6 SQL::Abstract - Generate SQL from Perl data structures
12 my $sql = SQL::Abstract->new;
14 my($stmt, @bind) = $sql->select($table, \@fields, \%where, \@order);
16 my($stmt, @bind) = $sql->insert($table, \%fieldvals || \@values);
18 my($stmt, @bind) = $sql->update($table, \%fieldvals, \%where);
20 my($stmt, @bind) = $sql->delete($table, \%where);
22 # Then, use these in your DBI statements
23 my $sth = $dbh->prepare($stmt);
26 # Just generate the WHERE clause
27 my($stmt, @bind) = $sql->where(\%where, \@order);
29 # Return values in the same order, for hashed queries
30 # See PERFORMANCE section for more details
31 my @bind = $sql->values(\%fieldvals);
35 This module was inspired by the excellent L<DBIx::Abstract>.
36 However, in using that module I found that what I really wanted
37 to do was generate SQL, but still retain complete control over my
38 statement handles and use the DBI interface. So, I set out to
39 create an abstract SQL generation module.
41 While based on the concepts used by L<DBIx::Abstract>, there are
42 several important differences, especially when it comes to WHERE
43 clauses. I have modified the concepts used to make the SQL easier
44 to generate from Perl data structures and, IMO, more intuitive.
45 The underlying idea is for this module to do what you mean, based
46 on the data structures you provide it. The big advantage is that
47 you don't have to modify your code every time your data changes,
48 as this module figures it out.
50 To begin with, an SQL INSERT is as easy as just specifying a hash
51 of C<key=value> pairs:
54 name => 'Jimbo Bobson',
55 phone => '123-456-7890',
56 address => '42 Sister Lane',
61 The SQL can then be generated with this:
63 my($stmt, @bind) = $sql->insert('people', \%data);
65 Which would give you something like this:
67 $stmt = "INSERT INTO people
68 (address, city, name, phone, state)
69 VALUES (?, ?, ?, ?, ?)";
70 @bind = ('42 Sister Lane', 'St. Louis', 'Jimbo Bobson',
71 '123-456-7890', 'Louisiana');
73 These are then used directly in your DBI code:
75 my $sth = $dbh->prepare($stmt);
78 In addition, you can apply SQL functions to elements of your C<%data>
79 by specifying an arrayref for the given hash value. For example, if
80 you need to execute the Oracle C<to_date> function on a value, you
81 can say something like this:
85 date_entered => ["to_date(?,'MM/DD/YYYY')", "03/02/2003"],
88 The first value in the array is the actual SQL. Any other values are
89 optional and would be included in the bind values array. This gives
92 my($stmt, @bind) = $sql->insert('people', \%data);
94 $stmt = "INSERT INTO people (name, date_entered)
95 VALUES (?, to_date(?,'MM/DD/YYYY'))";
96 @bind = ('Bill', '03/02/2003');
98 An UPDATE is just as easy, all you change is the name of the function:
100 my($stmt, @bind) = $sql->update('people', \%data);
102 Notice that your C<%data> isn't touched; the module will generate
103 the appropriately quirky SQL for you automatically. Usually you'll
104 want to specify a WHERE clause for your UPDATE, though, which is
105 where handling C<%where> hashes comes in handy...
107 This module can generate pretty complicated WHERE statements
108 easily. For example, simple C<key=value> pairs are taken to mean
109 equality, and if you want to see if a field is within a set
110 of values, you can use an arrayref. Let's say we wanted to
111 SELECT some data based on this criteria:
115 worker => ['nwiger', 'rcwe', 'sfz'],
116 status => { '!=', 'completed' }
119 my($stmt, @bind) = $sql->select('tickets', '*', \%where);
121 The above would give you something like this:
123 $stmt = "SELECT * FROM tickets WHERE
124 ( requestor = ? ) AND ( status != ? )
125 AND ( worker = ? OR worker = ? OR worker = ? )";
126 @bind = ('inna', 'completed', 'nwiger', 'rcwe', 'sfz');
128 Which you could then use in DBI code like so:
130 my $sth = $dbh->prepare($stmt);
131 $sth->execute(@bind);
137 The functions are simple. There's one for each major SQL operation,
138 and a constructor you use first. The arguments are specified in a
139 similar order to each function (table, then fields, then a where
140 clause) to try and simplify things.
