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/^ (?: not \s )? in $/ix, handler => sub { die "NOPE" }},
42 {regex => qr/^ ident $/ix, handler => sub { die "NOPE" }},
43 {regex => qr/^ value $/ix, handler => sub { die "NOPE" }},
44 {regex => qr/^ is (?: \s+ not )? $/ix, handler => sub { die "NOPE" }},
47 # unaryish operators - key maps to handler
48 my @BUILTIN_UNARY_OPS = (
49 # the digits are backcompat stuff
50 { regex => qr/^ and (?: [_\s]? \d+ )? $/xi, handler => '_where_op_ANDOR' },
51 { regex => qr/^ or (?: [_\s]? \d+ )? $/xi, handler => '_where_op_ANDOR' },
52 { regex => qr/^ nest (?: [_\s]? \d+ )? $/xi, handler => '_where_op_NEST' },
53 { regex => qr/^ (?: not \s )? bool $/xi, handler => '_where_op_BOOL' },
54 { regex => qr/^ ident $/xi, handler => '_where_op_IDENT' },
55 { regex => qr/^ value $/xi, handler => '_where_op_VALUE' },
56 { regex => qr/^ op $/xi, handler => '_where_op_OP' },
57 { regex => qr/^ bind $/xi, handler => '_where_op_BIND' },
58 { regex => qr/^ literal $/xi, handler => '_where_op_LITERAL' },
59 { regex => qr/^ func $/xi, handler => '_where_op_FUNC' },
62 #======================================================================
63 # DEBUGGING AND ERROR REPORTING
64 #======================================================================
67 return unless $_[0]->{debug}; shift; # a little faster
68 my $func = (caller(1))[3];
69 warn "[$func] ", @_, "\n";
73 my($func) = (caller(1))[3];
74 Carp::carp "[$func] Warning: ", @_;
78 my($func) = (caller(1))[3];
79 Carp::croak "[$func] Fatal: ", @_;
82 sub is_literal_value ($) {
83 ref $_[0] eq 'SCALAR' ? [ ${$_[0]} ]
84 : ( ref $_[0] eq 'REF' and ref ${$_[0]} eq 'ARRAY' ) ? [ @${ $_[0] } ]
88 # FIXME XSify - this can be done so much more efficiently
89 sub is_plain_value ($) {
91 ! length ref $_[0] ? \($_[0])
93 ref $_[0] eq 'HASH' and keys %{$_[0]} == 1
95 exists $_[0]->{-value}
96 ) ? \($_[0]->{-value})
98 # reuse @_ for even moar speedz
99 defined ( $_[1] = Scalar::Util::blessed $_[0] )
101 # deliberately not using Devel::OverloadInfo - the checks we are
102 # intersted in are much more limited than the fullblown thing, and
103 # this is a very hot piece of code
105 # simply using ->can('(""') can leave behind stub methods that
106 # break actually using the overload later (see L<perldiag/Stub
107 # found while resolving method "%s" overloading "%s" in package
108 # "%s"> and the source of overload::mycan())
110 # either has stringification which DBI SHOULD prefer out of the box
111 grep { *{ (qq[${_}::(""]) }{CODE} } @{ $_[2] = mro::get_linear_isa( $_[1] ) }
113 # has nummification or boolification, AND fallback is *not* disabled
115 SQL::Abstract::_ENV_::DETECT_AUTOGENERATED_STRINGIFICATION
118 grep { *{"${_}::(0+"}{CODE} } @{$_[2]}
120 grep { *{"${_}::(bool"}{CODE} } @{$_[2]}
124 # no fallback specified at all
125 ! ( ($_[3]) = grep { *{"${_}::()"}{CODE} } @{$_[2]} )
127 # fallback explicitly undef
128 ! defined ${"$_[3]::()"}
141 #======================================================================
143 #======================================================================
147 my $class = ref($self) || $self;
148 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
150 # choose our case by keeping an option around
151 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
153 # default logic for interpreting arrayrefs
154 $opt{logic} = $opt{logic} ? uc $opt{logic} : 'OR';
156 # how to return bind vars
157 $opt{bindtype} ||= 'normal';
159 # default comparison is "=", but can be overridden
162 # try to recognize which are the 'equality' and 'inequality' ops
163 # (temporary quickfix (in 2007), should go through a more seasoned API)
164 $opt{equality_op} = qr/^( \Q$opt{cmp}\E | \= )$/ix;
165 $opt{inequality_op} = qr/^( != | <> )$/ix;
167 $opt{like_op} = qr/^ (is\s+)? r?like $/xi;
168 $opt{not_like_op} = qr/^ (is\s+)? not \s+ r?like $/xi;
171 $opt{sqltrue} ||= '1=1';
172 $opt{sqlfalse} ||= '0=1';
175 $opt{user_special_ops} = [ @{$opt{special_ops} ||= []} ];
176 # regexes are applied in order, thus push after user-defines
177 push @{$opt{special_ops}}, @BUILTIN_SPECIAL_OPS;
180 $opt{unary_ops} ||= [];
181 push @{$opt{unary_ops}}, @BUILTIN_UNARY_OPS;
183 # rudimentary sanity-check for user supplied bits treated as functions/operators
184 # If a purported function matches this regular expression, an exception is thrown.
185 # Literal SQL is *NOT* subject to this check, only functions (and column names
186 # when quoting is not in effect)
189 # need to guard against ()'s in column names too, but this will break tons of
190 # hacks... ideas anyone?
191 $opt{injection_guard} ||= qr/
197 return bless \%opt, $class;
200 sub sqltrue { +{ -literal => [ $_[0]->{sqltrue} ] } }
201 sub sqlfalse { +{ -literal => [ $_[0]->{sqlfalse} ] } }
203 sub _assert_pass_injection_guard {
204 if ($_[1] =~ $_[0]->{injection_guard}) {
205 my $class = ref $_[0];
206 puke "Possible SQL injection attempt '$_[1]'. If this is indeed a part of the "
207 . "desired SQL use literal SQL ( \'...' or \[ '...' ] ) or supply your own "
208 . "{injection_guard} attribute to ${class}->new()"
213 #======================================================================
215 #======================================================================
219 my $table = $self->_table(shift);
220 my $data = shift || return;
223 my $method = $self->_METHOD_FOR_refkind("_insert", $data);
224 my ($sql, @bind) = $self->$method($data);
225 $sql = join " ", $self->_sqlcase('insert into'), $table, $sql;
227 if ($options->{returning}) {
228 my ($s, @b) = $self->_insert_returning($options);
233 return wantarray ? ($sql, @bind) : $sql;
236 # So that subclasses can override INSERT ... RETURNING separately from
237 # UPDATE and DELETE (e.g. DBIx::Class::SQLMaker::Oracle does this)
238 sub _insert_returning { shift->_returning(@_) }
241 my ($self, $options) = @_;
243 my $f = $options->{returning};
245 my $fieldlist = $self->_SWITCH_refkind($f, {
246 ARRAYREF => sub {join ', ', map { $self->_quote($_) } @$f;},
247 SCALAR => sub {$self->_quote($f)},
248 SCALARREF => sub {$$f},
250 return $self->_sqlcase(' returning ') . $fieldlist;
253 sub _insert_HASHREF { # explicit list of fields and then values
254 my ($self, $data) = @_;
256 my @fields = sort keys %$data;
258 my ($sql, @bind) = $self->_insert_values($data);
261 $_ = $self->_quote($_) foreach @fields;
262 $sql = "( ".join(", ", @fields).") ".$sql;
264 return ($sql, @bind);
267 sub _insert_ARRAYREF { # just generate values(?,?) part (no list of fields)
268 my ($self, $data) = @_;
270 # no names (arrayref) so can't generate bindtype
271 $self->{bindtype} ne 'columns'
272 or belch "can't do 'columns' bindtype when called with arrayref";
274 my (@values, @all_bind);
275 foreach my $value (@$data) {
276 my ($values, @bind) = $self->_insert_value(undef, $value);
277 push @values, $values;
278 push @all_bind, @bind;
280 my $sql = $self->_sqlcase('values')." ( ".join(", ", @values)." )";
281 return ($sql, @all_bind);
284 sub _insert_ARRAYREFREF { # literal SQL with bind
285 my ($self, $data) = @_;
287 my ($sql, @bind) = @${$data};
288 $self->_assert_bindval_matches_bindtype(@bind);
290 return ($sql, @bind);
294 sub _insert_SCALARREF { # literal SQL without bind
295 my ($self, $data) = @_;
301 my ($self, $data) = @_;
303 my (@values, @all_bind);
304 foreach my $column (sort keys %$data) {
305 my ($values, @bind) = $self->_insert_value($column, $data->{$column});
306 push @values, $values;
307 push @all_bind, @bind;
309 my $sql = $self->_sqlcase('values')." ( ".join(", ", @values)." )";
310 return ($sql, @all_bind);
314 my ($self, $column, $v) = @_;
316 my (@values, @all_bind);
317 $self->_SWITCH_refkind($v, {
320 if ($self->{array_datatypes}) { # if array datatype are activated
322 push @all_bind, $self->_bindtype($column, $v);
324 else { # else literal SQL with bind
325 my ($sql, @bind) = @$v;
326 $self->_assert_bindval_matches_bindtype(@bind);
328 push @all_bind, @bind;
332 ARRAYREFREF => sub { # literal SQL with bind
333 my ($sql, @bind) = @${$v};
334 $self->_assert_bindval_matches_bindtype(@bind);
336 push @all_bind, @bind;
339 # THINK: anything useful to do with a HASHREF ?
