @bind = ('%son%', 10, 20)
2828 =head3 Deprecated usage of Literal SQL
2830 Below are some examples of archaic use of literal SQL. It is shown only as
2831 reference for those who deal with legacy code. Each example has a much
2832 better, cleaner and safer alternative that users should opt for in new code.
2838 my %where = ( requestor => \'IS NOT NULL' )
2840 $stmt = "WHERE requestor IS NOT NULL"
2842 This used to be the way of generating NULL comparisons, before the handling
2843 of C<undef> got formalized. For new code please use the superior syntax as
2844 described in L</Tests for NULL values>.
2848 my %where = ( requestor => \'= submitter' )
2850 $stmt = "WHERE requestor = submitter"
2852 This used to be the only way to compare columns. Use the superior L</-ident>
2853 method for all new code. For example an identifier declared in such a way
2854 will be properly quoted if L</quote_char> is properly set, while the legacy
2855 form will remain as supplied.
2859 my %where = ( is_ready => \"", completed => { '>', '2012-12-21' } )
2861 $stmt = "WHERE completed > ? AND is_ready"
2862 @bind = ('2012-12-21')
2864 Using an empty string literal used to be the only way to express a boolean.
2865 For all new code please use the much more readable
2866 L<-bool|/Unary operators: bool> operator.
2872 These pages could go on for a while, since the nesting of the data
2873 structures this module can handle are pretty much unlimited (the
2874 module implements the C<WHERE> expansion as a recursive function
2875 internally). Your best bet is to "play around" with the module a
2876 little to see how the data structures behave, and choose the best
2877 format for your data based on that.
2879 And of course, all the values above will probably be replaced with
2880 variables gotten from forms or the command line. After all, if you
2881 knew everything ahead of time, you wouldn't have to worry about
2882 dynamically-generating SQL and could just hardwire it into your
2885 =head1 ORDER BY CLAUSES
2887 Some functions take an order by clause. This can either be a scalar (just a
2888 column name), a hashref of C<< { -desc => 'col' } >> or C<< { -asc => 'col' }
2889 >>, a scalarref, an arrayref-ref, or an arrayref of any of the previous
2892 Given | Will Generate
2893 ---------------------------------------------------------------
2895 'colA' | ORDER BY colA
2897 [qw/colA colB/] | ORDER BY colA, colB
2899 {-asc => 'colA'} | ORDER BY colA ASC
2901 {-desc => 'colB'} | ORDER BY colB DESC
2903 ['colA', {-asc => 'colB'}] | ORDER BY colA, colB ASC
2905 { -asc => [qw/colA colB/] } | ORDER BY colA ASC, colB ASC
2907 \'colA DESC' | ORDER BY colA DESC
2909 \[ 'FUNC(colA, ?)', $x ] | ORDER BY FUNC(colA, ?)
2910 | /* ...with $x bound to ? */
2913 { -asc => 'colA' }, | colA ASC,
2914 { -desc => [qw/colB/] }, | colB DESC,
2915 { -asc => [qw/colC colD/] },| colC ASC, colD ASC,
2916 \'colE DESC', | colE DESC,
2917 \[ 'FUNC(colF, ?)', $x ], | FUNC(colF, ?)
2918 ] | /* ...with $x bound to ? */
2919 ===============================================================
2923 =head1 SPECIAL OPERATORS
2925 my $sqlmaker = SQL::Abstract->new(special_ops => [
2929 my ($self, $field, $op, $arg) = @_;
2935 handler => 'method_name',
2939 A "special operator" is a SQL syntactic clause that can be
2940 applied to a field, instead of a usual binary operator.
2943 WHERE field IN (?, ?, ?)
2944 WHERE field BETWEEN ? AND ?
2945 WHERE MATCH(field) AGAINST (?, ?)
2947 Special operators IN and BETWEEN are fairly standard and therefore
2948 are builtin within C<SQL::Abstract> (as the overridable methods
2949 C<_where_field_IN> and C<_where_field_BETWEEN>). For other operators,
2950 like the MATCH .. AGAINST example above which is specific to MySQL,
2951 you can write your own operator handlers - supply a C<special_ops>
2952 argument to the C<new> method. That argument takes an arrayref of
2953 operator definitions; each operator definition is a hashref with two
2960 the regular expression to match the operator
2964 Either a coderef or a plain scalar method name. In both cases
2965 the expected return is C<< ($sql, @bind) >>.
