1 package SQL::Abstract; # see doc at end of file
3 # LDNOTE : this code is heavy refactoring from original SQLA.
4 # Several design decisions will need discussion during
5 # the test / diffusion / acceptance phase; those are marked with flag
6 # 'LDNOTE' (note by laurent.dami AT free.fr)
10 use warnings FATAL => 'all';
13 use Data::Query::Constants qw(
14 DQ_IDENTIFIER DQ_OPERATOR DQ_VALUE DQ_LITERAL DQ_JOIN DQ_SELECT DQ_ORDER
16 use Data::Query::ExprHelpers qw(perl_scalar_value);
18 #======================================================================
20 #======================================================================
22 our $VERSION = '1.72';
24 # This would confuse some packagers
25 $VERSION = eval $VERSION if $VERSION =~ /_/; # numify for warning-free dev releases
29 # special operators (-in, -between). May be extended/overridden by user.
30 # See section WHERE: BUILTIN SPECIAL OPERATORS below for implementation
31 my @BUILTIN_SPECIAL_OPS = ();
33 # unaryish operators - key maps to handler
34 my @BUILTIN_UNARY_OPS = (
35 # the digits are backcompat stuff
36 { regex => qr/^ and (?: [_\s]? \d+ )? $/xi, handler => '_where_op_ANDOR' },
37 { regex => qr/^ or (?: [_\s]? \d+ )? $/xi, handler => '_where_op_ANDOR' },
38 { regex => qr/^ nest (?: [_\s]? \d+ )? $/xi, handler => '_where_op_NEST' },
39 { regex => qr/^ (?: not \s )? bool $/xi, handler => '_where_op_BOOL' },
40 { regex => qr/^ ident $/xi, handler => '_where_op_IDENT' },
41 { regex => qr/^ value $/ix, handler => '_where_op_VALUE' },
44 #======================================================================
45 # DEBUGGING AND ERROR REPORTING
46 #======================================================================
49 return unless $_[0]->{debug}; shift; # a little faster
50 my $func = (caller(1))[3];
51 warn "[$func] ", @_, "\n";
55 my($func) = (caller(1))[3];
56 Carp::carp "[$func] Warning: ", @_;
60 my($func) = (caller(1))[3];
61 Carp::croak "[$func] Fatal: ", @_;
65 #======================================================================
67 #======================================================================
71 my $class = ref($self) || $self;
72 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
74 # choose our case by keeping an option around
75 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
77 # default logic for interpreting arrayrefs
78 $opt{logic} = $opt{logic} ? uc $opt{logic} : 'OR';
80 # how to return bind vars
81 # LDNOTE: changed nwiger code : why this 'delete' ??
82 # $opt{bindtype} ||= delete($opt{bind_type}) || 'normal';
83 $opt{bindtype} ||= 'normal';
85 # default comparison is "=", but can be overridden
88 # try to recognize which are the 'equality' and 'unequality' ops
89 # (temporary quickfix, should go through a more seasoned API)
90 $opt{equality_op} = qr/^(\Q$opt{cmp}\E|is|(is\s+)?like)$/i;
91 $opt{inequality_op} = qr/^(!=|<>|(is\s+)?not(\s+like)?)$/i;
94 $opt{sqltrue} ||= '1=1';
95 $opt{sqlfalse} ||= '0=1';
98 $opt{special_ops} ||= [];
99 # regexes are applied in order, thus push after user-defines
100 push @{$opt{special_ops}}, @BUILTIN_SPECIAL_OPS;
103 $opt{unary_ops} ||= [];
104 push @{$opt{unary_ops}}, @BUILTIN_UNARY_OPS;
106 # rudimentary saniy-check for user supplied bits treated as functions/operators
107 # If a purported function matches this regular expression, an exception is thrown.
108 # Literal SQL is *NOT* subject to this check, only functions (and column names
109 # when quoting is not in effect)
112 # need to guard against ()'s in column names too, but this will break tons of
113 # hacks... ideas anyone?
114 $opt{injection_guard} ||= qr/
120 $opt{name_sep} ||= '.';
122 $opt{renderer} ||= do {
123 require Data::Query::Renderer::SQL::Naive;
124 my ($always, $chars);
125 for ($opt{quote_char}) {
126 $chars = defined() ? (ref() ? $_ : [$_]) : ['',''];
129 Data::Query::Renderer::SQL::Naive->new({
130 quote_chars => $chars, always_quote => $always,
134 return bless \%opt, $class;
138 my ($self, $dq) = @_;
139 my ($sql, @bind) = @{$self->{renderer}->render($dq)};
141 ($self->{bindtype} eq 'normal'
142 ? ($sql, map $_->{value}, @bind)
143 : ($sql, map [ $_->{value_meta}, $_->{value} ], @bind)
149 my ($self, $literal) = @_;
151 ($literal, @bind) = @$literal if ref($literal) eq 'ARRAY';
156 (@bind ? (values => [ $self->_bind_to_dq(@bind) ]) : ()),
161 my ($self, @bind) = @_;
163 $self->{bindtype} eq 'normal'
164 ? map perl_scalar_value($_), @bind
166 $self->_assert_bindval_matches_bindtype(@bind);
167 map perl_scalar_value(reverse @$_), @bind
172 my ($self, $value) = @_;
173 perl_scalar_value($value, our $Cur_Col_Meta);
177 my ($self, $ident) = @_;
178 $self->_assert_pass_injection_guard($ident)
179 unless $self->{renderer}{always_quote};
181 type => DQ_IDENTIFIER,
182 elements => [ split /\Q$self->{name_sep}/, $ident ],
187 my ($self, $op, @args) = @_;
188 $self->_assert_pass_injection_guard($op);
191 operator => { 'SQL.Naive' => $op },
196 sub _assert_pass_injection_guard {
197 if ($_[1] =~ $_[0]->{injection_guard}) {
198 my $class = ref $_[0];
199 puke "Possible SQL injection attempt '$_[1]'. If this is indeed a part of the "
200 . "desired SQL use literal SQL ( \'...' or \[ '...' ] ) or supply your own "
201 . "{injection_guard} attribute to ${class}->new()"
206 #======================================================================
208 #======================================================================
212 my $table = $self->_table(shift);
213 my $data = shift || return;
216 my $method = $self->_METHOD_FOR_refkind("_insert", $data);
217 my ($sql, @bind) = $self->$method($data);
218 $sql = join " ", $self->_sqlcase('insert into'), $table, $sql;
220 if ($options->{returning}) {
221 my ($s, @b) = $self->_insert_returning ($options);
226 return wantarray ? ($sql, @bind) : $sql;
229 sub _insert_returning {
230 my ($self, $options) = @_;
232 my $f = $options->{returning};
234 my $fieldlist = $self->_SWITCH_refkind($f, {
235 ARRAYREF => sub {join ', ', map { $self->_quote($_) } @$f;},
236 SCALAR => sub {$self->_quote($f)},
237 SCALARREF => sub {$$f},
239 return $self->_sqlcase(' returning ') . $fieldlist;
242 sub _insert_HASHREF { # explicit list of fields and then values
243 my ($self, $data) = @_;
245 my @fields = sort keys %$data;
247 my ($sql, @bind) = $self->_insert_values($data);
250 $_ = $self->_quote($_) foreach @fields;
251 $sql = "( ".join(", ", @fields).") ".$sql;
253 return ($sql, @bind);
256 sub _insert_ARRAYREF { # just generate values(?,?) part (no list of fields)
257 my ($self, $data) = @_;
259 # no names (arrayref) so can't generate bindtype
260 $self->{bindtype} ne 'columns'
261 or belch "can't do 'columns' bindtype when called with arrayref";
263 # fold the list of values into a hash of column name - value pairs
264 # (where the column names are artificially generated, and their
265 # lexicographical ordering keep the ordering of the original list)
266 my $i = "a"; # incremented values will be in lexicographical order
267 my $data_in_hash = { map { ($i++ => $_) } @$data };
269 return $self->_insert_values($data_in_hash);
272 sub _insert_ARRAYREFREF { # literal SQL with bind
273 my ($self, $data) = @_;
275 my ($sql, @bind) = @${$data};
276 $self->_assert_bindval_matches_bindtype(@bind);
278 return ($sql, @bind);
282 sub _insert_SCALARREF { # literal SQL without bind
283 my ($self, $data) = @_;
289 my ($self, $data) = @_;
291 my (@values, @all_bind);
292 foreach my $column (sort keys %$data) {
293 my $v = $data->{$column};
295 $self->_SWITCH_refkind($v, {
298 if ($self->{array_datatypes}) { # if array datatype are activated
300 push @all_bind, $self->_bindtype($column, $v);
302 else { # else literal SQL with bind
303 my ($sql, @bind) = @$v;
304 $self->_assert_bindval_matches_bindtype(@bind);
306 push @all_bind, @bind;
310 ARRAYREFREF => sub { # literal SQL with bind
311 my ($sql, @bind) = @${$v};
312 $self->_assert_bindval_matches_bindtype(@bind);
314 push @all_bind, @bind;
317 # THINK : anything useful to do with a HASHREF ?