340 HASHREF => sub { # (nothing, but old SQLA passed it through)
341 #TODO in SQLA >= 2.0 it will die instead
342 belch "HASH ref as bind value in insert is not supported";
344 push @all_bind, $self->_bindtype($column, $v);
347 SCALARREF => sub { # literal SQL without bind
351 SCALAR_or_UNDEF => sub {
353 push @all_bind, $self->_bindtype($column, $v);
358 my $sql = join(", ", @values);
359 return ($sql, @all_bind);
364 #======================================================================
366 #======================================================================
371 my $table = $self->_table(shift);
372 my $data = shift || return;
376 # first build the 'SET' part of the sql statement
377 puke "Unsupported data type specified to \$sql->update"
378 unless ref $data eq 'HASH';
380 my ($sql, @all_bind) = $self->_update_set_values($data);
381 $sql = $self->_sqlcase('update ') . $table . $self->_sqlcase(' set ')
385 my($where_sql, @where_bind) = $self->where($where);
387 push @all_bind, @where_bind;
390 if ($options->{returning}) {
391 my ($returning_sql, @returning_bind) = $self->_update_returning($options);
392 $sql .= $returning_sql;
393 push @all_bind, @returning_bind;
396 return wantarray ? ($sql, @all_bind) : $sql;
399 sub _update_set_values {
400 my ($self, $data) = @_;
402 my (@set, @all_bind);
403 for my $k (sort keys %$data) {
406 my $label = $self->_quote($k);
408 $self->_SWITCH_refkind($v, {
410 if ($self->{array_datatypes}) { # array datatype
411 push @set, "$label = ?";
412 push @all_bind, $self->_bindtype($k, $v);
414 else { # literal SQL with bind
415 my ($sql, @bind) = @$v;
416 $self->_assert_bindval_matches_bindtype(@bind);
417 push @set, "$label = $sql";
418 push @all_bind, @bind;
421 ARRAYREFREF => sub { # literal SQL with bind
422 my ($sql, @bind) = @${$v};
423 $self->_assert_bindval_matches_bindtype(@bind);
424 push @set, "$label = $sql";
425 push @all_bind, @bind;
427 SCALARREF => sub { # literal SQL without bind
428 push @set, "$label = $$v";
431 my ($op, $arg, @rest) = %$v;
433 puke 'Operator calls in update must be in the form { -op => $arg }'
434 if (@rest or not $op =~ /^\-(.+)/);
436 local $self->{_nested_func_lhs} = $k;
437 my ($sql, @bind) = $self->_where_unary_op($1, $arg);
439 push @set, "$label = $sql";
440 push @all_bind, @bind;
442 SCALAR_or_UNDEF => sub {
443 push @set, "$label = ?";
444 push @all_bind, $self->_bindtype($k, $v);
450 my $sql = join ', ', @set;
452 return ($sql, @all_bind);
455 # So that subclasses can override UPDATE ... RETURNING separately from
457 sub _update_returning { shift->_returning(@_) }
461 #======================================================================
463 #======================================================================
468 my $table = $self->_table(shift);
469 my $fields = shift || '*';
473 my ($fields_sql, @bind) = $self->_select_fields($fields);
475 my ($where_sql, @where_bind) = $self->where($where, $order);
476 push @bind, @where_bind;
478 my $sql = join(' ', $self->_sqlcase('select'), $fields_sql,
479 $self->_sqlcase('from'), $table)
482 return wantarray ? ($sql, @bind) : $sql;
486 my ($self, $fields) = @_;
487 return ref $fields eq 'ARRAY' ? join ', ', map { $self->_quote($_) } @$fields
491 #======================================================================
493 #======================================================================
498 my $table = $self->_table(shift);
502 my($where_sql, @bind) = $self->where($where);
503 my $sql = $self->_sqlcase('delete from ') . $table . $where_sql;
505 if ($options->{returning}) {
506 my ($returning_sql, @returning_bind) = $self->_delete_returning($options);
507 $sql .= $returning_sql;
508 push @bind, @returning_bind;
511 return wantarray ? ($sql, @bind) : $sql;
514 # So that subclasses can override DELETE ... RETURNING separately from
516 sub _delete_returning { shift->_returning(@_) }
520 #======================================================================
522 #======================================================================
526 # Finally, a separate routine just to handle WHERE clauses
528 my ($self, $where, $order) = @_;
531 my ($sql, @bind) = defined($where)
532 ? $self->_recurse_where($where)
534 $sql = (defined $sql and length $sql) ? $self->_sqlcase(' where ') . "( $sql )" : '';
538 my ($order_sql, @order_bind) = $self->_order_by($order);
540 push @bind, @order_bind;
543 return wantarray ? ($sql, @bind) : $sql;
547 my ($self, $expr, $logic) = @_;
548 return undef unless defined($expr);
549 if (ref($expr) eq 'HASH') {
550 if (keys %$expr > 1) {
554 map $self->_expand_expr_hashpair($_ => $expr->{$_}, $logic),
558 return unless %$expr;
559 return $self->_expand_expr_hashpair(%$expr, $logic);
561 if (ref($expr) eq 'ARRAY') {
562 my $logic = lc($logic || $self->{logic});
563 $logic eq 'and' or $logic eq 'or' or puke "unknown logic: $logic";
569 while (my ($el) = splice @expr, 0, 1) {
570 puke "Supplying an empty left hand side argument is not supported in array-pairs"
571 unless defined($el) and length($el);
572 my $elref = ref($el);
574 push(@res, $self->_expand_expr({ $el, shift(@expr) }));
575 } elsif ($elref eq 'ARRAY') {
576 push(@res, $self->_expand_expr($el)) if @$el;
577 } elsif (my $l = is_literal_value($el)) {
578 push @res, { -literal => $l };
579 } elsif ($elref eq 'HASH') {
580 push @res, $self->_expand_expr($el);
585 return { -op => [ $logic, @res ] };
587 if (my $literal = is_literal_value($expr)) {
588 return +{ -literal => $literal };
590 if (!ref($expr) or Scalar::Util::blessed($expr)) {
591 if (my $m = our $Cur_Col_Meta) {
592 return +{ -bind => [ $m, $expr ] };
594 return +{ -value => $expr };
599 sub _expand_expr_hashpair {
600 my ($self, $k, $v, $logic) = @_;
601 unless (defined($k) and length($k)) {
602 if (defined($k) and my $literal = is_literal_value($v)) {
603 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
604 return { -literal => $literal };
606 puke "Supplying an empty left hand side argument is not supported";
609 $self->_assert_pass_injection_guard($k =~ /^-(.*)$/s);
610 if ($k =~ s/ [_\s]? \d+ $//x ) {
611 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
612 . "You probably wanted ...-and => [ $k => COND1, $k => COND2 ... ]";
615 return $self->_expand_expr($v);
619 return $self->_expand_expr($v);
621 puke "-bool => undef not supported" unless defined($v);
622 return { -ident => $v };
625 return { -op => [ 'not', $self->_expand_expr($v) ] };
627 if (my ($rest) = $k =~/^-not[_ ](.*)$/) {
630 $self->_expand_expr_hashpair("-${rest}", $v, $logic)
633 if (my ($logic) = $k =~ /^-(and|or)$/i) {
634 if (ref($v) eq 'HASH') {
635 return $self->_expand_expr($v, $logic);
637 if (ref($v) eq 'ARRAY') {
638 return $self->_expand_expr($v, $logic);
643 $op =~ s/^-// if length($op) > 1;
645 # top level special ops are illegal in general
646 puke "Illegal use of top-level '-$op'"
647 if !(defined $self->{_nested_func_lhs})
648 and List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}
649 and not List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}};
651 if ($k eq '-value' and my $m = our $Cur_Col_Meta) {
652 return +{ -bind => [ $m, $v ] };
654 if ($k eq '-op' or $k eq '-ident' or $k eq '-value' or $k eq '-bind' or $k eq '-literal' or $k eq '-func') {
660 and (keys %$v)[0] =~ /^-/
662 my ($func) = $k =~ /^-(.*)$/;
663 return +{ -func => [ $func, $self->_expand_expr($v) ] };
665 if (!ref($v) or is_literal_value($v)) {
666 return +{ -op => [ $k =~ /^-(.*)$/, $self->_expand_expr($v) ] };
673 and exists $v->{-value}
674 and not defined $v->{-value}
677 return $self->_expand_expr_hashpair($k => { $self->{cmp} => undef });
679 if (!ref($v) or Scalar::Util::blessed($v)) {
684 { -bind => [ $k, $v ] }
688 if (ref($v) eq 'HASH') {
692 map $self->_expand_expr_hashpair($k => { $_ => $v->{$_} }),
699 $self->_assert_pass_injection_guard($vk);
700 if ($vk =~ s/ [_\s]? \d+ $//x ) {
701 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
702 . "You probably wanted ...-and => [ -$vk => COND1, -$vk => COND2 ... ]";
704 if ($vk =~ /^(?:not[ _])?between$/) {
705 local our $Cur_Col_Meta = $k;
706 my @rhs = map $self->_expand_expr($_),
707 ref($vv) eq 'ARRAY' ? @$vv : $vv;
709 (@rhs == 1 and ref($rhs[0]) eq 'HASH' and $rhs[0]->{-literal})
711 (@rhs == 2 and defined($rhs[0]) and defined($rhs[1]))
713 puke "Operator '${\uc($vk)}' requires either an arrayref with two defined values or expressions, or a single literal scalarref/arrayref-ref";
716 join(' ', split '_', $vk),
721 if ($vk =~ /^(?:not[ _])?in$/) {
722 if (my $literal = is_literal_value($vv)) {
723 my ($sql, @bind) = @$literal;
724 my $opened_sql = $self->_open_outer_paren($sql);
726 $vk, { -ident => $k },
727 [ { -literal => [ $opened_sql, @bind ] } ]
731 'SQL::Abstract before v1.75 used to generate incorrect SQL when the '
732 . "-${\uc($vk)} operator was given an undef-containing list: !!!AUDIT YOUR CODE "
733 . 'AND DATA!!! (the upcoming Data::Query-based version of SQL::Abstract '
734 . 'will emit the logically correct SQL instead of raising this exception)'
736 puke("Argument passed to the '${\uc($vk)}' operator can not be undefined")
738 my @rhs = map $self->_expand_expr($_),
739 map { ref($_) ? $_ : { -bind => [ $k, $_ ] } }
740 map { defined($_) ? $_: puke($undef_err) }
741 (ref($vv) eq 'ARRAY' ? @$vv : $vv);
742 return $self->${\($vk =~ /^not/ ? 'sqltrue' : 'sqlfalse')} unless @rhs;
745 join(' ', split '_', $vk),
750 if ($vk eq 'ident') {
751 if (! defined $vv or ref $vv) {
752 puke "-$vk requires a single plain scalar argument (a quotable identifier)";
760 if ($vk eq 'value') {
761 return $self->_expand_expr_hashpair($k, undef) unless defined($vv);
765 { -bind => [ $k, $vv ] }
768 if ($vk =~ /^is(?:[ _]not)?$/) {
769 puke "$vk can only take undef as argument"
773 and exists($vv->{-value})
774 and !defined($vv->{-value})
777 return +{ -op => [ $vk.' null', { -ident => $k } ] };
779 if ($vk =~ /^(and|or)$/) {
780 if (ref($vv) eq 'HASH') {
783 map $self->_expand_expr_hashpair($k, { $_ => $vv->{$_} }),
788 if (my $us = List::Util::first { $vk =~ $_->{regex} } @{$self->{user_special_ops}}) {
789 return { -op => [ $vk, { -ident => $k }, $vv ] };
791 if (ref($vv) eq 'ARRAY') {
792 my ($logic, @values) = (
793 (defined($vv->[0]) and $vv->[0] =~ /^-(and|or)$/i)
798 $vk =~ $self->{inequality_op}
799 or join(' ', split '_', $vk) =~ $self->{not_like_op}
801 if (lc($logic) eq '-or' and @values > 1) {
802 my $op = uc join ' ', split '_', $vk;
803 belch "A multi-element arrayref as an argument to the inequality op '$op' "
804 . 'is technically equivalent to an always-true 1=1 (you probably wanted '
805 . "to say ...{ \$inequality_op => [ -and => \@values ] }... instead)"
810 # try to DWIM on equality operators
811 my $op = join ' ', split '_', $vk;
813 $op =~ $self->{equality_op} ? $self->sqlfalse
814 : $op =~ $self->{like_op} ? belch("Supplying an empty arrayref to '@{[ uc $op]}' is deprecated") && $self->sqlfalse
815 : $op =~ $self->{inequality_op} ? $self->sqltrue
816 : $op =~ $self->{not_like_op} ? belch("Supplying an empty arrayref to '@{[ uc $op]}' is deprecated") && $self->sqltrue
817 : puke "operator '$op' applied on an empty array (field '$k')";
821 map $self->_expand_expr_hashpair($k => { $vk => $_ }),
829 and exists $vv->{-value}
830 and not defined $vv->{-value}
833 my $op = join ' ', split '_', $vk;
835 $op =~ /^not$/i ? 'is not' # legacy
836 : $op =~ $self->{equality_op} ? 'is'
837 : $op =~ $self->{like_op} ? belch("Supplying an undefined argument to '@{[ uc $op]}' is deprecated") && 'is'
838 : $op =~ $self->{inequality_op} ? 'is not'
839 : $op =~ $self->{not_like_op} ? belch("Supplying an undefined argument to '@{[ uc $op]}' is deprecated") && 'is not'
840 : puke "unexpected operator '$op' with undef operand";
841 return +{ -op => [ $is.' null', { -ident => $k } ] };
843 local our $Cur_Col_Meta = $k;
847 $self->_expand_expr($vv)
850 if (ref($v) eq 'ARRAY') {
851 return $self->sqlfalse unless @$v;
852 $self->_debug("ARRAY($k) means distribute over elements");
854 $v->[0] =~ /^-((?:and|or))$/i
855 ? ($v = [ @{$v}[1..$#$v] ], $1)
856 : ($self->{logic} || 'or')
860 map $self->_expand_expr({ $k => $_ }, $this_logic), @$v
863 if (my $literal = is_literal_value($v)) {
865 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
868 my ($sql, @bind) = @$literal;
869 if ($self->{bindtype} eq 'columns') {
871 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
872 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
876 return +{ -literal => [ $self->_quote($k).' '.$sql, @bind ] };
882 my ($self, $expr) = @_;
883 my ($k, $v, @rest) = %$expr;
885 my %op = map +("-$_" => '_where_op_'.uc($_)),
886 qw(op func value bind ident literal);
887 if (my $meth = $op{$k}) {
888 return $self->$meth(undef, $v);
890 die "notreached: $k";
894 my ($self, $where, $logic) = @_;
896 #print STDERR Data::Dumper::Concise::Dumper([ $where, $logic ]);
898 my $where_exp = $self->_expand_expr($where, $logic);
900 #print STDERR Data::Dumper::Concise::Dumper([ EXP => $where_exp ]);
902 # dispatch on appropriate method according to refkind of $where
903 # my $method = $self->_METHOD_FOR_refkind("_where", $where_exp);
905 # my ($sql, @bind) = $self->$method($where_exp, $logic);
907 my ($sql, @bind) = defined($where_exp) ? $self->_render_expr($where_exp) : (undef);
909 # DBIx::Class used to call _recurse_where in scalar context
910 # something else might too...