2967 When supplied with a method name, it is simply called on the
2968 L<SQL::Abstract> object as:
2970 $self->$method_name($field, $op, $arg)
2974 $field is the LHS of the operator
2975 $op is the part that matched the handler regex
2978 When supplied with a coderef, it is called as:
2980 $coderef->($self, $field, $op, $arg)
2985 For example, here is an implementation
2986 of the MATCH .. AGAINST syntax for MySQL
2988 my $sqlmaker = SQL::Abstract->new(special_ops => [
2990 # special op for MySql MATCH (field) AGAINST(word1, word2, ...)
2991 {regex => qr/^match$/i,
2993 my ($self, $field, $op, $arg) = @_;
2994 $arg = [$arg] if not ref $arg;
2995 my $label = $self->_quote($field);
2996 my ($placeholder) = $self->_convert('?');
2997 my $placeholders = join ", ", (($placeholder) x @$arg);
2998 my $sql = $self->_sqlcase('match') . " ($label) "
2999 . $self->_sqlcase('against') . " ($placeholders) ";
3000 my @bind = $self->_bindtype($field, @$arg);
3001 return ($sql, @bind);
3008 =head1 UNARY OPERATORS
3010 my $sqlmaker = SQL::Abstract->new(unary_ops => [
3014 my ($self, $op, $arg) = @_;
3020 handler => 'method_name',
3024 A "unary operator" is a SQL syntactic clause that can be
3025 applied to a field - the operator goes before the field
3027 You can write your own operator handlers - supply a C<unary_ops>
3028 argument to the C<new> method. That argument takes an arrayref of
3029 operator definitions; each operator definition is a hashref with two
3036 the regular expression to match the operator
3040 Either a coderef or a plain scalar method name. In both cases
3041 the expected return is C<< $sql >>.
3043 When supplied with a method name, it is simply called on the
3044 L<SQL::Abstract> object as:
3046 $self->$method_name($op, $arg)
3050 $op is the part that matched the handler regex
3051 $arg is the RHS or argument of the operator
3053 When supplied with a coderef, it is called as:
3055 $coderef->($self, $op, $arg)
3063 Thanks to some benchmarking by Mark Stosberg, it turns out that
3064 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
3065 I must admit this wasn't an intentional design issue, but it's a
3066 byproduct of the fact that you get to control your C<DBI> handles
3069 To maximize performance, use a code snippet like the following:
3071 # prepare a statement handle using the first row
3072 # and then reuse it for the rest of the rows
3074 for my $href (@array_of_hashrefs) {
3075 $stmt ||= $sql->insert('table', $href);
3076 $sth ||= $dbh->prepare($stmt);
3077 $sth->execute($sql->values($href));
3080 The reason this works is because the keys in your C<$href> are sorted
3081 internally by B<SQL::Abstract>. Thus, as long as your data retains
3082 the same structure, you only have to generate the SQL the first time
3083 around. On subsequent queries, simply use the C<values> function provided
3084 by this module to return your values in the correct order.
3086 However this depends on the values having the same type - if, for
3087 example, the values of a where clause may either have values
3088 (resulting in sql of the form C<column = ?> with a single bind
3089 value), or alternatively the values might be C<undef> (resulting in
3090 sql of the form C<column IS NULL> with no bind value) then the
3091 caching technique suggested will not work.
3095 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
3096 really like this part (I do, at least). Building up a complex query
3097 can be as simple as the following:
3104 use CGI::FormBuilder;
3107 my $form = CGI::FormBuilder->new(...);
3108 my $sql = SQL::Abstract->new;
3110 if ($form->submitted) {
3111 my $field = $form->field;
3112 my $id = delete $field->{id};
3113 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
3116 Of course, you would still have to connect using C<DBI> to run the
3117 query, but the point is that if you make your form look like your
3118 table, the actual query script can be extremely simplistic.
3120 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
3121 a fast interface to returning and formatting data. I frequently
3122 use these three modules together to write complex database query
3123 apps in under 50 lines.