318 HASHREF => sub { # (nothing, but old SQLA passed it through)
319 #TODO in SQLA >= 2.0 it will die instead
320 belch "HASH ref as bind value in insert is not supported";
322 push @all_bind, $self->_bindtype($column, $v);
325 SCALARREF => sub { # literal SQL without bind
329 SCALAR_or_UNDEF => sub {
331 push @all_bind, $self->_bindtype($column, $v);
338 my $sql = $self->_sqlcase('values')." ( ".join(", ", @values)." )";
339 return ($sql, @all_bind);
344 #======================================================================
346 #======================================================================
351 my $table = $self->_table(shift);
352 my $data = shift || return;
355 # first build the 'SET' part of the sql statement
356 my (@set, @all_bind);
357 puke "Unsupported data type specified to \$sql->update"
358 unless ref $data eq 'HASH';
360 for my $k (sort keys %$data) {
363 my $label = $self->_quote($k);
365 $self->_SWITCH_refkind($v, {
367 if ($self->{array_datatypes}) { # array datatype
368 push @set, "$label = ?";
369 push @all_bind, $self->_bindtype($k, $v);
371 else { # literal SQL with bind
372 my ($sql, @bind) = @$v;
373 $self->_assert_bindval_matches_bindtype(@bind);
374 push @set, "$label = $sql";
375 push @all_bind, @bind;
378 ARRAYREFREF => sub { # literal SQL with bind
379 my ($sql, @bind) = @${$v};
380 $self->_assert_bindval_matches_bindtype(@bind);
381 push @set, "$label = $sql";
382 push @all_bind, @bind;
384 SCALARREF => sub { # literal SQL without bind
385 push @set, "$label = $$v";
388 my ($op, $arg, @rest) = %$v;
390 puke 'Operator calls in update must be in the form { -op => $arg }'
391 if (@rest or not $op =~ /^\-(.+)/);
393 local $self->{_nested_func_lhs} = $k;
394 my ($sql, @bind) = $self->_where_unary_op ($1, $arg);
396 push @set, "$label = $sql";
397 push @all_bind, @bind;
399 SCALAR_or_UNDEF => sub {
400 push @set, "$label = ?";
401 push @all_bind, $self->_bindtype($k, $v);
407 my $sql = $self->_sqlcase('update') . " $table " . $self->_sqlcase('set ')
411 my($where_sql, @where_bind) = $self->where($where);
413 push @all_bind, @where_bind;
416 return wantarray ? ($sql, @all_bind) : $sql;
422 #======================================================================
424 #======================================================================
430 my $fields = shift || '*';
434 my($where_sql, @bind) = $self->where($where, $order);
436 my $sql = $self->_render_dq({
439 map $self->_ident_to_dq($_),
440 ref($fields) eq 'ARRAY' ? @$fields : $fields
442 from => $self->_table_to_dq($table),
447 return wantarray ? ($sql, @bind) : $sql;
450 #======================================================================
452 #======================================================================
457 my $table = $self->_table(shift);
461 my($where_sql, @bind) = $self->where($where);
462 my $sql = $self->_sqlcase('delete from') . " $table" . $where_sql;
464 return wantarray ? ($sql, @bind) : $sql;
468 #======================================================================
470 #======================================================================
474 # Finally, a separate routine just to handle WHERE clauses
476 my ($self, $where, $order) = @_;
482 ($sql, @bind) = $self->_recurse_where($where) if defined($where);
483 $sql = $sql ? $self->_sqlcase(' where ') . "( $sql )" : '';
487 $sql .= $self->_order_by($order);
490 return wantarray ? ($sql, @bind) : $sql;
495 my ($self, $where, $logic) = @_;
497 return $self->_render_dq($self->_where_to_dq($where, $logic));
501 my ($self, $where, $logic) = @_;
503 if (ref($where) eq 'ARRAY') {
504 return $self->_where_to_dq_ARRAYREF($where, $logic);
505 } elsif (ref($where) eq 'HASH') {
506 return $self->_where_to_dq_HASHREF($where, $logic);
508 ref($where) eq 'SCALAR'
509 or (ref($where) eq 'REF' and ref($$where) eq 'ARRAY')
511 return $self->_literal_to_dq($$where);
512 } elsif (!ref($where) or Scalar::Util::blessed($where)) {
513 return $self->_value_to_dq($where);
515 die "Can't handle $where";
518 sub _where_to_dq_ARRAYREF {
519 my ($self, $where, $logic) = @_;
521 $logic = uc($logic || $self->{logic} || 'OR');
522 $logic eq 'AND' or $logic eq 'OR' or puke "unknown logic: $logic";
524 return unless @$where;
526 my ($first, @rest) = @$where;
528 return $self->_where_to_dq($first) unless @rest;
532 $self->_where_hashpair_to_dq($first => shift(@rest));
534 $self->_where_to_dq($first);
538 return $self->_where_to_dq_ARRAYREF(\@rest, $logic) unless $first_dq;
541 $logic, $first_dq, $self->_where_to_dq_ARRAYREF(\@rest, $logic)
545 sub _where_to_dq_HASHREF {
546 my ($self, $where, $logic) = @_;
548 $logic = uc($logic) if $logic;
551 $self->_where_hashpair_to_dq($_ => $where->{$_}, $logic)
554 return $dq[0] unless @dq > 1;
556 my $final = pop(@dq);
558 foreach my $dq (reverse @dq) {
559 $final = $self->_op_to_dq($logic||'AND', $dq, $final);
565 sub _where_to_dq_SCALAR {
566 shift->_value_to_dq(@_);
569 sub _where_op_IDENT {
571 my ($op, $rhs) = splice @_, -2;
573 puke "-$op takes a single scalar argument (a quotable identifier)";
576 # in case we are called as a top level special op (no '=')
579 $_ = $self->_convert($self->_quote($_)) for ($lhs, $rhs);
587 sub _where_op_VALUE {
589 my ($op, $rhs) = splice @_, -2;
591 # in case we are called as a top level special op (no '=')
596 ($lhs || $self->{_nested_func_lhs}),
603 $self->_convert($self->_quote($lhs)) . ' = ' . $self->_convert('?'),
607 $self->_convert('?'),
613 sub _where_hashpair_to_dq {
614 my ($self, $k, $v, $logic) = @_;
616 if ($k =~ /^-(.*)/s) {
618 if ($op eq 'AND' or $op eq 'OR') {
619 return $self->_where_to_dq($v, $op);
620 } elsif ($op eq 'NEST') {
621 return $self->_where_to_dq($v);
622 } elsif ($op eq 'NOT') {
623 return $self->_op_to_dq(NOT => $self->_where_to_dq($v));
624 } elsif ($op eq 'BOOL') {
625 return ref($v) ? $self->_where_to_dq($v) : $self->_ident_to_dq($v);
626 } elsif ($op eq 'NOT_BOOL') {
627 return $self->_op_to_dq(
628 NOT => ref($v) ? $self->_where_to_dq($v) : $self->_ident_to_dq($v)
630 } elsif ($op =~ /^(?:AND|OR|NEST)_?\d+/) {
631 die "Use of [and|or|nest]_N modifiers is no longer supported";
634 if (ref($v) eq 'HASH' and keys(%$v) == 1 and (keys %$v)[0] =~ /^-(.*)/s) {
636 my ($inner) = values %$v;
639 (map $self->_where_to_dq($_),
640 (ref($inner) eq 'ARRAY' ? @$inner : $inner))
643 (map $self->_where_to_dq($_), (ref($v) eq 'ARRAY' ? @$v : $v))
646 $self->_assert_pass_injection_guard($op);
647 return $self->_op_to_dq(
648 apply => $self->_ident_to_dq($op), @args
652 local our $Cur_Col_Meta = $k;
653 if (ref($v) eq 'ARRAY') {
655 return $self->_literal_to_dq($self->{sqlfalse});
656 } elsif (defined($v->[0]) && $v->[0] =~ /-(and|or)/i) {
657 return $self->_where_to_dq_ARRAYREF([
658 map +{ $k => $_ }, @{$v}[1..$#$v]
661 return $self->_where_to_dq_ARRAYREF([
662 map +{ $k => $_ }, @$v
664 } elsif (ref($v) eq 'SCALAR' or (ref($v) eq 'REF' and ref($$v) eq 'ARRAY')) {
668 parts => [ $self->_ident_to_dq($k), $self->_literal_to_dq($$v) ]
671 my ($op, $rhs) = do {
672 if (ref($v) eq 'HASH') {
674 return $self->_where_to_dq_ARRAYREF([
675 map +{ $k => { $_ => $v->{$_} } }, sort keys %$v
678 my ($op, $value) = %$v;
679 s/^-//, s/_/ /g for $op;
680 if ($op =~ /^(and|or)$/i) {
681 return $self->_where_to_dq({ $k => $value }, $op);
683 my $special_op = List::Util::first {$op =~ $_->{regex}}
684 @{$self->{special_ops}}
686 return $self->_literal_to_dq(
687 [ $self->${\$special_op->{handler}}($k, $op, $value) ]
689 } elsif ($op =~ /^(?:AND|OR|NEST)_?\d+$/i) {
690 die "Use of [and|or|nest]_N modifiers is no longer supported";
697 if ($op eq 'BETWEEN' or $op eq 'IN' or $op eq 'NOT IN' or $op eq 'NOT BETWEEN') {
698 if (ref($rhs) ne 'ARRAY') {
700 # have to add parens if none present because -in => \"SELECT ..."