912 return ($sql, @bind);
915 belch "Calling _recurse_where in scalar context is deprecated and will go away before 2.0";
922 #======================================================================
923 # WHERE: top-level ARRAYREF
924 #======================================================================
927 sub _where_ARRAYREF {
928 my ($self, $where, $logic) = @_;
930 $logic = uc($logic || $self->{logic});
931 $logic eq 'AND' or $logic eq 'OR' or puke "unknown logic: $logic";
933 my @clauses = @$where;
935 my (@sql_clauses, @all_bind);
936 # need to use while() so can shift() for pairs
938 my $el = shift @clauses;
940 $el = undef if (defined $el and ! length $el);
942 # switch according to kind of $el and get corresponding ($sql, @bind)
943 my ($sql, @bind) = $self->_SWITCH_refkind($el, {
945 # skip empty elements, otherwise get invalid trailing AND stuff
946 ARRAYREF => sub {$self->_recurse_where($el) if @$el},
950 $self->_assert_bindval_matches_bindtype(@b);
954 HASHREF => sub {$self->_recurse_where($el, 'and') if %$el},
956 SCALARREF => sub { ($$el); },
959 # top-level arrayref with scalars, recurse in pairs
960 $self->_recurse_where({$el => shift(@clauses)})
963 UNDEF => sub {puke "Supplying an empty left hand side argument is not supported in array-pairs" },
967 push @sql_clauses, $sql;
968 push @all_bind, @bind;
972 return $self->_join_sql_clauses($logic, \@sql_clauses, \@all_bind);
975 #======================================================================
976 # WHERE: top-level ARRAYREFREF
977 #======================================================================
979 sub _where_ARRAYREFREF {
980 my ($self, $where) = @_;
981 my ($sql, @bind) = @$$where;
982 $self->_assert_bindval_matches_bindtype(@bind);
983 return ($sql, @bind);
986 #======================================================================
987 # WHERE: top-level HASHREF
988 #======================================================================
991 my ($self, $where) = @_;
992 my (@sql_clauses, @all_bind);
994 for my $k (sort keys %$where) {
995 my $v = $where->{$k};
997 # ($k => $v) is either a special unary op or a regular hashpair
998 my ($sql, @bind) = do {
1000 # put the operator in canonical form
1002 $op = substr $op, 1; # remove initial dash
1003 $op =~ s/^\s+|\s+$//g;# remove leading/trailing space
1004 $op =~ s/\s+/ /g; # compress whitespace
1006 # so that -not_foo works correctly
1007 $op =~ s/^not_/NOT /i;
1009 $self->_debug("Unary OP(-$op) within hashref, recursing...");
1010 my ($s, @b) = $self->_where_unary_op($op, $v);
1012 # top level vs nested
1013 # we assume that handled unary ops will take care of their ()s
1014 $s = "($s)" unless (
1015 List::Util::first {$op =~ $_->{regex}} @{$self->{unary_ops}}
1017 ( defined $self->{_nested_func_lhs} and $self->{_nested_func_lhs} eq $k )
1023 if (is_literal_value ($v) ) {
1024 belch 'Hash-pairs consisting of an empty string with a literal are deprecated, and will be removed in 2.0: use -and => [ $literal ] instead';
1027 puke "Supplying an empty left hand side argument is not supported in hash-pairs";
1031 my $method = $self->_METHOD_FOR_refkind("_where_hashpair", $v);
1032 $self->$method($k, $v);
1036 push @sql_clauses, $sql;
1037 push @all_bind, @bind;
1040 return $self->_join_sql_clauses('and', \@sql_clauses, \@all_bind);
1043 sub _where_unary_op {
1044 my ($self, $op, $rhs) = @_;
1046 $op =~ s/^-// if length($op) > 1;
1048 # top level special ops are illegal in general
1049 puke "Illegal use of top-level '-$op'"
1050 if !(defined $self->{_nested_func_lhs})
1051 and List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}
1052 and not List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}};
1054 if (my $op_entry = List::Util::first { $op =~ $_->{regex} } @{$self->{unary_ops}}) {
1055 my $handler = $op_entry->{handler};
1057 if (not ref $handler) {
1058 if ($op =~ s/ [_\s]? \d+ $//x ) {
1059 belch 'Use of [and|or|nest]_N modifiers is deprecated and will be removed in SQLA v2.0. '
1060 . "You probably wanted ...-and => [ -$op => COND1, -$op => COND2 ... ]";
1062 return $self->$handler($op, $rhs);
1064 elsif (ref $handler eq 'CODE') {
1065 return $handler->($self, $op, $rhs);
1068 puke "Illegal handler for operator $op - expecting a method name or a coderef";
1072 $self->_debug("Generic unary OP: $op - recursing as function");
1074 $self->_assert_pass_injection_guard($op);
1076 my ($sql, @bind) = $self->_SWITCH_refkind($rhs, {
1078 puke "Illegal use of top-level '-$op'"
1079 unless defined $self->{_nested_func_lhs};
1082 $self->_convert('?'),
1083 $self->_bindtype($self->{_nested_func_lhs}, $rhs)
1087 $self->_recurse_where($rhs)
1091 $sql = sprintf('%s %s',
1092 $self->_sqlcase($op),
1096 return ($sql, @bind);
1099 sub _where_op_NEST {
1100 my ($self, $op, $v) = @_;
1102 $self->_SWITCH_refkind($v, {
1104 SCALAR => sub { # permissively interpreted as SQL
1105 belch "literal SQL should be -nest => \\'scalar' "
1106 . "instead of -nest => 'scalar' ";
1111 puke "-$op => undef not supported";
1115 $self->_recurse_where($v);
1122 sub _where_op_BOOL {
1123 my ($self, $op, $v) = @_;
1125 my ($s, @b) = $self->_SWITCH_refkind($v, {
1126 SCALAR => sub { # interpreted as SQL column
1127 $self->_convert($self->_quote($v));
1131 puke "-$op => undef not supported";
1135 $self->_recurse_where($v);
1139 $s = "(NOT $s)" if $op =~ /^not/i;
1144 sub _where_op_IDENT {
1146 my ($op, $rhs) = splice @_, -2;
1147 if (! defined $rhs or length ref $rhs) {
1148 puke "-$op requires a single plain scalar argument (a quotable identifier)";
1151 # in case we are called as a top level special op (no '=')
1152 my $has_lhs = my $lhs = shift;
1154 $_ = $self->_convert($self->_quote($_)) for ($lhs, $rhs);
1162 sub _where_op_VALUE {
1164 my ($op, $rhs) = splice @_, -2;
1166 # in case we are called as a top level special op (no '=')
1170 if (! defined $rhs) {
1172 ? $self->_where_hashpair_HASHREF($lhs, { -is => undef })
1179 (defined $lhs ? $lhs : $self->{_nested_func_lhs}),
1186 $self->_convert($self->_quote($lhs)) . ' = ' . $self->_convert('?'),
1190 $self->_convert('?'),
1197 my %unop_postfix = map +($_ => 1), 'is null', 'is not null';
1203 my ($self, $args) = @_;
1204 my ($left, $low, $high) = @$args;
1205 my ($rhsql, @rhbind) = do {
1207 puke "Single arg to between must be a literal"
1208 unless $low->{-literal};
1211 my ($l, $h) = map [ $self->_render_expr($_) ], $low, $high;
1212 (join(' ', $l->[0], $self->_sqlcase('and'), $h->[0]),
1213 @{$l}[1..$#$l], @{$h}[1..$#$h])
1216 my ($lhsql, @lhbind) = $self->_render_expr($left);
1218 join(' ', '(', $lhsql, $self->_sqlcase($op), $rhsql, ')'),
1222 }), 'between', 'not between'),
1226 my ($self, $args) = @_;
1227 my ($lhs, $rhs) = @$args;
1230 my ($sql, @bind) = $self->_render_expr($_);
1231 push @in_bind, @bind;
1234 my ($lhsql, @lbind) = $self->_render_expr($lhs);
1236 $lhsql.' '.$self->_sqlcase($op).' ( '
1237 .join(', ', @in_sql)
1242 }), 'in', 'not in'),
1246 my ($self, undef, $v) = @_;
1247 my ($op, @args) = @$v;
1248 $op =~ s/^-// if length($op) > 1;
1250 local $self->{_nested_func_lhs};
1251 if (my $h = $special{$op}) {
1252 return $self->$h(\@args);
1254 if (my $us = List::Util::first { $op =~ $_->{regex} } @{$self->{user_special_ops}}) {
1255 puke "Special op '${op}' requires first value to be identifier"
1256 unless my ($k) = map $_->{-ident}, grep ref($_) eq 'HASH', $args[0];
1257 return $self->${\($us->{handler})}($k, $op, $args[1]);
1259 my $final_op = $op =~ /^(?:is|not)_/ ? join(' ', split '_', $op) : $op;
1260 if (@args == 1 and $op !~ /^(and|or)$/) {
1261 my ($expr_sql, @bind) = $self->_render_expr($args[0]);
1262 my $op_sql = $self->_sqlcase($final_op);
1264 $unop_postfix{lc($final_op)}
1265 ? "${expr_sql} ${op_sql}"
1266 : "${op_sql} ${expr_sql}"
1268 return (($op eq 'not' ? '('.$final_sql.')' : $final_sql), @bind);
1270 my @parts = map [ $self->_render_expr($_) ], @args;
1271 my ($final_sql) = map +($op =~ /^(and|or)$/ ? "(${_})" : $_), join(
1272 ' '.$self->_sqlcase($final_op).' ',
1277 map @{$_}[1..$#$_], @parts
1283 sub _where_op_FUNC {
1284 my ($self, undef, $rest) = @_;
1285 my ($func, @args) = @$rest;
1289 push @arg_sql, shift @x;
1291 } map [ $self->_render_expr($_) ], @args;
1292 return ($self->_sqlcase($func).'('.join(', ', @arg_sql).')', @bind);
1295 sub _where_op_BIND {
1296 my ($self, undef, $bind) = @_;
1297 return ($self->_convert('?'), $self->_bindtype(@$bind));
1300 sub _where_op_LITERAL {
1301 my ($self, undef, $literal) = @_;
1302 $self->_assert_bindval_matches_bindtype(@{$literal}[1..$#$literal]);
1306 sub _where_hashpair_ARRAYREF {
1307 my ($self, $k, $v) = @_;
1310 my @v = @$v; # need copy because of shift below
1311 $self->_debug("ARRAY($k) means distribute over elements");
1313 # put apart first element if it is an operator (-and, -or)
1315 (defined $v[0] && $v[0] =~ /^ - (?: AND|OR ) $/ix)
1319 my @distributed = map { {$k => $_} } @v;
1322 $self->_debug("OP($op) reinjected into the distributed array");
1323 unshift @distributed, $op;
1326 my $logic = $op ? substr($op, 1) : '';
1328 return $self->_recurse_where(\@distributed, $logic);
1331 $self->_debug("empty ARRAY($k) means 0=1");
1332 return ($self->{sqlfalse});
1336 sub _where_hashpair_HASHREF {
1337 my ($self, $k, $v, $logic) = @_;
1340 local $self->{_nested_func_lhs} = defined $self->{_nested_func_lhs}
1341 ? $self->{_nested_func_lhs}
1345 my ($all_sql, @all_bind);
1347 for my $orig_op (sort keys %$v) {
1348 my $val = $v->{$orig_op};
1350 # put the operator in canonical form
1353 # FIXME - we need to phase out dash-less ops
1354 $op =~ s/^-//; # remove possible initial dash
1355 $op =~ s/^\s+|\s+$//g;# remove leading/trailing space
1356 $op =~ s/\s+/ /g; # compress whitespace
1358 $self->_assert_pass_injection_guard($op);
1361 $op =~ s/^is_not/IS NOT/i;
1363 # so that -not_foo works correctly
1364 $op =~ s/^not_/NOT /i;
1366 # another retarded special case: foo => { $op => { -value => undef } }
1367 if (ref $val eq 'HASH' and keys %$val == 1 and exists $val->{-value} and ! defined $val->{-value} ) {
1373 # CASE: col-value logic modifiers
1374 if ($orig_op =~ /^ \- (and|or) $/xi) {
1375 ($sql, @bind) = $self->_where_hashpair_HASHREF($k, $val, $1);
1377 # CASE: special operators like -in or -between
1378 elsif (my $special_op = List::Util::first { $op =~ $_->{regex} } @{$self->{special_ops}}) {
1379 my $handler = $special_op->{handler};
1381 puke "No handler supplied for special operator $orig_op";
1383 elsif (not ref $handler) {
1384 ($sql, @bind) = $self->$handler($k, $op, $val);
1386 elsif (ref $handler eq 'CODE') {
1387 ($sql, @bind) = $handler->($self, $k, $op, $val);
1390 puke "Illegal handler for special operator $orig_op - expecting a method name or a coderef";
1394 $self->_SWITCH_refkind($val, {
1396 ARRAYREF => sub { # CASE: col => {op => \@vals}
1397 ($sql, @bind) = $self->_where_field_op_ARRAYREF($k, $op, $val);
1400 ARRAYREFREF => sub { # CASE: col => {op => \[$sql, @bind]} (literal SQL with bind)
1401 my ($sub_sql, @sub_bind) = @$$val;
1402 $self->_assert_bindval_matches_bindtype(@sub_bind);
1403 $sql = join ' ', $self->_convert($self->_quote($k)),
1404 $self->_sqlcase($op),
1409 UNDEF => sub { # CASE: col => {op => undef} : sql "IS (NOT)? NULL"
1411 $op =~ /^not$/i ? 'is not' # legacy
1412 : $op =~ $self->{equality_op} ? 'is'