147 our $VERSION = '1.22';
148 our $REVISION = '$Id$';
151 # Fix SQL case, if so requested
154 return $self->{case} ? $_[0] : uc($_[0]);
157 # Anon copies of arrays/hashes
158 # Based on deep_copy example by merlyn
159 # http://www.stonehenge.com/merlyn/UnixReview/col30.html
162 return (ref $orig eq 'HASH') ? +{map { $_ => _anoncopy($orig->{$_}) } keys %$orig}
163 : (ref $orig eq 'ARRAY') ? [map _anoncopy($_), @$orig]
169 return unless $_[0]->{debug}; shift; # a little faster
170 my $func = (caller(1))[3];
171 warn "[$func] ", @_, "\n";
175 my($func) = (caller(1))[3];
176 carp "[$func] Warning: ", @_;
180 my($func) = (caller(1))[3];
181 croak "[$func] Fatal: ", @_;
188 if (ref $tab eq 'ARRAY') {
189 return join ', ', map { $self->_quote($_) } @$tab;
191 return $self->_quote($tab);
202 return $self->{quote_char} . $label . $self->{quote_char}
203 if !defined $self->{name_sep};
205 return join $self->{name_sep},
206 map { $self->{quote_char} . $_ . $self->{quote_char} }
207 split /\Q$self->{name_sep}\E/, $label;
210 # Conversion, if applicable
213 return @_ unless $self->{convert};
214 my $conv = $self->_sqlcase($self->{convert});
215 my @ret = map { $conv.'('.$_.')' } @_;
216 return wantarray ? @ret : $ret[0];
223 return $self->{bindtype} eq 'columns' ? [ @_ ] : @val;
226 # Modified -logic or -nest
229 my $sym = @_ ? lc(shift) : $self->{logic};
231 $sym = $self->{logic} if $sym eq 'nest';
232 return $self->_sqlcase($sym); # override join
235 =head2 new(option => 'value')
237 The C<new()> function takes a list of options and values, and returns
238 a new B<SQL::Abstract> object which can then be used to generate SQL
239 through the methods below. The options accepted are:
245 If set to 'lower', then SQL will be generated in all lowercase. By
246 default SQL is generated in "textbook" case meaning something like:
248 SELECT a_field FROM a_table WHERE some_field LIKE '%someval%'
252 This determines what the default comparison operator is. By default
253 it is C<=>, meaning that a hash like this:
255 %where = (name => 'nwiger', email => 'nate@wiger.org');
257 Will generate SQL like this:
259 WHERE name = 'nwiger' AND email = 'nate@wiger.org'
261 However, you may want loose comparisons by default, so if you set
262 C<cmp> to C<like> you would get SQL such as:
264 WHERE name like 'nwiger' AND email like 'nate@wiger.org'
266 You can also override the comparsion on an individual basis - see
267 the huge section on L</"WHERE CLAUSES"> at the bottom.
271 This determines the default logical operator for multiple WHERE
272 statements in arrays. By default it is "or", meaning that a WHERE
276 event_date => {'>=', '2/13/99'},
277 event_date => {'<=', '4/24/03'},
280 Will generate SQL like this:
282 WHERE event_date >= '2/13/99' OR event_date <= '4/24/03'
284 This is probably not what you want given this query, though (look
285 at the dates). To change the "OR" to an "AND", simply specify:
287 my $sql = SQL::Abstract->new(logic => 'and');
289 Which will change the above C<WHERE> to:
291 WHERE event_date >= '2/13/99' AND event_date <= '4/24/03'
295 This will automatically convert comparisons using the specified SQL
296 function for both column and value. This is mostly used with an argument
297 of C<upper> or C<lower>, so that the SQL will have the effect of
298 case-insensitive "searches". For example, this:
300 $sql = SQL::Abstract->new(convert => 'upper');
301 %where = (keywords => 'MaKe iT CAse inSeNSItive');
303 Will turn out the following SQL:
305 WHERE upper(keywords) like upper('MaKe iT CAse inSeNSItive')
307 The conversion can be C<upper()>, C<lower()>, or any other SQL function
308 that can be applied symmetrically to fields (actually B<SQL::Abstract> does
309 not validate this option; it will just pass through what you specify verbatim).
313 This is a kludge because many databases suck. For example, you can't
314 just bind values using DBI's C<execute()> for Oracle C<CLOB> or C<BLOB> fields.
315 Instead, you have to use C<bind_param()>:
317 $sth->bind_param(1, 'reg data');
318 $sth->bind_param(2, $lots, {ora_type => ORA_CLOB});
320 The problem is, B<SQL::Abstract> will normally just return a C<@bind> array,
321 which loses track of which field each slot refers to. Fear not.