701 # got documented. mst hates everything.
702 if (ref($rhs) eq 'SCALAR') {
704 1 while ($x =~ s/\A\s*\((.*)\)\s*\Z/$1/s);
707 my ($x, @rest) = @{$$rhs};
708 1 while ($x =~ s/\A\s*\((.*)\)\s*\Z/$1/s);
709 $rhs = \[ $x, @rest ];
712 return $self->_op_to_dq(
713 $op, $self->_ident_to_dq($k), $self->_literal_to_dq($$rhs)
716 return $self->_literal_to_dq($self->{sqlfalse}) unless @$rhs;
717 return $self->_op_to_dq(
718 $op, $self->_ident_to_dq($k), map $self->_where_to_dq($_), @$rhs
720 } elsif ($op =~ s/^NOT (?!LIKE)//) {
721 return $self->_where_hashpair_to_dq(-not => { $k => { $op => $rhs } });
722 } elsif (!defined($rhs)) {
724 if ($op eq '=' or $op eq 'LIKE') {
726 } elsif ($op eq '!=') {
729 die "Can't do undef -> NULL transform for operator ${op}";
732 return $self->_op_to_dq($null_op, $self->_ident_to_dq($k));
734 if (ref($rhs) eq 'ARRAY') {
736 return $self->_literal_to_dq(
737 $op eq '!=' ? $self->{sqltrue} : $self->{sqlfalse}
739 } elsif (defined($rhs->[0]) and $rhs->[0] =~ /^-(and|or)$/i) {
740 return $self->_where_to_dq_ARRAYREF([
741 map +{ $k => { $op => $_ } }, @{$rhs}[1..$#$rhs]
743 } elsif ($op =~ /^-(?:AND|OR|NEST)_?\d+/) {
744 die "Use of [and|or|nest]_N modifiers is no longer supported";
746 return $self->_where_to_dq_ARRAYREF([
747 map +{ $k => { $op => $_ } }, @$rhs
750 return $self->_op_to_dq(
751 $op, $self->_ident_to_dq($k), $self->_where_to_dq($rhs)
756 #======================================================================
758 #======================================================================
761 my ($self, $arg) = @_;
762 if (my $dq = $self->_order_by_to_dq($arg)) {
763 # SQLA generates ' ORDER BY foo'. The hilarity.
765 ? do { my @r = $self->_render_dq($dq); $r[0] = ' '.$r[0]; @r }
766 : ' '.$self->_render_dq($dq);
772 sub _order_by_to_dq {
773 my ($self, $arg, $dir) = @_;
779 ($dir ? (direction => $dir) : ()),
783 $dq->{by} = $self->_ident_to_dq($arg);
784 } elsif (ref($arg) eq 'ARRAY') {
786 local our $Order_Inner unless our $Order_Recursing;
787 local $Order_Recursing = 1;
789 foreach my $member (@$arg) {
791 my $next = $self->_order_by_to_dq($member, $dir);
793 $inner->{from} = $next if $inner;
794 $inner = $Order_Inner || $next;
796 $Order_Inner = $inner;
798 } elsif (ref($arg) eq 'REF' and ref($$arg) eq 'ARRAY') {
799 $dq->{by} = $self->_literal_to_dq($$arg);
800 } elsif (ref($arg) eq 'SCALAR') {
801 $dq->{by} = $self->_literal_to_dq($$arg);
802 } elsif (ref($arg) eq 'HASH') {
803 my ($key, $val, @rest) = %$arg;
807 if (@rest or not $key =~ /^-(desc|asc)/i) {
808 puke "hash passed to _order_by must have exactly one key (-desc or -asc)";
811 return $self->_order_by_to_dq($val, $dir);
813 die "Can't handle $arg in _order_by_to_dq";
818 #======================================================================
819 # DATASOURCE (FOR NOW, JUST PLAIN TABLE OR LIST OF TABLES)
820 #======================================================================
823 my ($self, $from) = @_;
824 $self->_render_dq($self->_table_to_dq($from));
828 my ($self, $from) = @_;
829 $self->_SWITCH_refkind($from, {
831 die "Empty FROM list" unless my @f = @$from;
832 my $dq = $self->_ident_to_dq(shift @f);
833 while (my $x = shift @f) {
836 join => [ $dq, $self->_ident_to_dq($x) ]
841 SCALAR => sub { $self->_ident_to_dq($from) },
853 #======================================================================
855 #======================================================================
857 # highly optimized, as it's called way too often
859 # my ($self, $label) = @_;
861 return '' unless defined $_[1];
862 return ${$_[1]} if ref($_[1]) eq 'SCALAR';
864 unless ($_[0]->{quote_char}) {
865 $_[0]->_assert_pass_injection_guard($_[1]);
869 my $qref = ref $_[0]->{quote_char};
872 ($l, $r) = ( $_[0]->{quote_char}, $_[0]->{quote_char} );
874 elsif ($qref eq 'ARRAY') {
875 ($l, $r) = @{$_[0]->{quote_char}};
878 puke "Unsupported quote_char format: $_[0]->{quote_char}";
881 # parts containing * are naturally unquoted
882 return join( $_[0]->{name_sep}||'', map
883 { $_ eq '*' ? $_ : $l . $_ . $r }
884 ( $_[0]->{name_sep} ? split (/\Q$_[0]->{name_sep}\E/, $_[1] ) : $_[1] )
889 # Conversion, if applicable
891 #my ($self, $arg) = @_;
893 # LDNOTE : modified the previous implementation below because
894 # it was not consistent : the first "return" is always an array,
895 # the second "return" is context-dependent. Anyway, _convert
896 # seems always used with just a single argument, so make it a
898 # return @_ unless $self->{convert};
899 # my $conv = $self->_sqlcase($self->{convert});
900 # my @ret = map { $conv.'('.$_.')' } @_;
901 # return wantarray ? @ret : $ret[0];
902 if ($_[0]->{convert}) {
903 return $_[0]->_sqlcase($_[0]->{convert}) .'(' . $_[1] . ')';
910 #my ($self, $col, @vals) = @_;
912 #LDNOTE : changed original implementation below because it did not make
913 # sense when bindtype eq 'columns' and @vals > 1.
914 # return $self->{bindtype} eq 'columns' ? [ $col, @vals ] : @vals;
916 # called often - tighten code
917 return $_[0]->{bindtype} eq 'columns'
918 ? map {[$_[1], $_]} @_[2 .. $#_]
923 # Dies if any element of @bind is not in [colname => value] format
924 # if bindtype is 'columns'.
925 sub _assert_bindval_matches_bindtype {
926 # my ($self, @bind) = @_;
928 if ($self->{bindtype} eq 'columns') {
930 if (!defined $_ || ref($_) ne 'ARRAY' || @$_ != 2) {
931 puke "bindtype 'columns' selected, you need to pass: [column_name => bind_value]"
937 sub _join_sql_clauses {
938 my ($self, $logic, $clauses_aref, $bind_aref) = @_;
940 if (@$clauses_aref > 1) {
941 my $join = " " . $self->_sqlcase($logic) . " ";
942 my $sql = '( ' . join($join, @$clauses_aref) . ' )';
943 return ($sql, @$bind_aref);
945 elsif (@$clauses_aref) {
946 return ($clauses_aref->[0], @$bind_aref); # no parentheses
949 return (); # if no SQL, ignore @$bind_aref
954 # Fix SQL case, if so requested
956 # LDNOTE: if $self->{case} is true, then it contains 'lower', so we
957 # don't touch the argument ... crooked logic, but let's not change it!