1413 : $op =~ $self->{like_op} ? belch("Supplying an undefined argument to '@{[ uc $op]}' is deprecated") && 'is'
1414 : $op =~ $self->{inequality_op} ? 'is not'
1415 : $op =~ $self->{not_like_op} ? belch("Supplying an undefined argument to '@{[ uc $op]}' is deprecated") && 'is not'
1416 : puke "unexpected operator '$orig_op' with undef operand";
1418 $sql = $self->_quote($k) . $self->_sqlcase(" $is null");
1421 FALLBACK => sub { # CASE: col => {op/func => $stuff}
1422 ($sql, @bind) = $self->_where_unary_op($op, $val);
1425 $self->_convert($self->_quote($k)),
1426 $self->{_nested_func_lhs} eq $k ? $sql : "($sql)", # top level vs nested
1432 ($all_sql) = (defined $all_sql and $all_sql) ? $self->_join_sql_clauses($logic, [$all_sql, $sql], []) : $sql;
1433 push @all_bind, @bind;
1435 return ($all_sql, @all_bind);
1438 sub _where_field_IS {
1439 my ($self, $k, $op, $v) = @_;
1441 my ($s) = $self->_SWITCH_refkind($v, {
1444 $self->_convert($self->_quote($k)),
1445 map { $self->_sqlcase($_)} ($op, 'null')
1448 puke "$op can only take undef as argument";
1455 sub _where_field_op_ARRAYREF {
1456 my ($self, $k, $op, $vals) = @_;
1458 my @vals = @$vals; #always work on a copy
1461 $self->_debug(sprintf '%s means multiple elements: [ %s ]',
1463 join(', ', map { defined $_ ? "'$_'" : 'NULL' } @vals ),
1466 # see if the first element is an -and/-or op
1468 if (defined $vals[0] && $vals[0] =~ /^ - (AND|OR) $/ix) {
1473 # a long standing API wart - an attempt to change this behavior during
1474 # the 1.50 series failed *spectacularly*. Warn instead and leave the
1479 (!$logic or $logic eq 'OR')
1481 ($op =~ $self->{inequality_op} or $op =~ $self->{not_like_op})
1484 belch "A multi-element arrayref as an argument to the inequality op '$o' "
1485 . 'is technically equivalent to an always-true 1=1 (you probably wanted '
1486 . "to say ...{ \$inequality_op => [ -and => \@values ] }... instead)"
1490 # distribute $op over each remaining member of @vals, append logic if exists
1491 return $self->_recurse_where([map { {$k => {$op, $_}} } @vals], $logic);
1495 # try to DWIM on equality operators
1497 $op =~ $self->{equality_op} ? $self->{sqlfalse}
1498 : $op =~ $self->{like_op} ? belch("Supplying an empty arrayref to '@{[ uc $op]}' is deprecated") && $self->{sqlfalse}
1499 : $op =~ $self->{inequality_op} ? $self->{sqltrue}
1500 : $op =~ $self->{not_like_op} ? belch("Supplying an empty arrayref to '@{[ uc $op]}' is deprecated") && $self->{sqltrue}
1501 : puke "operator '$op' applied on an empty array (field '$k')";
1506 sub _where_hashpair_SCALARREF {
1507 my ($self, $k, $v) = @_;
1508 $self->_debug("SCALAR($k) means literal SQL: $$v");
1509 my $sql = $self->_quote($k) . " " . $$v;
1513 # literal SQL with bind
1514 sub _where_hashpair_ARRAYREFREF {
1515 my ($self, $k, $v) = @_;
1516 $self->_debug("REF($k) means literal SQL: @${$v}");
1517 my ($sql, @bind) = @$$v;
1518 $self->_assert_bindval_matches_bindtype(@bind);
1519 $sql = $self->_quote($k) . " " . $sql;
1520 return ($sql, @bind );
1523 # literal SQL without bind
1524 sub _where_hashpair_SCALAR {
1525 my ($self, $k, $v) = @_;
1526 $self->_debug("NOREF($k) means simple key=val: $k $self->{cmp} $v");
1527 return ($self->_where_hashpair_HASHREF($k, { $self->{cmp} => $v }));
1531 sub _where_hashpair_UNDEF {
1532 my ($self, $k, $v) = @_;
1533 $self->_debug("UNDEF($k) means IS NULL");
1534 return $self->_where_hashpair_HASHREF($k, { -is => undef });
1537 #======================================================================
1538 # WHERE: TOP-LEVEL OTHERS (SCALARREF, SCALAR, UNDEF)
1539 #======================================================================
1542 sub _where_SCALARREF {
1543 my ($self, $where) = @_;
1546 $self->_debug("SCALAR(*top) means literal SQL: $$where");
1552 my ($self, $where) = @_;
1555 $self->_debug("NOREF(*top) means literal SQL: $where");
1566 #======================================================================
1567 # WHERE: BUILTIN SPECIAL OPERATORS (-in, -between)
1568 #======================================================================
1571 sub _where_field_BETWEEN {
1572 my ($self, $k, $op, $vals) = @_;
1574 my ($label, $and, $placeholder);
1575 $label = $self->_convert($self->_quote($k));
1576 $and = ' ' . $self->_sqlcase('and') . ' ';
1577 $placeholder = $self->_convert('?');
1578 $op = $self->_sqlcase($op);
1580 my $invalid_args = "Operator '$op' requires either an arrayref with two defined values or expressions, or a single literal scalarref/arrayref-ref";
1582 my ($clause, @bind) = $self->_SWITCH_refkind($vals, {
1583 ARRAYREFREF => sub {
1584 my ($s, @b) = @$$vals;
1585 $self->_assert_bindval_matches_bindtype(@b);
1592 puke $invalid_args if @$vals != 2;
1594 my (@all_sql, @all_bind);
1595 foreach my $val (@$vals) {
1596 my ($sql, @bind) = $self->_SWITCH_refkind($val, {
1598 return ($placeholder, $self->_bindtype($k, $val) );
1603 ARRAYREFREF => sub {
1604 my ($sql, @bind) = @$$val;
1605 $self->_assert_bindval_matches_bindtype(@bind);
1606 return ($sql, @bind);
1609 my ($func, $arg, @rest) = %$val;
1610 puke "Only simple { -func => arg } functions accepted as sub-arguments to BETWEEN"
1611 if (@rest or $func !~ /^ \- (.+)/x);
1612 $self->_where_unary_op($1 => $arg);
1618 push @all_sql, $sql;
1619 push @all_bind, @bind;
1623 (join $and, @all_sql),
1632 my $sql = "( $label $op $clause )";
1633 return ($sql, @bind)
1637 sub _where_field_IN {
1638 my ($self, $k, $op, $vals) = @_;
1640 # backwards compatibility: if scalar, force into an arrayref
1641 $vals = [$vals] if defined $vals && ! ref $vals;
1643 my ($label) = $self->_convert($self->_quote($k));
1644 my ($placeholder) = $self->_convert('?');
1645 $op = $self->_sqlcase($op);
1647 my ($sql, @bind) = $self->_SWITCH_refkind($vals, {
1648 ARRAYREF => sub { # list of choices
1649 if (@$vals) { # nonempty list
1650 my (@all_sql, @all_bind);
1652 for my $val (@$vals) {
1653 my ($sql, @bind) = $self->_SWITCH_refkind($val, {
1655 return ($placeholder, $val);
1660 ARRAYREFREF => sub {
1661 my ($sql, @bind) = @$$val;
1662 $self->_assert_bindval_matches_bindtype(@bind);
1663 return ($sql, @bind);
1666 my ($func, $arg, @rest) = %$val;
1667 puke "Only simple { -func => arg } functions accepted as sub-arguments to IN"
1668 if (@rest or $func !~ /^ \- (.+)/x);
1669 $self->_where_unary_op($1 => $arg);
1673 'SQL::Abstract before v1.75 used to generate incorrect SQL when the '
1674 . "-$op operator was given an undef-containing list: !!!AUDIT YOUR CODE "
1675 . 'AND DATA!!! (the upcoming Data::Query-based version of SQL::Abstract '
1676 . 'will emit the logically correct SQL instead of raising this exception)'
1680 push @all_sql, $sql;
1681 push @all_bind, @bind;
1685 sprintf('%s %s ( %s )',
1688 join(', ', @all_sql)
1690 $self->_bindtype($k, @all_bind),
1693 else { # empty list: some databases won't understand "IN ()", so DWIM
1694 my $sql = ($op =~ /\bnot\b/i) ? $self->{sqltrue} : $self->{sqlfalse};
1699 SCALARREF => sub { # literal SQL
1700 my $sql = $self->_open_outer_paren($$vals);
1701 return ("$label $op ( $sql )");
1703 ARRAYREFREF => sub { # literal SQL with bind
1704 my ($sql, @bind) = @$$vals;
1705 $self->_assert_bindval_matches_bindtype(@bind);
1706 $sql = $self->_open_outer_paren($sql);
1707 return ("$label $op ( $sql )", @bind);
1711 puke "Argument passed to the '$op' operator can not be undefined";
1715 puke "special op $op requires an arrayref (or scalarref/arrayref-ref)";
1719 return ($sql, @bind);
1722 # Some databases (SQLite) treat col IN (1, 2) different from
1723 # col IN ( (1, 2) ). Use this to strip all outer parens while
1724 # adding them back in the corresponding method
1725 sub _open_outer_paren {
1726 my ($self, $sql) = @_;
1728 while (my ($inner) = $sql =~ /^ \s* \( (.*) \) \s* $/xs) {
1730 # there are closing parens inside, need the heavy duty machinery
1731 # to reevaluate the extraction starting from $sql (full reevaluation)
1732 if ($inner =~ /\)/) {
1733 require Text::Balanced;
1735 my (undef, $remainder) = do {
1736 # idiotic design - writes to $@ but *DOES NOT* throw exceptions
1738 Text::Balanced::extract_bracketed($sql, '()', qr/\s*/);
1741 # the entire expression needs to be a balanced bracketed thing
1742 # (after an extract no remainder sans trailing space)
1743 last if defined $remainder and $remainder =~ /\S/;
1753 #======================================================================
1755 #======================================================================
1758 my ($self, $arg) = @_;
1761 for my $c ($self->_order_by_chunks($arg) ) {
1762 $self->_SWITCH_refkind($c, {
1763 SCALAR => sub { push @sql, $c },
1764 ARRAYREF => sub { push @sql, shift @$c; push @bind, @$c },
1770 $self->_sqlcase(' order by'),
1776 return wantarray ? ($sql, @bind) : $sql;
1779 sub _order_by_chunks {
1780 my ($self, $arg) = @_;
1782 return $self->_SWITCH_refkind($arg, {
1785 map { $self->_order_by_chunks($_ ) } @$arg;
1788 ARRAYREFREF => sub {
1789 my ($s, @b) = @$$arg;
1790 $self->_assert_bindval_matches_bindtype(@b);
1794 SCALAR => sub {$self->_quote($arg)},
1796 UNDEF => sub {return () },
1798 SCALARREF => sub {$$arg}, # literal SQL, no quoting
1801 # get first pair in hash
1802 my ($key, $val, @rest) = %$arg;
1804 return () unless $key;
1806 if (@rest or not $key =~ /^-(desc|asc)/i) {
1807 puke "hash passed to _order_by must have exactly one key (-desc or -asc)";
1813 for my $c ($self->_order_by_chunks($val)) {
1816 $self->_SWITCH_refkind($c, {
1821 ($sql, @bind) = @$c;
1825 $sql = $sql . ' ' . $self->_sqlcase($direction);
1827 push @ret, [ $sql, @bind];
1836 #======================================================================
1837 # DATASOURCE (FOR NOW, JUST PLAIN TABLE OR LIST OF TABLES)
1838 #======================================================================
1843 $self->_SWITCH_refkind($from, {
1844 ARRAYREF => sub {join ', ', map { $self->_quote($_) } @$from;},
1845 SCALAR => sub {$self->_quote($from)},
1846 SCALARREF => sub {$$from},
1851 #======================================================================
1853 #======================================================================
1855 # highly optimized, as it's called way too often
1857 # my ($self, $label) = @_;
1859 return '' unless defined $_[1];
1860 return ${$_[1]} if ref($_[1]) eq 'SCALAR';
1862 $_[0]->{quote_char} or
1863 ($_[0]->_assert_pass_injection_guard($_[1]), return $_[1]);
1865 my $qref = ref $_[0]->{quote_char};
1867 !$qref ? ($_[0]->{quote_char}, $_[0]->{quote_char})
1868 : ($qref eq 'ARRAY') ? @{$_[0]->{quote_char}}
1869 : puke "Unsupported quote_char format: $_[0]->{quote_char}";
1871 my $esc = $_[0]->{escape_char} || $r;
1873 # parts containing * are naturally unquoted
1874 return join($_[0]->{name_sep}||'', map
1875 +( $_ eq '*' ? $_ : do { (my $n = $_) =~ s/(\Q$esc\E|\Q$r\E)/$esc$1/g; $l . $n . $r } ),
1876 ( $_[0]->{name_sep} ? split (/\Q$_[0]->{name_sep}\E/, $_[1] ) : $_[1] )
1881 # Conversion, if applicable
1883 #my ($self, $arg) = @_;
1884 if ($_[0]->{convert}) {
1885 return $_[0]->_sqlcase($_[0]->{convert}) .'(' . $_[1] . ')';
1892 #my ($self, $col, @vals) = @_;
1893 # called often - tighten code
1894 return $_[0]->{bindtype} eq 'columns'
1895 ? map {[$_[1], $_]} @_[2 .. $#_]
1900 # Dies if any element of @bind is not in [colname => value] format
1901 # if bindtype is 'columns'.