3125 =head1 HOW TO CONTRIBUTE
3127 Contributions are always welcome, in all usable forms (we especially
3128 welcome documentation improvements). The delivery methods include git-
3129 or unified-diff formatted patches, GitHub pull requests, or plain bug
3130 reports either via RT or the Mailing list. Contributors are generally
3131 granted full access to the official repository after their first several
3132 patches pass successful review.
3134 This project is maintained in a git repository. The code and related tools are
3135 accessible at the following locations:
3139 =item * Official repo: L<git://git.shadowcat.co.uk/dbsrgits/SQL-Abstract.git>
3141 =item * Official gitweb: L<http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=dbsrgits/SQL-Abstract.git>
3143 =item * GitHub mirror: L<https://github.com/dbsrgits/sql-abstract>
3145 =item * Authorized committers: L<ssh://dbsrgits@git.shadowcat.co.uk/SQL-Abstract.git>
3151 Version 1.50 was a major internal refactoring of C<SQL::Abstract>.
3152 Great care has been taken to preserve the I<published> behavior
3153 documented in previous versions in the 1.* family; however,
3154 some features that were previously undocumented, or behaved
3155 differently from the documentation, had to be changed in order
3156 to clarify the semantics. Hence, client code that was relying
3157 on some dark areas of C<SQL::Abstract> v1.*
3158 B<might behave differently> in v1.50.
3160 The main changes are:
3166 support for literal SQL through the C<< \ [ $sql, @bind ] >> syntax.
3170 support for the { operator => \"..." } construct (to embed literal SQL)
3174 support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values)
3178 optional support for L<array datatypes|/"Inserting and Updating Arrays">
3182 defensive programming: check arguments
3186 fixed bug with global logic, which was previously implemented
3187 through global variables yielding side-effects. Prior versions would
3188 interpret C<< [ {cond1, cond2}, [cond3, cond4] ] >>
3189 as C<< "(cond1 AND cond2) OR (cond3 AND cond4)" >>.
3190 Now this is interpreted
3191 as C<< "(cond1 AND cond2) OR (cond3 OR cond4)" >>.
3196 fixed semantics of _bindtype on array args
3200 dropped the C<_anoncopy> of the %where tree. No longer necessary,
3201 we just avoid shifting arrays within that tree.
3205 dropped the C<_modlogic> function
3209 =head1 ACKNOWLEDGEMENTS
3211 There are a number of individuals that have really helped out with
3212 this module. Unfortunately, most of them submitted bugs via CPAN
3213 so I have no idea who they are! But the people I do know are:
3215 Ash Berlin (order_by hash term support)
3216 Matt Trout (DBIx::Class support)
3217 Mark Stosberg (benchmarking)
3218 Chas Owens (initial "IN" operator support)
3219 Philip Collins (per-field SQL functions)
3220 Eric Kolve (hashref "AND" support)
3221 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
3222 Dan Kubb (support for "quote_char" and "name_sep")
3223 Guillermo Roditi (patch to cleanup "IN" and "BETWEEN", fix and tests for _order_by)
3224 Laurent Dami (internal refactoring, extensible list of special operators, literal SQL)
3225 Norbert Buchmuller (support for literal SQL in hashpair, misc. fixes & tests)
3226 Peter Rabbitson (rewrite of SQLA::Test, misc. fixes & tests)
3227 Oliver Charles (support for "RETURNING" after "INSERT")
3233 L<DBIx::Class>, L<DBIx::Abstract>, L<CGI::FormBuilder>, L<HTML::QuickTable>.
3237 Copyright (c) 2001-2007 Nathan Wiger <nwiger@cpan.org>. All Rights Reserved.
3239 This module is actively maintained by Matt Trout <mst@shadowcatsystems.co.uk>
3241 For support, your best bet is to try the C<DBIx::Class> users mailing list.
3242 While not an official support venue, C<DBIx::Class> makes heavy use of
3243 C<SQL::Abstract>, and as such list members there are very familiar with
3244 how to create queries.
3248 This module is free software; you may copy this under the same
3249 terms as perl itself (either the GNU General Public License or
3250 the Artistic License)