958 return $_[0]->{case} ? $_[1] : uc($_[1]);
962 #======================================================================
963 # DISPATCHING FROM REFKIND
964 #======================================================================
967 my ($self, $data) = @_;
969 return 'UNDEF' unless defined $data;
971 # blessed objects are treated like scalars
972 my $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
974 return 'SCALAR' unless $ref;
977 while ($ref eq 'REF') {
979 $ref = (Scalar::Util::blessed $data) ? '' : ref $data;
983 return ($ref||'SCALAR') . ('REF' x $n_steps);
987 my ($self, $data) = @_;
988 my @try = ($self->_refkind($data));
989 push @try, 'SCALAR_or_UNDEF' if $try[0] eq 'SCALAR' || $try[0] eq 'UNDEF';
990 push @try, 'FALLBACK';
994 sub _METHOD_FOR_refkind {
995 my ($self, $meth_prefix, $data) = @_;
998 for (@{$self->_try_refkind($data)}) {
999 $method = $self->can($meth_prefix."_".$_)
1003 return $method || puke "cannot dispatch on '$meth_prefix' for ".$self->_refkind($data);
1007 sub _SWITCH_refkind {
1008 my ($self, $data, $dispatch_table) = @_;
1011 for (@{$self->_try_refkind($data)}) {
1012 $coderef = $dispatch_table->{$_}
1016 puke "no dispatch entry for ".$self->_refkind($data)
1025 #======================================================================
1026 # VALUES, GENERATE, AUTOLOAD
1027 #======================================================================
1029 # LDNOTE: original code from nwiger, didn't touch code in that section
1030 # I feel the AUTOLOAD stuff should not be the default, it should
1031 # only be activated on explicit demand by user.
1035 my $data = shift || return;
1036 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
1037 unless ref $data eq 'HASH';
1040 foreach my $k ( sort keys %$data ) {
1041 my $v = $data->{$k};
1042 $self->_SWITCH_refkind($v, {
1044 if ($self->{array_datatypes}) { # array datatype
1045 push @all_bind, $self->_bindtype($k, $v);
1047 else { # literal SQL with bind
1048 my ($sql, @bind) = @$v;
1049 $self->_assert_bindval_matches_bindtype(@bind);
1050 push @all_bind, @bind;
1053 ARRAYREFREF => sub { # literal SQL with bind
1054 my ($sql, @bind) = @${$v};
1055 $self->_assert_bindval_matches_bindtype(@bind);
1056 push @all_bind, @bind;
1058 SCALARREF => sub { # literal SQL without bind
1060 SCALAR_or_UNDEF => sub {
1061 push @all_bind, $self->_bindtype($k, $v);
1072 my(@sql, @sqlq, @sqlv);
1076 if ($ref eq 'HASH') {
1077 for my $k (sort keys %$_) {
1080 my $label = $self->_quote($k);
1081 if ($r eq 'ARRAY') {
1082 # literal SQL with bind
1083 my ($sql, @bind) = @$v;
1084 $self->_assert_bindval_matches_bindtype(@bind);
1085 push @sqlq, "$label = $sql";
1087 } elsif ($r eq 'SCALAR') {
1088 # literal SQL without bind
1089 push @sqlq, "$label = $$v";
1091 push @sqlq, "$label = ?";
1092 push @sqlv, $self->_bindtype($k, $v);
1095 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
1096 } elsif ($ref eq 'ARRAY') {
1097 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
1100 if ($r eq 'ARRAY') { # literal SQL with bind
1101 my ($sql, @bind) = @$v;
1102 $self->_assert_bindval_matches_bindtype(@bind);
1105 } elsif ($r eq 'SCALAR') { # literal SQL without bind
1106 # embedded literal SQL
1113 push @sql, '(' . join(', ', @sqlq) . ')';
1114 } elsif ($ref eq 'SCALAR') {
1118 # strings get case twiddled
1119 push @sql, $self->_sqlcase($_);
1123 my $sql = join ' ', @sql;
1125 # this is pretty tricky
1126 # if ask for an array, return ($stmt, @bind)
1127 # otherwise, s/?/shift @sqlv/ to put it inline
1129 return ($sql, @sqlv);
1131 1 while $sql =~ s/\?/my $d = shift(@sqlv);
1132 ref $d ? $d->[1] : $d/e;
1141 # This allows us to check for a local, then _form, attr
1143 my($name) = $AUTOLOAD =~ /.*::(.+)/;
1144 return $self->generate($name, @_);
1155 SQL::Abstract - Generate SQL from Perl data structures
1161 my $sql = SQL::Abstract->new;
1163 my($stmt, @bind) = $sql->select($table, \@fields, \%where, \@order);
1165 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
1167 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
1169 my($stmt, @bind) = $sql->delete($table, \%where);
1171 # Then, use these in your DBI statements
1172 my $sth = $dbh->prepare($stmt);
1173 $sth->execute(@bind);
1175 # Just generate the WHERE clause
1176 my($stmt, @bind) = $sql->where(\%where, \@order);
1178 # Return values in the same order, for hashed queries
1179 # See PERFORMANCE section for more details
1180 my @bind = $sql->values(\%fieldvals);
1184 This module was inspired by the excellent L<DBIx::Abstract>.
1185 However, in using that module I found that what I really wanted
1186 to do was generate SQL, but still retain complete control over my
1187 statement handles and use the DBI interface. So, I set out to
1188 create an abstract SQL generation module.
1190 While based on the concepts used by L<DBIx::Abstract>, there are
1191 several important differences, especially when it comes to WHERE
1192 clauses. I have modified the concepts used to make the SQL easier
1193 to generate from Perl data structures and, IMO, more intuitive.
1194 The underlying idea is for this module to do what you mean, based
1195 on the data structures you provide it. The big advantage is that
1196 you don't have to modify your code every time your data changes,
1197 as this module figures it out.
1199 To begin with, an SQL INSERT is as easy as just specifying a hash
1200 of C<key=value> pairs:
1203 name => 'Jimbo Bobson',
1204 phone => '123-456-7890',
1205 address => '42 Sister Lane',
1206 city => 'St. Louis',
1207 state => 'Louisiana',
1210 The SQL can then be generated with this:
1212 my($stmt, @bind) = $sql->insert('people', \%data);
1214 Which would give you something like this:
1216 $stmt = "INSERT INTO people
1217 (address, city, name, phone, state)
1218 VALUES (?, ?, ?, ?, ?)";
1219 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
1220 '123-456-7890', 'Louisiana');
1222 These are then used directly in your DBI code:
1224 my $sth = $dbh->prepare($stmt);
1225 $sth->execute(@bind);
1227 =head2 Inserting and Updating Arrays
1229 If your database has array types (like for example Postgres),
1230 activate the special option C<< array_datatypes => 1 >>
1231 when creating the C<SQL::Abstract> object.
1232 Then you may use an arrayref to insert and update database array types:
1234 my $sql = SQL::Abstract->new(array_datatypes => 1);
1236 planets => [qw/Mercury Venus Earth Mars/]
1239 my($stmt, @bind) = $sql->insert('solar_system', \%data);
1243 $stmt = "INSERT INTO solar_system (planets) VALUES (?)"
1245 @bind = (['Mercury', 'Venus', 'Earth', 'Mars']);
1248 =head2 Inserting and Updating SQL
1250 In order to apply SQL functions to elements of your C<%data> you may
1251 specify a reference to an arrayref for the given hash value. For example,
1252 if you need to execute the Oracle C<to_date> function on a value, you can
1253 say something like this:
1257 date_entered => \["to_date(?,'MM/DD/YYYY')", "03/02/2003"],
1260 The first value in the array is the actual SQL. Any other values are
1261 optional and would be included in the bind values array. This gives
1264 my($stmt, @bind) = $sql->insert('people', \%data);
1266 $stmt = "INSERT INTO people (name, date_entered)
1267 VALUES (?, to_date(?,'MM/DD/YYYY'))";
1268 @bind = ('Bill', '03/02/2003');
1270 An UPDATE is just as easy, all you change is the name of the function:
1272 my($stmt, @bind) = $sql->update('people', \%data);
1274 Notice that your C<%data> isn't touched; the module will generate
1275 the appropriately quirky SQL for you automatically. Usually you'll
1276 want to specify a WHERE clause for your UPDATE, though, which is
1277 where handling C<%where> hashes comes in handy...
1279 =head2 Complex where statements
1281 This module can generate pretty complicated WHERE statements
1282 easily. For example, simple C<key=value> pairs are taken to mean
1283 equality, and if you want to see if a field is within a set
1284 of values, you can use an arrayref. Let's say we wanted to
1285 SELECT some data based on this criteria:
1288 requestor => 'inna',
1289 worker => ['nwiger', 'rcwe', 'sfz'],
1290 status => { '!=', 'completed' }
1293 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
1295 The above would give you something like this:
1297 $stmt = "SELECT * FROM tickets WHERE
1298 ( requestor = ? ) AND ( status != ? )
1299 AND ( worker = ? OR worker = ? OR worker = ? )";
1300 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
1302 Which you could then use in DBI code like so:
1304 my $sth = $dbh->prepare($stmt);
1305 $sth->execute(@bind);
1311 The functions are simple. There's one for each major SQL operation,
1312 and a constructor you use first. The arguments are specified in a
1313 similar order to each function (table, then fields, then a where
1314 clause) to try and simplify things.