1902 sub _assert_bindval_matches_bindtype {
1903 # my ($self, @bind) = @_;
1905 if ($self->{bindtype} eq 'columns') {
1907 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
1908 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
1914 sub _join_sql_clauses {
1915 my ($self, $logic, $clauses_aref, $bind_aref) = @_;
1917 if (@$clauses_aref > 1) {
1918 my $join = " " . $self->_sqlcase($logic) . " ";
1919 my $sql = '( ' . join($join, @$clauses_aref) . ' )';
1920 return ($sql, @$bind_aref);
1922 elsif (@$clauses_aref) {
1923 return ($clauses_aref->[0], @$bind_aref); # no parentheses
1926 return (); # if no SQL, ignore @$bind_aref
1931 # Fix SQL case, if so requested
1933 # LDNOTE: if $self->{case} is true, then it contains 'lower', so we
1934 # don't touch the argument ... crooked logic, but let's not change it!
1935 return $_[0]->{case} ? $_[1] : uc($_[1]);
1939 #======================================================================
1940 # DISPATCHING FROM REFKIND
1941 #======================================================================
1944 my ($self, $data) = @_;
1946 return 'UNDEF' unless defined $data;
1948 # blessed objects are treated like scalars
1949 my $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1951 return 'SCALAR' unless $ref;
1954 while ($ref eq 'REF') {
1956 $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
1960 return ($ref||'SCALAR') . ('REF' x $n_steps);
1964 my ($self, $data) = @_;
1965 my @try = ($self->_refkind($data));
1966 push @try, 'SCALAR_or_UNDEF' if $try[0] eq 'SCALAR' || $try[0] eq 'UNDEF';
1967 push @try, 'FALLBACK';
1971 sub _METHOD_FOR_refkind {
1972 my ($self, $meth_prefix, $data) = @_;
1975 for (@{$self->_try_refkind($data)}) {
1976 $method = $self->can($meth_prefix."_".$_)
1980 return $method || puke "cannot dispatch on '$meth_prefix' for ".$self->_refkind($data);
1984 sub _SWITCH_refkind {
1985 my ($self, $data, $dispatch_table) = @_;
1988 for (@{$self->_try_refkind($data)}) {
1989 $coderef = $dispatch_table->{$_}
1993 puke "no dispatch entry for ".$self->_refkind($data)
2002 #======================================================================
2003 # VALUES, GENERATE, AUTOLOAD
2004 #======================================================================
2006 # LDNOTE: original code from nwiger, didn't touch code in that section
2007 # I feel the AUTOLOAD stuff should not be the default, it should
2008 # only be activated on explicit demand by user.
2012 my $data = shift || return;
2013 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
2014 unless ref $data eq 'HASH';
2017 foreach my $k (sort keys %$data) {
2018 my $v = $data->{$k};
2019 $self->_SWITCH_refkind($v, {
2021 if ($self->{array_datatypes}) { # array datatype
2022 push @all_bind, $self->_bindtype($k, $v);
2024 else { # literal SQL with bind
2025 my ($sql, @bind) = @$v;
2026 $self->_assert_bindval_matches_bindtype(@bind);
2027 push @all_bind, @bind;
2030 ARRAYREFREF => sub { # literal SQL with bind
2031 my ($sql, @bind) = @${$v};
2032 $self->_assert_bindval_matches_bindtype(@bind);
2033 push @all_bind, @bind;
2035 SCALARREF => sub { # literal SQL without bind
2037 SCALAR_or_UNDEF => sub {
2038 push @all_bind, $self->_bindtype($k, $v);
2049 my(@sql, @sqlq, @sqlv);
2053 if ($ref eq 'HASH') {
2054 for my $k (sort keys %$_) {
2057 my $label = $self->_quote($k);
2058 if ($r eq 'ARRAY') {
2059 # literal SQL with bind
2060 my ($sql, @bind) = @$v;
2061 $self->_assert_bindval_matches_bindtype(@bind);
2062 push @sqlq, "$label = $sql";
2064 } elsif ($r eq 'SCALAR') {
2065 # literal SQL without bind
2066 push @sqlq, "$label = $$v";
2068 push @sqlq, "$label = ?";
2069 push @sqlv, $self->_bindtype($k, $v);
2072 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
2073 } elsif ($ref eq 'ARRAY') {
2074 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
2077 if ($r eq 'ARRAY') { # literal SQL with bind
2078 my ($sql, @bind) = @$v;
2079 $self->_assert_bindval_matches_bindtype(@bind);
2082 } elsif ($r eq 'SCALAR') { # literal SQL without bind
2083 # embedded literal SQL
2090 push @sql, '(' . join(', ', @sqlq) . ')';
2091 } elsif ($ref eq 'SCALAR') {
2095 # strings get case twiddled
2096 push @sql, $self->_sqlcase($_);
2100 my $sql = join ' ', @sql;
2102 # this is pretty tricky
2103 # if ask for an array, return ($stmt, @bind)
2104 # otherwise, s/?/shift @sqlv/ to put it inline
2106 return ($sql, @sqlv);
2108 1 while $sql =~ s/\?/my $d = shift(@sqlv);
2109 ref $d ? $d->[1] : $d/e;
2118 # This allows us to check for a local, then _form, attr
2120 my($name) = $AUTOLOAD =~ /.*::(.+)/;
2121 return $self->generate($name, @_);
2132 SQL::Abstract - Generate SQL from Perl data structures
2138 my $sql = SQL::Abstract->new;
2140 my($stmt, @bind) = $sql->select($source, \@fields, \%where, $order);
2142 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
2144 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
2146 my($stmt, @bind) = $sql->delete($table, \%where);
2148 # Then, use these in your DBI statements
2149 my $sth = $dbh->prepare($stmt);
2150 $sth->execute(@bind);
2152 # Just generate the WHERE clause
2153 my($stmt, @bind) = $sql->where(\%where, $order);
2155 # Return values in the same order, for hashed queries
2156 # See PERFORMANCE section for more details
2157 my @bind = $sql->values(\%fieldvals);
2161 This module was inspired by the excellent L<DBIx::Abstract>.
2162 However, in using that module I found that what I really wanted
2163 to do was generate SQL, but still retain complete control over my
2164 statement handles and use the DBI interface. So, I set out to
2165 create an abstract SQL generation module.
2167 While based on the concepts used by L<DBIx::Abstract>, there are
2168 several important differences, especially when it comes to WHERE
2169 clauses. I have modified the concepts used to make the SQL easier
2170 to generate from Perl data structures and, IMO, more intuitive.
2171 The underlying idea is for this module to do what you mean, based
2172 on the data structures you provide it. The big advantage is that
2173 you don't have to modify your code every time your data changes,
2174 as this module figures it out.
2176 To begin with, an SQL INSERT is as easy as just specifying a hash
2177 of C<key=value> pairs:
2180 name => 'Jimbo Bobson',
2181 phone => '123-456-7890',
2182 address => '42 Sister Lane',
2183 city => 'St. Louis',
2184 state => 'Louisiana',
2187 The SQL can then be generated with this:
2189 my($stmt, @bind) = $sql->insert('people', \%data);
2191 Which would give you something like this:
2193 $stmt = "INSERT INTO people
2194 (address, city, name, phone, state)
2195 VALUES (?, ?, ?, ?, ?)";
2196 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
2197 '123-456-7890', 'Louisiana');
2199 These are then used directly in your DBI code:
2201 my $sth = $dbh->prepare($stmt);
2202 $sth->execute(@bind);
2204 =head2 Inserting and Updating Arrays
2206 If your database has array types (like for example Postgres),
2207 activate the special option C<< array_datatypes => 1 >>
2208 when creating the C<SQL::Abstract> object.
2209 Then you may use an arrayref to insert and update database array types:
2211 my $sql = SQL::Abstract->new(array_datatypes => 1);
2213 planets => [qw/Mercury Venus Earth Mars/]
2216 my($stmt, @bind) = $sql->insert('solar_system', \%data);
2220 $stmt = "INSERT INTO solar_system (planets) VALUES (?)"
2222 @bind = (['Mercury', 'Venus', 'Earth', 'Mars']);
2225 =head2 Inserting and Updating SQL
2227 In order to apply SQL functions to elements of your C<%data> you may
2228 specify a reference to an arrayref for the given hash value. For example,
2229 if you need to execute the Oracle C<to_date> function on a value, you can
2230 say something like this:
2234 date_entered => \[ "to_date(?,'MM/DD/YYYY')", "03/02/2003" ],
2237 The first value in the array is the actual SQL. Any other values are
2238 optional and would be included in the bind values array. This gives
2241 my($stmt, @bind) = $sql->insert('people', \%data);
2243 $stmt = "INSERT INTO people (name, date_entered)
2244 VALUES (?, to_date(?,'MM/DD/YYYY'))";
2245 @bind = ('Bill', '03/02/2003');
2247 An UPDATE is just as easy, all you change is the name of the function:
2249 my($stmt, @bind) = $sql->update('people', \%data);
2251 Notice that your C<%data> isn't touched; the module will generate
2252 the appropriately quirky SQL for you automatically. Usually you'll
2253 want to specify a WHERE clause for your UPDATE, though, which is
2254 where handling C<%where> hashes comes in handy...
2256 =head2 Complex where statements
2258 This module can generate pretty complicated WHERE statements
2259 easily. For example, simple C<key=value> pairs are taken to mean
2260 equality, and if you want to see if a field is within a set
2261 of values, you can use an arrayref. Let's say we wanted to
2262 SELECT some data based on this criteria:
2265 requestor => 'inna',
2266 worker => ['nwiger', 'rcwe', 'sfz'],
2267 status => { '!=', 'completed' }
2270 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
2272 The above would give you something like this:
2274 $stmt = "SELECT * FROM tickets WHERE
2275 ( requestor = ? ) AND ( status != ? )
2276 AND ( worker = ? OR worker = ? OR worker = ? )";
2277 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
2279 Which you could then use in DBI code like so:
2281 my $sth = $dbh->prepare($stmt);
2282 $sth->execute(@bind);
2288 The methods are simple. There's one for every major SQL operation,
2289 and a constructor you use first. The arguments are specified in a
2290 similar order for each method (table, then fields, then a where
2291 clause) to try and simplify things.
2293 =head2 new(option => 'value')
2295 The C<new()> function takes a list of options and values, and returns
2296 a new B<SQL::Abstract> object which can then be used to generate SQL
2297 through the methods below. The options accepted are:
2303 If set to 'lower', then SQL will be generated in all lowercase. By
2304 default SQL is generated in "textbook" case meaning something like:
2306 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
2308 Any setting other than 'lower' is ignored.
2312 This determines what the default comparison operator is. By default
2313 it is C<=>, meaning that a hash like this:
2315 %where = (name => 'nwiger', email => 'nate@wiger.org');
2317 Will generate SQL like this:
2319 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
2321 However, you may want loose comparisons by default, so if you set
2322 C<cmp> to C<like> you would get SQL such as:
2324 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
2326 You can also override the comparison on an individual basis - see
2327 the huge section on L</"WHERE CLAUSES"> at the bottom.
2329 =item sqltrue, sqlfalse
2331 Expressions for inserting boolean values within SQL statements.
2332 By default these are C<1=1> and C<1=0>. They are used
2333 by the special operators C<-in> and C<-not_in> for generating
2334 correct SQL even when the argument is an empty array (see below).