1319 =head2 new(option => 'value')
1321 The C<new()> function takes a list of options and values, and returns
1322 a new B<SQL::Abstract> object which can then be used to generate SQL
1323 through the methods below. The options accepted are:
1329 If set to 'lower', then SQL will be generated in all lowercase. By
1330 default SQL is generated in "textbook" case meaning something like:
1332 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
1334 Any setting other than 'lower' is ignored.
1338 This determines what the default comparison operator is. By default
1339 it is C<=>, meaning that a hash like this:
1341 %where = (name => 'nwiger', email => 'nate@wiger.org');
1343 Will generate SQL like this:
1345 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
1347 However, you may want loose comparisons by default, so if you set
1348 C<cmp> to C<like> you would get SQL such as:
1350 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
1352 You can also override the comparsion on an individual basis - see
1353 the huge section on L</"WHERE CLAUSES"> at the bottom.
1355 =item sqltrue, sqlfalse
1357 Expressions for inserting boolean values within SQL statements.
1358 By default these are C<1=1> and C<1=0>. They are used
1359 by the special operators C<-in> and C<-not_in> for generating
1360 correct SQL even when the argument is an empty array (see below).
1364 This determines the default logical operator for multiple WHERE
1365 statements in arrays or hashes. If absent, the default logic is "or"
1366 for arrays, and "and" for hashes. This means that a WHERE
1370 event_date => {'>=', '2/13/99'},
1371 event_date => {'<=', '4/24/03'},
1374 will generate SQL like this:
1376 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
1378 This is probably not what you want given this query, though (look
1379 at the dates). To change the "OR" to an "AND", simply specify:
1381 my $sql = SQL::Abstract->new(logic => 'and');
1383 Which will change the above C<WHERE> to:
1385 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
1387 The logic can also be changed locally by inserting
1388 a modifier in front of an arrayref :
1390 @where = (-and => [event_date => {'>=', '2/13/99'},
1391 event_date => {'<=', '4/24/03'} ]);
1393 See the L</"WHERE CLAUSES"> section for explanations.
1397 This will automatically convert comparisons using the specified SQL
1398 function for both column and value. This is mostly used with an argument
1399 of C<upper> or C<lower>, so that the SQL will have the effect of
1400 case-insensitive "searches". For example, this:
1402 $sql = SQL::Abstract->new(convert => 'upper');
1403 %where = (keywords => 'MaKe iT CAse inSeNSItive');
1405 Will turn out the following SQL:
1407 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
1409 The conversion can be C<upper()>, C<lower()>, or any other SQL function
1410 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
1411 not validate this option; it will just pass through what you specify verbatim).
1415 This is a kludge because many databases suck. For example, you can't
1416 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
1417 Instead, you have to use C<bind_param()>:
1419 $sth->bind_param(1, 'reg data');
1420 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
1422 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
1423 which loses track of which field each slot refers to. Fear not.
1425 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
1426 Currently, you can specify either C<normal> (default) or C<columns>. If you
1427 specify C<columns>, you will get an array that looks like this:
1429 my $sql = SQL::Abstract->new(bindtype => 'columns');
1430 my($stmt, @bind) = $sql->insert(...);
1433 [ 'column1', 'value1' ],
1434 [ 'column2', 'value2' ],
1435 [ 'column3', 'value3' ],
1438 You can then iterate through this manually, using DBI's C<bind_param()>.
1440 $sth->prepare($stmt);
1443 my($col, $data) = @$_;
1444 if ($col eq 'details' || $col eq 'comments') {
1445 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
1446 } elsif ($col eq 'image') {
1447 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
1449 $sth->bind_param($i, $data);
1453 $sth->execute; # execute without @bind now
1455 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
1456 Basically, the advantage is still that you don't have to care which fields
1457 are or are not included. You could wrap that above C<for> loop in a simple
1458 sub called C<bind_fields()> or something and reuse it repeatedly. You still
1459 get a layer of abstraction over manual SQL specification.
1461 Note that if you set L</bindtype> to C<columns>, the C<\[$sql, @bind]>
1462 construct (see L</Literal SQL with placeholders and bind values (subqueries)>)
1463 will expect the bind values in this format.
1467 This is the character that a table or column name will be quoted
1468 with. By default this is an empty string, but you could set it to
1469 the character C<`>, to generate SQL like this:
1471 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
1473 Alternatively, you can supply an array ref of two items, the first being the left
1474 hand quote character, and the second the right hand quote character. For
1475 example, you could supply C<['[',']']> for SQL Server 2000 compliant quotes
1476 that generates SQL like this:
1478 SELECT [a_field] FROM [a_table] WHERE [some_field] LIKE '%someval%'
1480 Quoting is useful if you have tables or columns names that are reserved
1481 words in your database's SQL dialect.
1485 This is the character that separates a table and column name. It is
1486 necessary to specify this when the C<quote_char> option is selected,
1487 so that tables and column names can be individually quoted like this:
1489 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
1491 =item injection_guard
1493 A regular expression C<qr/.../> that is applied to any C<-function> and unquoted
1494 column name specified in a query structure. This is a safety mechanism to avoid
1495 injection attacks when mishandling user input e.g.:
1497 my %condition_as_column_value_pairs = get_values_from_user();
1498 $sqla->select( ... , \%condition_as_column_value_pairs );
1500 If the expression matches an exception is thrown. Note that literal SQL
1501 supplied via C<\'...'> or C<\['...']> is B<not> checked in any way.
1503 Defaults to checking for C<;> and the C<GO> keyword (TransactSQL)
1505 =item array_datatypes
1507 When this option is true, arrayrefs in INSERT or UPDATE are
1508 interpreted as array datatypes and are passed directly
1510 When this option is false, arrayrefs are interpreted
1511 as literal SQL, just like refs to arrayrefs
1512 (but this behavior is for backwards compatibility; when writing
1513 new queries, use the "reference to arrayref" syntax
1519 Takes a reference to a list of "special operators"
1520 to extend the syntax understood by L<SQL::Abstract>.
1521 See section L</"SPECIAL OPERATORS"> for details.
1525 Takes a reference to a list of "unary operators"
1526 to extend the syntax understood by L<SQL::Abstract>.
1527 See section L</"UNARY OPERATORS"> for details.
1533 =head2 insert($table, \@values || \%fieldvals, \%options)
1535 This is the simplest function. You simply give it a table name
1536 and either an arrayref of values or hashref of field/value pairs.
1537 It returns an SQL INSERT statement and a list of bind values.
1538 See the sections on L</"Inserting and Updating Arrays"> and
1539 L</"Inserting and Updating SQL"> for information on how to insert
1540 with those data types.
1542 The optional C<\%options> hash reference may contain additional
1543 options to generate the insert SQL. Currently supported options
1550 Takes either a scalar of raw SQL fields, or an array reference of
1551 field names, and adds on an SQL C<RETURNING> statement at the end.
1552 This allows you to return data generated by the insert statement
1553 (such as row IDs) without performing another C<SELECT> statement.
1554 Note, however, this is not part of the SQL standard and may not
1555 be supported by all database engines.
1559 =head2 update($table, \%fieldvals, \%where)
1561 This takes a table, hashref of field/value pairs, and an optional
1562 hashref L<WHERE clause|/WHERE CLAUSES>. It returns an SQL UPDATE function and a list
1564 See the sections on L</"Inserting and Updating Arrays"> and
1565 L</"Inserting and Updating SQL"> for information on how to insert
1566 with those data types.
1568 =head2 select($source, $fields, $where, $order)
1570 This returns a SQL SELECT statement and associated list of bind values, as
1571 specified by the arguments :
1577 Specification of the 'FROM' part of the statement.
1578 The argument can be either a plain scalar (interpreted as a table
1579 name, will be quoted), or an arrayref (interpreted as a list
1580 of table names, joined by commas, quoted), or a scalarref
1581 (literal table name, not quoted), or a ref to an arrayref
1582 (list of literal table names, joined by commas, not quoted).
1586 Specification of the list of fields to retrieve from
1588 The argument can be either an arrayref (interpreted as a list
1589 of field names, will be joined by commas and quoted), or a
1590 plain scalar (literal SQL, not quoted).
1591 Please observe that this API is not as flexible as for
1592 the first argument C<$table>, for backwards compatibility reasons.
1596 Optional argument to specify the WHERE part of the query.
1597 The argument is most often a hashref, but can also be
1598 an arrayref or plain scalar --
1599 see section L<WHERE clause|/"WHERE CLAUSES"> for details.
1603 Optional argument to specify the ORDER BY part of the query.
1604 The argument can be a scalar, a hashref or an arrayref
1605 -- see section L<ORDER BY clause|/"ORDER BY CLAUSES">
1611 =head2 delete($table, \%where)
1613 This takes a table name and optional hashref L<WHERE clause|/WHERE CLAUSES>.
1614 It returns an SQL DELETE statement and list of bind values.
1616 =head2 where(\%where, \@order)
1618 This is used to generate just the WHERE clause. For example,
1619 if you have an arbitrary data structure and know what the
1620 rest of your SQL is going to look like, but want an easy way
1621 to produce a WHERE clause, use this. It returns an SQL WHERE
1622 clause and list of bind values.