2338 This determines the default logical operator for multiple WHERE
2339 statements in arrays or hashes. If absent, the default logic is "or"
2340 for arrays, and "and" for hashes. This means that a WHERE
2344 event_date => {'>=', '2/13/99'},
2345 event_date => {'<=', '4/24/03'},
2348 will generate SQL like this:
2350 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
2352 This is probably not what you want given this query, though (look
2353 at the dates). To change the "OR" to an "AND", simply specify:
2355 my $sql = SQL::Abstract->new(logic => 'and');
2357 Which will change the above C<WHERE> to:
2359 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
2361 The logic can also be changed locally by inserting
2362 a modifier in front of an arrayref:
2364 @where = (-and => [event_date => {'>=', '2/13/99'},
2365 event_date => {'<=', '4/24/03'} ]);
2367 See the L</"WHERE CLAUSES"> section for explanations.
2371 This will automatically convert comparisons using the specified SQL
2372 function for both column and value. This is mostly used with an argument
2373 of C<upper> or C<lower>, so that the SQL will have the effect of
2374 case-insensitive "searches". For example, this:
2376 $sql = SQL::Abstract->new(convert => 'upper');
2377 %where = (keywords => 'MaKe iT CAse inSeNSItive');
2379 Will turn out the following SQL:
2381 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
2383 The conversion can be C<upper()>, C<lower()>, or any other SQL function
2384 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
2385 not validate this option; it will just pass through what you specify verbatim).
2389 This is a kludge because many databases suck. For example, you can't
2390 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
2391 Instead, you have to use C<bind_param()>:
2393 $sth->bind_param(1, 'reg data');
2394 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
2396 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
2397 which loses track of which field each slot refers to. Fear not.
2399 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
2400 Currently, you can specify either C<normal> (default) or C<columns>. If you
2401 specify C<columns>, you will get an array that looks like this:
2403 my $sql = SQL::Abstract->new(bindtype => 'columns');
2404 my($stmt, @bind) = $sql->insert(...);
2407 [ 'column1', 'value1' ],
2408 [ 'column2', 'value2' ],
2409 [ 'column3', 'value3' ],
2412 You can then iterate through this manually, using DBI's C<bind_param()>.
2414 $sth->prepare($stmt);
2417 my($col, $data) = @$_;
2418 if ($col eq 'details' || $col eq 'comments') {
2419 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
2420 } elsif ($col eq 'image') {
2421 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
2423 $sth->bind_param($i, $data);
2427 $sth->execute; # execute without @bind now
2429 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
2430 Basically, the advantage is still that you don't have to care which fields
2431 are or are not included. You could wrap that above C<for> loop in a simple
2432 sub called C<bind_fields()> or something and reuse it repeatedly. You still
2433 get a layer of abstraction over manual SQL specification.
2435 Note that if you set L</bindtype> to C<columns>, the C<\[ $sql, @bind ]>
2436 construct (see L</Literal SQL with placeholders and bind values (subqueries)>)
2437 will expect the bind values in this format.
2441 This is the character that a table or column name will be quoted
2442 with. By default this is an empty string, but you could set it to
2443 the character C<`>, to generate SQL like this:
2445 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
2447 Alternatively, you can supply an array ref of two items, the first being the left
2448 hand quote character, and the second the right hand quote character. For
2449 example, you could supply C<['[',']']> for SQL Server 2000 compliant quotes
2450 that generates SQL like this:
2452 SELECT [a_field] FROM [a_table] WHERE [some_field] LIKE '%someval%'
2454 Quoting is useful if you have tables or columns names that are reserved
2455 words in your database's SQL dialect.
2459 This is the character that will be used to escape L</quote_char>s appearing
2460 in an identifier before it has been quoted.
2462 The parameter default in case of a single L</quote_char> character is the quote
2465 When opening-closing-style quoting is used (L</quote_char> is an arrayref)
2466 this parameter defaults to the B<closing (right)> L</quote_char>. Occurrences
2467 of the B<opening (left)> L</quote_char> within the identifier are currently left
2468 untouched. The default for opening-closing-style quotes may change in future
2469 versions, thus you are B<strongly encouraged> to specify the escape character
2474 This is the character that separates a table and column name. It is
2475 necessary to specify this when the C<quote_char> option is selected,
2476 so that tables and column names can be individually quoted like this:
2478 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
2480 =item injection_guard
2482 A regular expression C<qr/.../> that is applied to any C<-function> and unquoted
2483 column name specified in a query structure. This is a safety mechanism to avoid
2484 injection attacks when mishandling user input e.g.:
2486 my %condition_as_column_value_pairs = get_values_from_user();
2487 $sqla->select( ... , \%condition_as_column_value_pairs );
2489 If the expression matches an exception is thrown. Note that literal SQL
2490 supplied via C<\'...'> or C<\['...']> is B<not> checked in any way.
2492 Defaults to checking for C<;> and the C<GO> keyword (TransactSQL)
2494 =item array_datatypes
2496 When this option is true, arrayrefs in INSERT or UPDATE are
2497 interpreted as array datatypes and are passed directly
2499 When this option is false, arrayrefs are interpreted
2500 as literal SQL, just like refs to arrayrefs
2501 (but this behavior is for backwards compatibility; when writing
2502 new queries, use the "reference to arrayref" syntax
2508 Takes a reference to a list of "special operators"
2509 to extend the syntax understood by L<SQL::Abstract>.
2510 See section L</"SPECIAL OPERATORS"> for details.
2514 Takes a reference to a list of "unary operators"
2515 to extend the syntax understood by L<SQL::Abstract>.
2516 See section L</"UNARY OPERATORS"> for details.
2522 =head2 insert($table, \@values || \%fieldvals, \%options)
2524 This is the simplest function. You simply give it a table name
2525 and either an arrayref of values or hashref of field/value pairs.
2526 It returns an SQL INSERT statement and a list of bind values.
2527 See the sections on L</"Inserting and Updating Arrays"> and
2528 L</"Inserting and Updating SQL"> for information on how to insert
2529 with those data types.
2531 The optional C<\%options> hash reference may contain additional
2532 options to generate the insert SQL. Currently supported options
2539 Takes either a scalar of raw SQL fields, or an array reference of
2540 field names, and adds on an SQL C<RETURNING> statement at the end.
2541 This allows you to return data generated by the insert statement
2542 (such as row IDs) without performing another C<SELECT> statement.
2543 Note, however, this is not part of the SQL standard and may not
2544 be supported by all database engines.
2548 =head2 update($table, \%fieldvals, \%where, \%options)
2550 This takes a table, hashref of field/value pairs, and an optional
2551 hashref L<WHERE clause|/WHERE CLAUSES>. It returns an SQL UPDATE function and a list
2553 See the sections on L</"Inserting and Updating Arrays"> and
2554 L</"Inserting and Updating SQL"> for information on how to insert
2555 with those data types.
2557 The optional C<\%options> hash reference may contain additional
2558 options to generate the update SQL. Currently supported options
2565 See the C<returning> option to
2566 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2570 =head2 select($source, $fields, $where, $order)
2572 This returns a SQL SELECT statement and associated list of bind values, as
2573 specified by the arguments:
2579 Specification of the 'FROM' part of the statement.
2580 The argument can be either a plain scalar (interpreted as a table
2581 name, will be quoted), or an arrayref (interpreted as a list
2582 of table names, joined by commas, quoted), or a scalarref
2583 (literal SQL, not quoted).
2587 Specification of the list of fields to retrieve from
2589 The argument can be either an arrayref (interpreted as a list
2590 of field names, will be joined by commas and quoted), or a
2591 plain scalar (literal SQL, not quoted).
2592 Please observe that this API is not as flexible as that of
2593 the first argument C<$source>, for backwards compatibility reasons.
2597 Optional argument to specify the WHERE part of the query.
2598 The argument is most often a hashref, but can also be
2599 an arrayref or plain scalar --
2600 see section L<WHERE clause|/"WHERE CLAUSES"> for details.
2604 Optional argument to specify the ORDER BY part of the query.
2605 The argument can be a scalar, a hashref or an arrayref
2606 -- see section L<ORDER BY clause|/"ORDER BY CLAUSES">
2612 =head2 delete($table, \%where, \%options)
2614 This takes a table name and optional hashref L<WHERE clause|/WHERE CLAUSES>.
2615 It returns an SQL DELETE statement and list of bind values.
2617 The optional C<\%options> hash reference may contain additional
2618 options to generate the delete SQL. Currently supported options
2625 See the C<returning> option to
2626 L<insert|/insert($table, \@values || \%fieldvals, \%options)>.
2630 =head2 where(\%where, $order)
2632 This is used to generate just the WHERE clause. For example,
2633 if you have an arbitrary data structure and know what the
2634 rest of your SQL is going to look like, but want an easy way
2635 to produce a WHERE clause, use this. It returns an SQL WHERE
2636 clause and list of bind values.
2639 =head2 values(\%data)
2641 This just returns the values from the hash C<%data>, in the same
2642 order that would be returned from any of the other above queries.
2643 Using this allows you to markedly speed up your queries if you
2644 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
2646 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
2648 Warning: This is an experimental method and subject to change.
2650 This returns arbitrarily generated SQL. It's a really basic shortcut.
2651 It will return two different things, depending on return context:
2653 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
2654 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
2656 These would return the following:
2658 # First calling form
2659 $stmt = "CREATE TABLE test (?, ?)";
2660 @bind = (field1, field2);
2662 # Second calling form
2663 $stmt_and_val = "CREATE TABLE test (field1, field2)";
2665 Depending on what you're trying to do, it's up to you to choose the correct
2666 format. In this example, the second form is what you would want.
2670 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
2674 ALTER SESSION SET nls_date_format = 'MM/YY'
2676 You get the idea. Strings get their case twiddled, but everything
2677 else remains verbatim.
2679 =head1 EXPORTABLE FUNCTIONS
2681 =head2 is_plain_value
2683 Determines if the supplied argument is a plain value as understood by this
2688 =item * The value is C<undef>
2690 =item * The value is a non-reference
2692 =item * The value is an object with stringification overloading
2694 =item * The value is of the form C<< { -value => $anything } >>
2698 On failure returns C<undef>, on success returns a B<scalar> reference
2699 to the original supplied argument.
2705 The stringification overloading detection is rather advanced: it takes
2706 into consideration not only the presence of a C<""> overload, but if that
2707 fails also checks for enabled
2708 L<autogenerated versions of C<"">|overload/Magic Autogeneration>, based
2709 on either C<0+> or C<bool>.
2711 Unfortunately testing in the field indicates that this
2712 detection B<< may tickle a latent bug in perl versions before 5.018 >>,
2713 but only when very large numbers of stringifying objects are involved.
2714 At the time of writing ( Sep 2014 ) there is no clear explanation of
2715 the direct cause, nor is there a manageably small test case that reliably
2716 reproduces the problem.
2718 If you encounter any of the following exceptions in B<random places within
2719 your application stack> - this module may be to blame:
2721 Operation "ne": no method found,
2722 left argument in overloaded package <something>,
2723 right argument in overloaded package <something>
2727 Stub found while resolving method "???" overloading """" in package <something>
2729 If you fall victim to the above - please attempt to reduce the problem
2730 to something that could be sent to the L<SQL::Abstract developers
2731 |DBIx::Class/GETTING HELP/SUPPORT>
2732 (either publicly or privately). As a workaround in the meantime you can
2733 set C<$ENV{SQLA_ISVALUE_IGNORE_AUTOGENERATED_STRINGIFICATION}> to a true
2734 value, which will most likely eliminate your problem (at the expense of
2735 not being able to properly detect exotic forms of stringification).
2737 This notice and environment variable will be removed in a future version,
2738 as soon as the underlying problem is found and a reliable workaround is
2743 =head2 is_literal_value
2745 Determines if the supplied argument is a literal value as understood by this
2750 =item * C<\$sql_string>
2752 =item * C<\[ $sql_string, @bind_values ]>
2756 On failure returns C<undef>, on success returns an B<array> reference
2757 containing the unpacked version of the supplied literal SQL and bind values.
2759 =head1 WHERE CLAUSES
2763 This module uses a variation on the idea from L<DBIx::Abstract>. It
2764 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
2765 module is that things in arrays are OR'ed, and things in hashes
2768 The easiest way to explain is to show lots of examples. After
2769 each C<%where> hash shown, it is assumed you used:
2771 my($stmt, @bind) = $sql->where(\%where);
2773 However, note that the C<%where> hash can be used directly in any
2774 of the other functions as well, as described above.
2776 =head2 Key-value pairs
2778 So, let's get started. To begin, a simple hash:
2782 status => 'completed'
2785 Is converted to SQL C<key = val> statements:
2787 $stmt = "WHERE user = ? AND status = ?";
2788 @bind = ('nwiger', 'completed');
2790 One common thing I end up doing is having a list of values that
2791 a field can be in. To do this, simply specify a list inside of
2796 status => ['assigned', 'in-progress', 'pending'];
2799 This simple code will create the following:
2801 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
2802 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
2804 A field associated to an empty arrayref will be considered a
2805 logical false and will generate 0=1.