1625 =head2 values(\%data)
1627 This just returns the values from the hash C<%data>, in the same
1628 order that would be returned from any of the other above queries.
1629 Using this allows you to markedly speed up your queries if you
1630 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
1632 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
1634 Warning: This is an experimental method and subject to change.
1636 This returns arbitrarily generated SQL. It's a really basic shortcut.
1637 It will return two different things, depending on return context:
1639 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
1640 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
1642 These would return the following:
1644 # First calling form
1645 $stmt = "CREATE TABLE test (?, ?)";
1646 @bind = (field1, field2);
1648 # Second calling form
1649 $stmt_and_val = "CREATE TABLE test (field1, field2)";
1651 Depending on what you're trying to do, it's up to you to choose the correct
1652 format. In this example, the second form is what you would want.
1656 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
1660 ALTER SESSION SET nls_date_format = 'MM/YY'
1662 You get the idea. Strings get their case twiddled, but everything
1663 else remains verbatim.
1665 =head1 WHERE CLAUSES
1669 This module uses a variation on the idea from L<DBIx::Abstract>. It
1670 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
1671 module is that things in arrays are OR'ed, and things in hashes
1674 The easiest way to explain is to show lots of examples. After
1675 each C<%where> hash shown, it is assumed you used:
1677 my($stmt, @bind) = $sql->where(\%where);
1679 However, note that the C<%where> hash can be used directly in any
1680 of the other functions as well, as described above.
1682 =head2 Key-value pairs
1684 So, let's get started. To begin, a simple hash:
1688 status => 'completed'
1691 Is converted to SQL C<key = val> statements:
1693 $stmt = "WHERE user = ? AND status = ?";
1694 @bind = ('nwiger', 'completed');
1696 One common thing I end up doing is having a list of values that
1697 a field can be in. To do this, simply specify a list inside of
1702 status => ['assigned', 'in-progress', 'pending'];
1705 This simple code will create the following:
1707 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
1708 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
1710 A field associated to an empty arrayref will be considered a
1711 logical false and will generate 0=1.
1713 =head2 Tests for NULL values
1715 If the value part is C<undef> then this is converted to SQL <IS NULL>
1724 $stmt = "WHERE user = ? AND status IS NULL";
1727 To test if a column IS NOT NULL:
1731 status => { '!=', undef },
1734 =head2 Specific comparison operators
1736 If you want to specify a different type of operator for your comparison,
1737 you can use a hashref for a given column:
1741 status => { '!=', 'completed' }
1744 Which would generate:
1746 $stmt = "WHERE user = ? AND status != ?";
1747 @bind = ('nwiger', 'completed');
1749 To test against multiple values, just enclose the values in an arrayref:
1751 status => { '=', ['assigned', 'in-progress', 'pending'] };
1753 Which would give you:
1755 "WHERE status = ? OR status = ? OR status = ?"
1758 The hashref can also contain multiple pairs, in which case it is expanded
1759 into an C<AND> of its elements:
1763 status => { '!=', 'completed', -not_like => 'pending%' }
1766 # Or more dynamically, like from a form
1767 $where{user} = 'nwiger';
1768 $where{status}{'!='} = 'completed';
1769 $where{status}{'-not_like'} = 'pending%';
1771 # Both generate this
1772 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
1773 @bind = ('nwiger', 'completed', 'pending%');
1776 To get an OR instead, you can combine it with the arrayref idea:
1780 priority => [ { '=', 2 }, { '>', 5 } ]
1783 Which would generate:
1785 $stmt = "WHERE ( priority = ? OR priority > ? ) AND user = ?";
1786 @bind = ('2', '5', 'nwiger');
1788 If you want to include literal SQL (with or without bind values), just use a
1789 scalar reference or array reference as the value:
1792 date_entered => { '>' => \["to_date(?, 'MM/DD/YYYY')", "11/26/2008"] },
1793 date_expires => { '<' => \"now()" }
1796 Which would generate:
1798 $stmt = "WHERE date_entered > "to_date(?, 'MM/DD/YYYY') AND date_expires < now()";
1799 @bind = ('11/26/2008');
1802 =head2 Logic and nesting operators
1804 In the example above,
1805 there is a subtle trap if you want to say something like
1806 this (notice the C<AND>):
1808 WHERE priority != ? AND priority != ?
1810 Because, in Perl you I<can't> do this:
1812 priority => { '!=', 2, '!=', 1 }
1814 As the second C<!=> key will obliterate the first. The solution
1815 is to use the special C<-modifier> form inside an arrayref:
1817 priority => [ -and => {'!=', 2},
1821 Normally, these would be joined by C<OR>, but the modifier tells it
1822 to use C<AND> instead. (Hint: You can use this in conjunction with the
1823 C<logic> option to C<new()> in order to change the way your queries
1824 work by default.) B<Important:> Note that the C<-modifier> goes
1825 B<INSIDE> the arrayref, as an extra first element. This will
1826 B<NOT> do what you think it might:
1828 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
1830 Here is a quick list of equivalencies, since there is some overlap:
1833 status => {'!=', 'completed', 'not like', 'pending%' }
1834 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
1837 status => {'=', ['assigned', 'in-progress']}
1838 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
1839 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
1843 =head2 Special operators : IN, BETWEEN, etc.
1845 You can also use the hashref format to compare a list of fields using the
1846 C<IN> comparison operator, by specifying the list as an arrayref:
1849 status => 'completed',
1850 reportid => { -in => [567, 2335, 2] }
1853 Which would generate:
1855 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
1856 @bind = ('completed', '567', '2335', '2');
1858 The reverse operator C<-not_in> generates SQL C<NOT IN> and is used in
1861 If the argument to C<-in> is an empty array, 'sqlfalse' is generated
1862 (by default : C<1=0>). Similarly, C<< -not_in => [] >> generates
1863 'sqltrue' (by default : C<1=1>).
1865 In addition to the array you can supply a chunk of literal sql or
1866 literal sql with bind:
1869 customer => { -in => \[
1870 'SELECT cust_id FROM cust WHERE balance > ?',
1873 status => { -in => \'SELECT status_codes FROM states' },
1879 customer IN ( SELECT cust_id FROM cust WHERE balance > ? )
1880 AND status IN ( SELECT status_codes FROM states )
1886 Another pair of operators is C<-between> and C<-not_between>,
1887 used with an arrayref of two values:
1891 completion_date => {
1892 -not_between => ['2002-10-01', '2003-02-06']
1898 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
1900 Just like with C<-in> all plausible combinations of literal SQL
1904 start0 => { -between => [ 1, 2 ] },
1905 start1 => { -between => \["? AND ?", 1, 2] },
1906 start2 => { -between => \"lower(x) AND upper(y)" },
1907 start3 => { -between => [
1909 \["upper(?)", 'stuff' ],
1916 ( start0 BETWEEN ? AND ? )
1917 AND ( start1 BETWEEN ? AND ? )
1918 AND ( start2 BETWEEN lower(x) AND upper(y) )
1919 AND ( start3 BETWEEN lower(x) AND upper(?) )
1921 @bind = (1, 2, 1, 2, 'stuff');
1924 These are the two builtin "special operators"; but the
1925 list can be expanded : see section L</"SPECIAL OPERATORS"> below.
1927 =head2 Unary operators: bool
1929 If you wish to test against boolean columns or functions within your
1930 database you can use the C<-bool> and C<-not_bool> operators. For
1931 example to test the column C<is_user> being true and the column
1932 C<is_enabled> being false you would use:-
1936 -not_bool => 'is_enabled',
1941 WHERE is_user AND NOT is_enabled
1943 If a more complex combination is required, testing more conditions,
1944 then you should use the and/or operators:-
1951 -not_bool => 'four',
1957 WHERE one AND two AND three AND NOT four
1960 =head2 Nested conditions, -and/-or prefixes
1962 So far, we've seen how multiple conditions are joined with a top-level
1963 C<AND>. We can change this by putting the different conditions we want in
1964 hashes and then putting those hashes in an array. For example:
1969 status => { -like => ['pending%', 'dispatched'] },
1973 status => 'unassigned',
1977 This data structure would create the following:
1979 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
1980 OR ( user = ? AND status = ? ) )";
1981 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
1984 Clauses in hashrefs or arrayrefs can be prefixed with an C<-and> or C<-or>
1985 to change the logic inside :
1991 -and => [ workhrs => {'>', 20}, geo => 'ASIA' ],
1992 -or => { workhrs => {'<', 50}, geo => 'EURO' },
1999 WHERE ( user = ? AND (
2000 ( workhrs > ? AND geo = ? )
2001 OR ( workhrs < ? OR geo = ? )
2004 =head3 Algebraic inconsistency, for historical reasons
2006 C<Important note>: when connecting several conditions, the C<-and->|C<-or>
2007 operator goes C<outside> of the nested structure; whereas when connecting
2008 several constraints on one column, the C<-and> operator goes
2009 C<inside> the arrayref. Here is an example combining both features :
2012 -and => [a => 1, b => 2],
2013 -or => [c => 3, d => 4],
2014 e => [-and => {-like => 'foo%'}, {-like => '%bar'} ]
2019 WHERE ( ( ( a = ? AND b = ? )
2020 OR ( c = ? OR d = ? )
2021 OR ( e LIKE ? AND e LIKE ? ) ) )
2023 This difference in syntax is unfortunate but must be preserved for
2024 historical reasons. So be careful : the two examples below would
2025 seem algebraically equivalent, but they are not
2027 {col => [-and => {-like => 'foo%'}, {-like => '%bar'}]}
2028 # yields : WHERE ( ( col LIKE ? AND col LIKE ? ) )
2030 [-and => {col => {-like => 'foo%'}, {col => {-like => '%bar'}}]]
2031 # yields : WHERE ( ( col LIKE ? OR col LIKE ? ) )
2034 =head2 Literal SQL and value type operators
2036 The basic premise of SQL::Abstract is that in WHERE specifications the "left
2037 side" is a column name and the "right side" is a value (normally rendered as
2038 a placeholder). This holds true for both hashrefs and arrayref pairs as you
2039 see in the L</WHERE CLAUSES> examples above. Sometimes it is necessary to
2040 alter this behavior. There are several ways of doing so.