2807 =head2 Tests for NULL values
2809 If the value part is C<undef> then this is converted to SQL <IS NULL>
2818 $stmt = "WHERE user = ? AND status IS NULL";
2821 To test if a column IS NOT NULL:
2825 status => { '!=', undef },
2828 =head2 Specific comparison operators
2830 If you want to specify a different type of operator for your comparison,
2831 you can use a hashref for a given column:
2835 status => { '!=', 'completed' }
2838 Which would generate:
2840 $stmt = "WHERE user = ? AND status != ?";
2841 @bind = ('nwiger', 'completed');
2843 To test against multiple values, just enclose the values in an arrayref:
2845 status => { '=', ['assigned', 'in-progress', 'pending'] };
2847 Which would give you:
2849 "WHERE status = ? OR status = ? OR status = ?"
2852 The hashref can also contain multiple pairs, in which case it is expanded
2853 into an C<AND> of its elements:
2857 status => { '!=', 'completed', -not_like => 'pending%' }
2860 # Or more dynamically, like from a form
2861 $where{user} = 'nwiger';
2862 $where{status}{'!='} = 'completed';
2863 $where{status}{'-not_like'} = 'pending%';
2865 # Both generate this
2866 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
2867 @bind = ('nwiger', 'completed', 'pending%');
2870 To get an OR instead, you can combine it with the arrayref idea:
2874 priority => [ { '=', 2 }, { '>', 5 } ]
2877 Which would generate:
2879 $stmt = "WHERE ( priority = ? OR priority > ? ) AND user = ?";
2880 @bind = ('2', '5', 'nwiger');
2882 If you want to include literal SQL (with or without bind values), just use a
2883 scalar reference or reference to an arrayref as the value:
2886 date_entered => { '>' => \["to_date(?, 'MM/DD/YYYY')", "11/26/2008"] },
2887 date_expires => { '<' => \"now()" }
2890 Which would generate:
2892 $stmt = "WHERE date_entered > to_date(?, 'MM/DD/YYYY') AND date_expires < now()";
2893 @bind = ('11/26/2008');
2896 =head2 Logic and nesting operators
2898 In the example above,
2899 there is a subtle trap if you want to say something like
2900 this (notice the C<AND>):
2902 WHERE priority != ? AND priority != ?
2904 Because, in Perl you I<can't> do this:
2906 priority => { '!=' => 2, '!=' => 1 }
2908 As the second C<!=> key will obliterate the first. The solution
2909 is to use the special C<-modifier> form inside an arrayref:
2911 priority => [ -and => {'!=', 2},
2915 Normally, these would be joined by C<OR>, but the modifier tells it
2916 to use C<AND> instead. (Hint: You can use this in conjunction with the
2917 C<logic> option to C<new()> in order to change the way your queries
2918 work by default.) B<Important:> Note that the C<-modifier> goes
2919 B<INSIDE> the arrayref, as an extra first element. This will
2920 B<NOT> do what you think it might:
2922 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
2924 Here is a quick list of equivalencies, since there is some overlap:
2927 status => {'!=', 'completed', 'not like', 'pending%' }
2928 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
2931 status => {'=', ['assigned', 'in-progress']}
2932 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
2933 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
2937 =head2 Special operators: IN, BETWEEN, etc.
2939 You can also use the hashref format to compare a list of fields using the
2940 C<IN> comparison operator, by specifying the list as an arrayref:
2943 status => 'completed',
2944 reportid => { -in => [567, 2335, 2] }
2947 Which would generate:
2949 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
2950 @bind = ('completed', '567', '2335', '2');
2952 The reverse operator C<-not_in> generates SQL C<NOT IN> and is used in
2955 If the argument to C<-in> is an empty array, 'sqlfalse' is generated
2956 (by default: C<1=0>). Similarly, C<< -not_in => [] >> generates
2957 'sqltrue' (by default: C<1=1>).
2959 In addition to the array you can supply a chunk of literal sql or
2960 literal sql with bind:
2963 customer => { -in => \[
2964 'SELECT cust_id FROM cust WHERE balance > ?',
2967 status => { -in => \'SELECT status_codes FROM states' },
2973 customer IN ( SELECT cust_id FROM cust WHERE balance > ? )
2974 AND status IN ( SELECT status_codes FROM states )
2978 Finally, if the argument to C<-in> is not a reference, it will be
2979 treated as a single-element array.
2981 Another pair of operators is C<-between> and C<-not_between>,
2982 used with an arrayref of two values:
2986 completion_date => {
2987 -not_between => ['2002-10-01', '2003-02-06']
2993 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
2995 Just like with C<-in> all plausible combinations of literal SQL
2999 start0 => { -between => [ 1, 2 ] },
3000 start1 => { -between => \["? AND ?", 1, 2] },
3001 start2 => { -between => \"lower(x) AND upper(y)" },
3002 start3 => { -between => [
3004 \["upper(?)", 'stuff' ],
3011 ( start0 BETWEEN ? AND ? )
3012 AND ( start1 BETWEEN ? AND ? )
3013 AND ( start2 BETWEEN lower(x) AND upper(y) )
3014 AND ( start3 BETWEEN lower(x) AND upper(?) )
3016 @bind = (1, 2, 1, 2, 'stuff');
3019 These are the two builtin "special operators"; but the
3020 list can be expanded: see section L</"SPECIAL OPERATORS"> below.
3022 =head2 Unary operators: bool
3024 If you wish to test against boolean columns or functions within your
3025 database you can use the C<-bool> and C<-not_bool> operators. For
3026 example to test the column C<is_user> being true and the column
3027 C<is_enabled> being false you would use:-
3031 -not_bool => 'is_enabled',
3036 WHERE is_user AND NOT is_enabled
3038 If a more complex combination is required, testing more conditions,
3039 then you should use the and/or operators:-
3044 -not_bool => { two=> { -rlike => 'bar' } },
3045 -not_bool => { three => [ { '=', 2 }, { '>', 5 } ] },
3056 (NOT ( three = ? OR three > ? ))
3059 =head2 Nested conditions, -and/-or prefixes
3061 So far, we've seen how multiple conditions are joined with a top-level
3062 C<AND>. We can change this by putting the different conditions we want in
3063 hashes and then putting those hashes in an array. For example:
3068 status => { -like => ['pending%', 'dispatched'] },
3072 status => 'unassigned',
3076 This data structure would create the following:
3078 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
3079 OR ( user = ? AND status = ? ) )";
3080 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
3083 Clauses in hashrefs or arrayrefs can be prefixed with an C<-and> or C<-or>
3084 to change the logic inside:
3090 -and => [ workhrs => {'>', 20}, geo => 'ASIA' ],
3091 -or => { workhrs => {'<', 50}, geo => 'EURO' },
3098 $stmt = "WHERE ( user = ?
3099 AND ( ( workhrs > ? AND geo = ? )
3100 OR ( workhrs < ? OR geo = ? ) ) )";
3101 @bind = ('nwiger', '20', 'ASIA', '50', 'EURO');
3103 =head3 Algebraic inconsistency, for historical reasons
3105 C<Important note>: when connecting several conditions, the C<-and->|C<-or>
3106 operator goes C<outside> of the nested structure; whereas when connecting
3107 several constraints on one column, the C<-and> operator goes
3108 C<inside> the arrayref. Here is an example combining both features:
3111 -and => [a => 1, b => 2],
3112 -or => [c => 3, d => 4],
3113 e => [-and => {-like => 'foo%'}, {-like => '%bar'} ]
3118 WHERE ( ( ( a = ? AND b = ? )
3119 OR ( c = ? OR d = ? )
3120 OR ( e LIKE ? AND e LIKE ? ) ) )
3122 This difference in syntax is unfortunate but must be preserved for
3123 historical reasons. So be careful: the two examples below would
3124 seem algebraically equivalent, but they are not
3127 { -like => 'foo%' },
3128 { -like => '%bar' },
3130 # yields: WHERE ( ( col LIKE ? AND col LIKE ? ) )
3133 { col => { -like => 'foo%' } },
3134 { col => { -like => '%bar' } },
3136 # yields: WHERE ( ( col LIKE ? OR col LIKE ? ) )
3139 =head2 Literal SQL and value type operators
3141 The basic premise of SQL::Abstract is that in WHERE specifications the "left
3142 side" is a column name and the "right side" is a value (normally rendered as
3143 a placeholder). This holds true for both hashrefs and arrayref pairs as you
3144 see in the L</WHERE CLAUSES> examples above. Sometimes it is necessary to
3145 alter this behavior. There are several ways of doing so.
3149 This is a virtual operator that signals the string to its right side is an
3150 identifier (a column name) and not a value. For example to compare two
3151 columns you would write:
3154 priority => { '<', 2 },
3155 requestor => { -ident => 'submitter' },
3160 $stmt = "WHERE priority < ? AND requestor = submitter";
3163 If you are maintaining legacy code you may see a different construct as
3164 described in L</Deprecated usage of Literal SQL>, please use C<-ident> in new
3169 This is a virtual operator that signals that the construct to its right side
3170 is a value to be passed to DBI. This is for example necessary when you want
3171 to write a where clause against an array (for RDBMS that support such
3172 datatypes). For example:
3175 array => { -value => [1, 2, 3] }
3180 $stmt = 'WHERE array = ?';
3181 @bind = ([1, 2, 3]);
3183 Note that if you were to simply say:
3189 the result would probably not be what you wanted:
3191 $stmt = 'WHERE array = ? OR array = ? OR array = ?';
3196 Finally, sometimes only literal SQL will do. To include a random snippet
3197 of SQL verbatim, you specify it as a scalar reference. Consider this only
3198 as a last resort. Usually there is a better way. For example:
3201 priority => { '<', 2 },
3202 requestor => { -in => \'(SELECT name FROM hitmen)' },
3207 $stmt = "WHERE priority < ? AND requestor IN (SELECT name FROM hitmen)"
3210 Note that in this example, you only get one bind parameter back, since
3211 the verbatim SQL is passed as part of the statement.
3215 Never use untrusted input as a literal SQL argument - this is a massive
3216 security risk (there is no way to check literal snippets for SQL
3217 injections and other nastyness). If you need to deal with untrusted input
3218 use literal SQL with placeholders as described next.
3220 =head3 Literal SQL with placeholders and bind values (subqueries)
3222 If the literal SQL to be inserted has placeholders and bind values,
3223 use a reference to an arrayref (yes this is a double reference --
3224 not so common, but perfectly legal Perl). For example, to find a date
3225 in Postgres you can use something like this:
3228 date_column => \[ "= date '2008-09-30' - ?::integer", 10 ]
3233 $stmt = "WHERE ( date_column = date '2008-09-30' - ?::integer )"
3236 Note that you must pass the bind values in the same format as they are returned
3237 by L<where|/where(\%where, $order)>. This means that if you set L</bindtype>
3238 to C<columns>, you must provide the bind values in the
3239 C<< [ column_meta => value ] >> format, where C<column_meta> is an opaque
3240 scalar value; most commonly the column name, but you can use any scalar value
3241 (including references and blessed references), L<SQL::Abstract> will simply
3242 pass it through intact. So if C<bindtype> is set to C<columns> the above
3243 example will look like:
3246 date_column => \[ "= date '2008-09-30' - ?::integer", [ {} => 10 ] ]
3249 Literal SQL is especially useful for nesting parenthesized clauses in the
3250 main SQL query. Here is a first example:
3252 my ($sub_stmt, @sub_bind) = ("SELECT c1 FROM t1 WHERE c2 < ? AND c3 LIKE ?",
3256 bar => \["IN ($sub_stmt)" => @sub_bind],
3261 $stmt = "WHERE (foo = ? AND bar IN (SELECT c1 FROM t1
3262 WHERE c2 < ? AND c3 LIKE ?))";
3263 @bind = (1234, 100, "foo%");
3265 Other subquery operators, like for example C<"E<gt> ALL"> or C<"NOT IN">,
3266 are expressed in the same way. Of course the C<$sub_stmt> and
3267 its associated bind values can be generated through a former call
3270 my ($sub_stmt, @sub_bind)
3271 = $sql->select("t1", "c1", {c2 => {"<" => 100},
3272 c3 => {-like => "foo%"}});
3275 bar => \["> ALL ($sub_stmt)" => @sub_bind],
3278 In the examples above, the subquery was used as an operator on a column;
3279 but the same principle also applies for a clause within the main C<%where>
3280 hash, like an EXISTS subquery:
3282 my ($sub_stmt, @sub_bind)
3283 = $sql->select("t1", "*", {c1 => 1, c2 => \"> t0.c0"});
3284 my %where = ( -and => [
3286 \["EXISTS ($sub_stmt)" => @sub_bind],
3291 $stmt = "WHERE (foo = ? AND EXISTS (SELECT * FROM t1
3292 WHERE c1 = ? AND c2 > t0.c0))";
3296 Observe that the condition on C<c2> in the subquery refers to
3297 column C<t0.c0> of the main query: this is I<not> a bind
3298 value, so we have to express it through a scalar ref.