2044 This is a virtual operator that signals the string to its right side is an
2045 identifier (a column name) and not a value. For example to compare two
2046 columns you would write:
2049 priority => { '<', 2 },
2050 requestor => { -ident => 'submitter' },
2055 $stmt = "WHERE priority < ? AND requestor = submitter";
2058 If you are maintaining legacy code you may see a different construct as
2059 described in L</Deprecated usage of Literal SQL>, please use C<-ident> in new
2064 This is a virtual operator that signals that the construct to its right side
2065 is a value to be passed to DBI. This is for example necessary when you want
2066 to write a where clause against an array (for RDBMS that support such
2067 datatypes). For example:
2070 array => { -value => [1, 2, 3] }
2075 $stmt = 'WHERE array = ?';
2076 @bind = ([1, 2, 3]);
2078 Note that if you were to simply say:
2084 the result would porbably be not what you wanted:
2086 $stmt = 'WHERE array = ? OR array = ? OR array = ?';
2091 Finally, sometimes only literal SQL will do. To include a random snippet
2092 of SQL verbatim, you specify it as a scalar reference. Consider this only
2093 as a last resort. Usually there is a better way. For example:
2096 priority => { '<', 2 },
2097 requestor => { -in => \'(SELECT name FROM hitmen)' },
2102 $stmt = "WHERE priority < ? AND requestor IN (SELECT name FROM hitmen)"
2105 Note that in this example, you only get one bind parameter back, since
2106 the verbatim SQL is passed as part of the statement.
2110 Never use untrusted input as a literal SQL argument - this is a massive
2111 security risk (there is no way to check literal snippets for SQL
2112 injections and other nastyness). If you need to deal with untrusted input
2113 use literal SQL with placeholders as described next.
2115 =head3 Literal SQL with placeholders and bind values (subqueries)
2117 If the literal SQL to be inserted has placeholders and bind values,
2118 use a reference to an arrayref (yes this is a double reference --
2119 not so common, but perfectly legal Perl). For example, to find a date
2120 in Postgres you can use something like this:
2123 date_column => \[q/= date '2008-09-30' - ?::integer/, 10/]
2128 $stmt = "WHERE ( date_column = date '2008-09-30' - ?::integer )"
2131 Note that you must pass the bind values in the same format as they are returned
2132 by L</where>. That means that if you set L</bindtype> to C<columns>, you must
2133 provide the bind values in the C<< [ column_meta => value ] >> format, where
2134 C<column_meta> is an opaque scalar value; most commonly the column name, but
2135 you can use any scalar value (including references and blessed references),
2136 L<SQL::Abstract> will simply pass it through intact. So if C<bindtype> is set
2137 to C<columns> the above example will look like:
2140 date_column => \[q/= date '2008-09-30' - ?::integer/, [ dummy => 10 ]/]
2143 Literal SQL is especially useful for nesting parenthesized clauses in the
2144 main SQL query. Here is a first example :
2146 my ($sub_stmt, @sub_bind) = ("SELECT c1 FROM t1 WHERE c2 < ? AND c3 LIKE ?",
2150 bar => \["IN ($sub_stmt)" => @sub_bind],
2155 $stmt = "WHERE (foo = ? AND bar IN (SELECT c1 FROM t1
2156 WHERE c2 < ? AND c3 LIKE ?))";
2157 @bind = (1234, 100, "foo%");
2159 Other subquery operators, like for example C<"E<gt> ALL"> or C<"NOT IN">,
2160 are expressed in the same way. Of course the C<$sub_stmt> and
2161 its associated bind values can be generated through a former call
2164 my ($sub_stmt, @sub_bind)
2165 = $sql->select("t1", "c1", {c2 => {"<" => 100},
2166 c3 => {-like => "foo%"}});
2169 bar => \["> ALL ($sub_stmt)" => @sub_bind],
2172 In the examples above, the subquery was used as an operator on a column;
2173 but the same principle also applies for a clause within the main C<%where>
2174 hash, like an EXISTS subquery :
2176 my ($sub_stmt, @sub_bind)
2177 = $sql->select("t1", "*", {c1 => 1, c2 => \"> t0.c0"});
2178 my %where = ( -and => [
2180 \["EXISTS ($sub_stmt)" => @sub_bind],
2185 $stmt = "WHERE (foo = ? AND EXISTS (SELECT * FROM t1
2186 WHERE c1 = ? AND c2 > t0.c0))";
2190 Observe that the condition on C<c2> in the subquery refers to
2191 column C<t0.c0> of the main query : this is I<not> a bind
2192 value, so we have to express it through a scalar ref.
2193 Writing C<< c2 => {">" => "t0.c0"} >> would have generated
2194 C<< c2 > ? >> with bind value C<"t0.c0"> ... not exactly
2195 what we wanted here.
2197 Finally, here is an example where a subquery is used
2198 for expressing unary negation:
2200 my ($sub_stmt, @sub_bind)
2201 = $sql->where({age => [{"<" => 10}, {">" => 20}]});
2202 $sub_stmt =~ s/^ where //i; # don't want "WHERE" in the subclause
2204 lname => {like => '%son%'},
2205 \["NOT ($sub_stmt)" => @sub_bind],
2210 $stmt = "lname LIKE ? AND NOT ( age < ? OR age > ? )"
2211 @bind = ('%son%', 10, 20)
2213 =head3 Deprecated usage of Literal SQL
2215 Below are some examples of archaic use of literal SQL. It is shown only as
2216 reference for those who deal with legacy code. Each example has a much
2217 better, cleaner and safer alternative that users should opt for in new code.
2223 my %where = ( requestor => \'IS NOT NULL' )
2225 $stmt = "WHERE requestor IS NOT NULL"
2227 This used to be the way of generating NULL comparisons, before the handling
2228 of C<undef> got formalized. For new code please use the superior syntax as
2229 described in L</Tests for NULL values>.
2233 my %where = ( requestor => \'= submitter' )
2235 $stmt = "WHERE requestor = submitter"
2237 This used to be the only way to compare columns. Use the superior L</-ident>
2238 method for all new code. For example an identifier declared in such a way
2239 will be properly quoted if L</quote_char> is properly set, while the legacy
2240 form will remain as supplied.
2244 my %where = ( is_ready => \"", completed => { '>', '2012-12-21' } )
2246 $stmt = "WHERE completed > ? AND is_ready"
2247 @bind = ('2012-12-21')
2249 Using an empty string literal used to be the only way to express a boolean.
2250 For all new code please use the much more readable
2251 L<-bool|/Unary operators: bool> operator.
2257 These pages could go on for a while, since the nesting of the data
2258 structures this module can handle are pretty much unlimited (the
2259 module implements the C<WHERE> expansion as a recursive function
2260 internally). Your best bet is to "play around" with the module a
2261 little to see how the data structures behave, and choose the best
2262 format for your data based on that.