3299 Writing C<< c2 => {">" => "t0.c0"} >> would have generated
3300 C<< c2 > ? >> with bind value C<"t0.c0"> ... not exactly
3301 what we wanted here.
3303 Finally, here is an example where a subquery is used
3304 for expressing unary negation:
3306 my ($sub_stmt, @sub_bind)
3307 = $sql->where({age => [{"<" => 10}, {">" => 20}]});
3308 $sub_stmt =~ s/^ where //i; # don't want "WHERE" in the subclause
3310 lname => {like => '%son%'},
3311 \["NOT ($sub_stmt)" => @sub_bind],
3316 $stmt = "lname LIKE ? AND NOT ( age < ? OR age > ? )"
3317 @bind = ('%son%', 10, 20)
3319 =head3 Deprecated usage of Literal SQL
3321 Below are some examples of archaic use of literal SQL. It is shown only as
3322 reference for those who deal with legacy code. Each example has a much
3323 better, cleaner and safer alternative that users should opt for in new code.
3329 my %where = ( requestor => \'IS NOT NULL' )
3331 $stmt = "WHERE requestor IS NOT NULL"
3333 This used to be the way of generating NULL comparisons, before the handling
3334 of C<undef> got formalized. For new code please use the superior syntax as
3335 described in L</Tests for NULL values>.
3339 my %where = ( requestor => \'= submitter' )
3341 $stmt = "WHERE requestor = submitter"
3343 This used to be the only way to compare columns. Use the superior L</-ident>
3344 method for all new code. For example an identifier declared in such a way
3345 will be properly quoted if L</quote_char> is properly set, while the legacy
3346 form will remain as supplied.
3350 my %where = ( is_ready => \"", completed => { '>', '2012-12-21' } )
3352 $stmt = "WHERE completed > ? AND is_ready"
3353 @bind = ('2012-12-21')
3355 Using an empty string literal used to be the only way to express a boolean.
3356 For all new code please use the much more readable
3357 L<-bool|/Unary operators: bool> operator.
3363 These pages could go on for a while, since the nesting of the data
3364 structures this module can handle are pretty much unlimited (the
3365 module implements the C<WHERE> expansion as a recursive function
3366 internally). Your best bet is to "play around" with the module a
3367 little to see how the data structures behave, and choose the best
3368 format for your data based on that.
3370 And of course, all the values above will probably be replaced with
3371 variables gotten from forms or the command line. After all, if you
3372 knew everything ahead of time, you wouldn't have to worry about
3373 dynamically-generating SQL and could just hardwire it into your
3376 =head1 ORDER BY CLAUSES
3378 Some functions take an order by clause. This can either be a scalar (just a
3379 column name), a hashref of C<< { -desc => 'col' } >> or C<< { -asc => 'col' }
3380 >>, a scalarref, an arrayref-ref, or an arrayref of any of the previous
3383 Given | Will Generate
3384 ---------------------------------------------------------------
3386 'colA' | ORDER BY colA
3388 [qw/colA colB/] | ORDER BY colA, colB
3390 {-asc => 'colA'} | ORDER BY colA ASC
3392 {-desc => 'colB'} | ORDER BY colB DESC
3394 ['colA', {-asc => 'colB'}] | ORDER BY colA, colB ASC
3396 { -asc => [qw/colA colB/] } | ORDER BY colA ASC, colB ASC
3398 \'colA DESC' | ORDER BY colA DESC
3400 \[ 'FUNC(colA, ?)', $x ] | ORDER BY FUNC(colA, ?)
3401 | /* ...with $x bound to ? */
3404 { -asc => 'colA' }, | colA ASC,
3405 { -desc => [qw/colB/] }, | colB DESC,
3406 { -asc => [qw/colC colD/] },| colC ASC, colD ASC,
3407 \'colE DESC', | colE DESC,
3408 \[ 'FUNC(colF, ?)', $x ], | FUNC(colF, ?)
3409 ] | /* ...with $x bound to ? */
3410 ===============================================================
3414 =head1 SPECIAL OPERATORS
3416 my $sqlmaker = SQL::Abstract->new(special_ops => [
3420 my ($self, $field, $op, $arg) = @_;
3426 handler => 'method_name',
3430 A "special operator" is a SQL syntactic clause that can be
3431 applied to a field, instead of a usual binary operator.
3434 WHERE field IN (?, ?, ?)
3435 WHERE field BETWEEN ? AND ?
3436 WHERE MATCH(field) AGAINST (?, ?)
3438 Special operators IN and BETWEEN are fairly standard and therefore
3439 are builtin within C<SQL::Abstract> (as the overridable methods
3440 C<_where_field_IN> and C<_where_field_BETWEEN>). For other operators,
3441 like the MATCH .. AGAINST example above which is specific to MySQL,
3442 you can write your own operator handlers - supply a C<special_ops>
3443 argument to the C<new> method. That argument takes an arrayref of
3444 operator definitions; each operator definition is a hashref with two
3451 the regular expression to match the operator
3455 Either a coderef or a plain scalar method name. In both cases
3456 the expected return is C<< ($sql, @bind) >>.
3458 When supplied with a method name, it is simply called on the
3459 L<SQL::Abstract> object as:
3461 $self->$method_name($field, $op, $arg)
3465 $field is the LHS of the operator
3466 $op is the part that matched the handler regex
3469 When supplied with a coderef, it is called as:
3471 $coderef->($self, $field, $op, $arg)
3476 For example, here is an implementation
3477 of the MATCH .. AGAINST syntax for MySQL
3479 my $sqlmaker = SQL::Abstract->new(special_ops => [
3481 # special op for MySql MATCH (field) AGAINST(word1, word2, ...)
3482 {regex => qr/^match$/i,
3484 my ($self, $field, $op, $arg) = @_;
3485 $arg = [$arg] if not ref $arg;
3486 my $label = $self->_quote($field);
3487 my ($placeholder) = $self->_convert('?');
3488 my $placeholders = join ", ", (($placeholder) x @$arg);
3489 my $sql = $self->_sqlcase('match') . " ($label) "
3490 . $self->_sqlcase('against') . " ($placeholders) ";
3491 my @bind = $self->_bindtype($field, @$arg);
3492 return ($sql, @bind);
3499 =head1 UNARY OPERATORS
3501 my $sqlmaker = SQL::Abstract->new(unary_ops => [
3505 my ($self, $op, $arg) = @_;
3511 handler => 'method_name',
3515 A "unary operator" is a SQL syntactic clause that can be
3516 applied to a field - the operator goes before the field
3518 You can write your own operator handlers - supply a C<unary_ops>
3519 argument to the C<new> method. That argument takes an arrayref of
3520 operator definitions; each operator definition is a hashref with two
3527 the regular expression to match the operator
3531 Either a coderef or a plain scalar method name. In both cases
3532 the expected return is C<< $sql >>.
3534 When supplied with a method name, it is simply called on the
3535 L<SQL::Abstract> object as:
3537 $self->$method_name($op, $arg)
3541 $op is the part that matched the handler regex
3542 $arg is the RHS or argument of the operator
3544 When supplied with a coderef, it is called as:
3546 $coderef->($self, $op, $arg)
3554 Thanks to some benchmarking by Mark Stosberg, it turns out that
3555 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
3556 I must admit this wasn't an intentional design issue, but it's a
3557 byproduct of the fact that you get to control your C<DBI> handles
3560 To maximize performance, use a code snippet like the following:
3562 # prepare a statement handle using the first row
3563 # and then reuse it for the rest of the rows
3565 for my $href (@array_of_hashrefs) {
3566 $stmt ||= $sql->insert('table', $href);
3567 $sth ||= $dbh->prepare($stmt);
3568 $sth->execute($sql->values($href));
3571 The reason this works is because the keys in your C<$href> are sorted
3572 internally by B<SQL::Abstract>. Thus, as long as your data retains
3573 the same structure, you only have to generate the SQL the first time
3574 around. On subsequent queries, simply use the C<values> function provided
3575 by this module to return your values in the correct order.
3577 However this depends on the values having the same type - if, for
3578 example, the values of a where clause may either have values
3579 (resulting in sql of the form C<column = ?> with a single bind
3580 value), or alternatively the values might be C<undef> (resulting in
3581 sql of the form C<column IS NULL> with no bind value) then the
3582 caching technique suggested will not work.
3586 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
3587 really like this part (I do, at least). Building up a complex query
3588 can be as simple as the following:
3595 use CGI::FormBuilder;
3598 my $form = CGI::FormBuilder->new(...);
3599 my $sql = SQL::Abstract->new;
3601 if ($form->submitted) {
3602 my $field = $form->field;
3603 my $id = delete $field->{id};
3604 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
3607 Of course, you would still have to connect using C<DBI> to run the
3608 query, but the point is that if you make your form look like your
3609 table, the actual query script can be extremely simplistic.
3611 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
3612 a fast interface to returning and formatting data. I frequently
3613 use these three modules together to write complex database query
3614 apps in under 50 lines.
3616 =head1 HOW TO CONTRIBUTE
3618 Contributions are always welcome, in all usable forms (we especially
3619 welcome documentation improvements). The delivery methods include git-
3620 or unified-diff formatted patches, GitHub pull requests, or plain bug
3621 reports either via RT or the Mailing list. Contributors are generally
3622 granted full access to the official repository after their first several
3623 patches pass successful review.
3625 This project is maintained in a git repository. The code and related tools are
3626 accessible at the following locations:
3630 =item * Official repo: L<git://git.shadowcat.co.uk/dbsrgits/SQL-Abstract.git>
3632 =item * Official gitweb: L<http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=dbsrgits/SQL-Abstract.git>
3634 =item * GitHub mirror: L<https://github.com/dbsrgits/sql-abstract>
3636 =item * Authorized committers: L<ssh://dbsrgits@git.shadowcat.co.uk/SQL-Abstract.git>
3642 Version 1.50 was a major internal refactoring of C<SQL::Abstract>.
3643 Great care has been taken to preserve the I<published> behavior
3644 documented in previous versions in the 1.* family; however,
3645 some features that were previously undocumented, or behaved
3646 differently from the documentation, had to be changed in order
3647 to clarify the semantics. Hence, client code that was relying
3648 on some dark areas of C<SQL::Abstract> v1.*
3649 B<might behave differently> in v1.50.
3651 The main changes are:
3657 support for literal SQL through the C<< \ [ $sql, @bind ] >> syntax.
3661 support for the { operator => \"..." } construct (to embed literal SQL)
3665 support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values)
3669 optional support for L<array datatypes|/"Inserting and Updating Arrays">
3673 defensive programming: check arguments
3677 fixed bug with global logic, which was previously implemented
3678 through global variables yielding side-effects. Prior versions would
3679 interpret C<< [ {cond1, cond2}, [cond3, cond4] ] >>
3680 as C<< "(cond1 AND cond2) OR (cond3 AND cond4)" >>.
3681 Now this is interpreted
3682 as C<< "(cond1 AND cond2) OR (cond3 OR cond4)" >>.
3687 fixed semantics of _bindtype on array args
3691 dropped the C<_anoncopy> of the %where tree. No longer necessary,
3692 we just avoid shifting arrays within that tree.
3696 dropped the C<_modlogic> function
3700 =head1 ACKNOWLEDGEMENTS
3702 There are a number of individuals that have really helped out with
3703 this module. Unfortunately, most of them submitted bugs via CPAN
3704 so I have no idea who they are! But the people I do know are:
3706 Ash Berlin (order_by hash term support)
3707 Matt Trout (DBIx::Class support)
3708 Mark Stosberg (benchmarking)
3709 Chas Owens (initial "IN" operator support)
3710 Philip Collins (per-field SQL functions)
3711 Eric Kolve (hashref "AND" support)
3712 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
3713 Dan Kubb (support for "quote_char" and "name_sep")
3714 Guillermo Roditi (patch to cleanup "IN" and "BETWEEN", fix and tests for _order_by)
3715 Laurent Dami (internal refactoring, extensible list of special operators, literal SQL)
3716 Norbert Buchmuller (support for literal SQL in hashpair, misc. fixes & tests)
3717 Peter Rabbitson (rewrite of SQLA::Test, misc. fixes & tests)
3718 Oliver Charles (support for "RETURNING" after "INSERT")
3724 L<DBIx::Class>, L<DBIx::Abstract>, L<CGI::FormBuilder>, L<HTML::QuickTable>.
3728 Copyright (c) 2001-2007 Nathan Wiger <nwiger@cpan.org>. All Rights Reserved.
3730 This module is actively maintained by Matt Trout <mst@shadowcatsystems.co.uk>
3732 For support, your best bet is to try the C<DBIx::Class> users mailing list.
3733 While not an official support venue, C<DBIx::Class> makes heavy use of
3734 C<SQL::Abstract>, and as such list members there are very familiar with
3735 how to create queries.
3739 This module is free software; you may copy this under the same
3740 terms as perl itself (either the GNU General Public License or
3741 the Artistic License)