2264 And of course, all the values above will probably be replaced with
2265 variables gotten from forms or the command line. After all, if you
2266 knew everything ahead of time, you wouldn't have to worry about
2267 dynamically-generating SQL and could just hardwire it into your
2270 =head1 ORDER BY CLAUSES
2272 Some functions take an order by clause. This can either be a scalar (just a
2273 column name,) a hash of C<< { -desc => 'col' } >> or C<< { -asc => 'col' } >>,
2274 or an array of either of the two previous forms. Examples:
2276 Given | Will Generate
2277 ----------------------------------------------------------
2279 \'colA DESC' | ORDER BY colA DESC
2281 'colA' | ORDER BY colA
2283 [qw/colA colB/] | ORDER BY colA, colB
2285 {-asc => 'colA'} | ORDER BY colA ASC
2287 {-desc => 'colB'} | ORDER BY colB DESC
2289 ['colA', {-asc => 'colB'}] | ORDER BY colA, colB ASC
2291 { -asc => [qw/colA colB/] } | ORDER BY colA ASC, colB ASC
2294 { -asc => 'colA' }, | ORDER BY colA ASC, colB DESC,
2295 { -desc => [qw/colB/], | colC ASC, colD ASC
2296 { -asc => [qw/colC colD/],|
2298 ===========================================================
2302 =head1 SPECIAL OPERATORS
2304 my $sqlmaker = SQL::Abstract->new(special_ops => [
2308 my ($self, $field, $op, $arg) = @_;
2314 handler => 'method_name',
2318 A "special operator" is a SQL syntactic clause that can be
2319 applied to a field, instead of a usual binary operator.
2322 WHERE field IN (?, ?, ?)
2323 WHERE field BETWEEN ? AND ?
2324 WHERE MATCH(field) AGAINST (?, ?)
2326 Special operators IN and BETWEEN are fairly standard and therefore
2327 are builtin within C<SQL::Abstract> (as the overridable methods
2328 C<_where_field_IN> and C<_where_field_BETWEEN>). For other operators,
2329 like the MATCH .. AGAINST example above which is specific to MySQL,
2330 you can write your own operator handlers - supply a C<special_ops>
2331 argument to the C<new> method. That argument takes an arrayref of
2332 operator definitions; each operator definition is a hashref with two
2339 the regular expression to match the operator
2343 Either a coderef or a plain scalar method name. In both cases
2344 the expected return is C<< ($sql, @bind) >>.
2346 When supplied with a method name, it is simply called on the
2347 L<SQL::Abstract/> object as:
2349 $self->$method_name ($field, $op, $arg)
2353 $op is the part that matched the handler regex
2354 $field is the LHS of the operator
2357 When supplied with a coderef, it is called as:
2359 $coderef->($self, $field, $op, $arg)
2364 For example, here is an implementation
2365 of the MATCH .. AGAINST syntax for MySQL
2367 my $sqlmaker = SQL::Abstract->new(special_ops => [
2369 # special op for MySql MATCH (field) AGAINST(word1, word2, ...)
2370 {regex => qr/^match$/i,
2372 my ($self, $field, $op, $arg) = @_;
2373 $arg = [$arg] if not ref $arg;
2374 my $label = $self->_quote($field);
2375 my ($placeholder) = $self->_convert('?');
2376 my $placeholders = join ", ", (($placeholder) x @$arg);
2377 my $sql = $self->_sqlcase('match') . " ($label) "
2378 . $self->_sqlcase('against') . " ($placeholders) ";
2379 my @bind = $self->_bindtype($field, @$arg);
2380 return ($sql, @bind);
2387 =head1 UNARY OPERATORS
2389 my $sqlmaker = SQL::Abstract->new(unary_ops => [
2393 my ($self, $op, $arg) = @_;
2399 handler => 'method_name',
2403 A "unary operator" is a SQL syntactic clause that can be
2404 applied to a field - the operator goes before the field
2406 You can write your own operator handlers - supply a C<unary_ops>
2407 argument to the C<new> method. That argument takes an arrayref of
2408 operator definitions; each operator definition is a hashref with two
2415 the regular expression to match the operator
2419 Either a coderef or a plain scalar method name. In both cases
2420 the expected return is C<< $sql >>.
2422 When supplied with a method name, it is simply called on the
2423 L<SQL::Abstract/> object as:
2425 $self->$method_name ($op, $arg)
2429 $op is the part that matched the handler regex
2430 $arg is the RHS or argument of the operator
2432 When supplied with a coderef, it is called as:
2434 $coderef->($self, $op, $arg)
2442 Thanks to some benchmarking by Mark Stosberg, it turns out that
2443 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
2444 I must admit this wasn't an intentional design issue, but it's a
2445 byproduct of the fact that you get to control your C<DBI> handles
2448 To maximize performance, use a code snippet like the following:
2450 # prepare a statement handle using the first row
2451 # and then reuse it for the rest of the rows
2453 for my $href (@array_of_hashrefs) {
2454 $stmt ||= $sql->insert('table', $href);
2455 $sth ||= $dbh->prepare($stmt);
2456 $sth->execute($sql->values($href));
2459 The reason this works is because the keys in your C<$href> are sorted
2460 internally by B<SQL::Abstract>. Thus, as long as your data retains
2461 the same structure, you only have to generate the SQL the first time
2462 around. On subsequent queries, simply use the C<values> function provided
2463 by this module to return your values in the correct order.
2465 However this depends on the values having the same type - if, for
2466 example, the values of a where clause may either have values
2467 (resulting in sql of the form C<column = ?> with a single bind
2468 value), or alternatively the values might be C<undef> (resulting in
2469 sql of the form C<column IS NULL> with no bind value) then the
2470 caching technique suggested will not work.
2474 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
2475 really like this part (I do, at least). Building up a complex query
2476 can be as simple as the following:
2480 use CGI::FormBuilder;
2483 my $form = CGI::FormBuilder->new(...);
2484 my $sql = SQL::Abstract->new;
2486 if ($form->submitted) {
2487 my $field = $form->field;
2488 my $id = delete $field->{id};
2489 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
2492 Of course, you would still have to connect using C<DBI> to run the
2493 query, but the point is that if you make your form look like your
2494 table, the actual query script can be extremely simplistic.
2496 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
2497 a fast interface to returning and formatting data. I frequently
2498 use these three modules together to write complex database query
2499 apps in under 50 lines.
2505 =item * gitweb: L<http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=dbsrgits/SQL-Abstract.git>
2507 =item * git: L<git://git.shadowcat.co.uk/dbsrgits/SQL-Abstract.git>
2513 Version 1.50 was a major internal refactoring of C<SQL::Abstract>.
2514 Great care has been taken to preserve the I<published> behavior
2515 documented in previous versions in the 1.* family; however,
2516 some features that were previously undocumented, or behaved
2517 differently from the documentation, had to be changed in order
2518 to clarify the semantics. Hence, client code that was relying
2519 on some dark areas of C<SQL::Abstract> v1.*
2520 B<might behave differently> in v1.50.
2522 The main changes are :
2528 support for literal SQL through the C<< \ [$sql, bind] >> syntax.
2532 support for the { operator => \"..." } construct (to embed literal SQL)
2536 support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values)
2540 optional support for L<array datatypes|/"Inserting and Updating Arrays">
2544 defensive programming : check arguments
2548 fixed bug with global logic, which was previously implemented
2549 through global variables yielding side-effects. Prior versions would
2550 interpret C<< [ {cond1, cond2}, [cond3, cond4] ] >>
2551 as C<< "(cond1 AND cond2) OR (cond3 AND cond4)" >>.
2552 Now this is interpreted
2553 as C<< "(cond1 AND cond2) OR (cond3 OR cond4)" >>.
2558 fixed semantics of _bindtype on array args
2562 dropped the C<_anoncopy> of the %where tree. No longer necessary,
2563 we just avoid shifting arrays within that tree.
2567 dropped the C<_modlogic> function
2571 =head1 ACKNOWLEDGEMENTS
2573 There are a number of individuals that have really helped out with
2574 this module. Unfortunately, most of them submitted bugs via CPAN
2575 so I have no idea who they are! But the people I do know are:
2577 Ash Berlin (order_by hash term support)
2578 Matt Trout (DBIx::Class support)
2579 Mark Stosberg (benchmarking)
2580 Chas Owens (initial "IN" operator support)
2581 Philip Collins (per-field SQL functions)
2582 Eric Kolve (hashref "AND" support)
2583 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
2584 Dan Kubb (support for "quote_char" and "name_sep")
2585 Guillermo Roditi (patch to cleanup "IN" and "BETWEEN", fix and tests for _order_by)
2586 Laurent Dami (internal refactoring, extensible list of special operators, literal SQL)
2587 Norbert Buchmuller (support for literal SQL in hashpair, misc. fixes & tests)
2588 Peter Rabbitson (rewrite of SQLA::Test, misc. fixes & tests)
2589 Oliver Charles (support for "RETURNING" after "INSERT")
2595 L<DBIx::Class>, L<DBIx::Abstract>, L<CGI::FormBuilder>, L<HTML::QuickTable>.
2599 Copyright (c) 2001-2007 Nathan Wiger <nwiger@cpan.org>. All Rights Reserved.
2601 This module is actively maintained by Matt Trout <mst@shadowcatsystems.co.uk>
2603 For support, your best bet is to try the C<DBIx::Class> users mailing list.
2604 While not an official support venue, C<DBIx::Class> makes heavy use of
2605 C<SQL::Abstract>, and as such list members there are very familiar with
2606 how to create queries.
2610 This module is free software; you may copy this under the same
2611 terms as perl itself (either the GNU General Public License or
2612 the Artistic License)