323 If you specify C<bindtype> in new, you can determine how C<@bind> is returned.
324 Currently, you can specify either C<normal> (default) or C<columns>. If you
325 specify C<columns>, you will get an array that looks like this:
327 my $sql = SQL::Abstract->new(bindtype => 'columns');
328 my($stmt, @bind) = $sql->insert(...);
331 [ 'column1', 'value1' ],
332 [ 'column2', 'value2' ],
333 [ 'column3', 'value3' ],
336 You can then iterate through this manually, using DBI's C<bind_param()>.
338 $sth->prepare($stmt);
341 my($col, $data) = @$_;
342 if ($col eq 'details' || $col eq 'comments') {
343 $sth->bind_param($i, $data, {ora_type => ORA_CLOB});
344 } elsif ($col eq 'image') {
345 $sth->bind_param($i, $data, {ora_type => ORA_BLOB});
347 $sth->bind_param($i, $data);
351 $sth->execute; # execute without @bind now
353 Now, why would you still use B<SQL::Abstract> if you have to do this crap?
354 Basically, the advantage is still that you don't have to care which fields
355 are or are not included. You could wrap that above C<for> loop in a simple
356 sub called C<bind_fields()> or something and reuse it repeatedly. You still
357 get a layer of abstraction over manual SQL specification.
361 This is the character that a table or column name will be quoted
362 with. By default this is an empty string, but you could set it to
363 the character C<`>, to generate SQL like this:
365 SELECT `a_field` FROM `a_table` WHERE `some_field` LIKE '%someval%'
367 This is useful if you have tables or columns that are reserved words
368 in your database's SQL dialect.
372 This is the character that separates a table and column name. It is
373 necessary to specify this when the C<quote_char> option is selected,
374 so that tables and column names can be individually quoted like this:
376 SELECT `table`.`one_field` FROM `table` WHERE `table`.`other_field` = 1
384 my $class = ref($self) || $self;
385 my %opt = (ref $_[0] eq 'HASH') ? %{$_[0]} : @_;
387 # choose our case by keeping an option around
388 delete $opt{case} if $opt{case} && $opt{case} ne 'lower';
390 # override logical operator
391 $opt{logic} = uc $opt{logic} if $opt{logic};
393 # how to return bind vars
394 $opt{bindtype} ||= delete($opt{bind_type}) || 'normal';
396 # default comparison is "=", but can be overridden
399 # default quotation character around tables/columns
400 $opt{quote_char} ||= '';
402 return bless \%opt, $class;
405 =head2 insert($table, \@values || \%fieldvals)
407 This is the simplest function. You simply give it a table name
408 and either an arrayref of values or hashref of field/value pairs.
409 It returns an SQL INSERT statement and a list of bind values.
415 my $table = $self->_table(shift);
416 my $data = shift || return;
418 my $sql = $self->_sqlcase('insert into') . " $table ";
419 my(@sqlf, @sqlv, @sqlq) = ();
422 if ($ref eq 'HASH') {
423 for my $k (sort keys %$data) {
426 # named fields, so must save names in order
427 push @sqlf, $self->_quote($k);
429 # SQL included for values
431 push @sqlq, shift @val;
432 push @sqlv, $self->_bindtype($k, @val);
433 } elsif ($r eq 'SCALAR') {
434 # embedded literal SQL
438 push @sqlv, $self->_bindtype($k, $v);
441 $sql .= '(' . join(', ', @sqlf) .') '. $self->_sqlcase('values') . ' ('. join(', ', @sqlq) .')';
442 } elsif ($ref eq 'ARRAY') {
443 # just generate values(?,?) part
444 # no names (arrayref) so can't generate bindtype
445 carp "Warning: ",__PACKAGE__,"->insert called with arrayref when bindtype set"
446 if $self->{bindtype} ne 'normal';
451 push @sqlq, shift @val;
453 } elsif ($r eq 'SCALAR') {
454 # embedded literal SQL
461 $sql .= $self->_sqlcase('values') . ' ('. join(', ', @sqlq) .')';
462 } elsif ($ref eq 'SCALAR') {
466 puke "Unsupported data type specified to \$sql->insert";
469 return wantarray ? ($sql, @sqlv) : $sql;
472 =head2 update($table, \%fieldvals, \%where)
474 This takes a table, hashref of field/value pairs, and an optional
475 hashref WHERE clause. It returns an SQL UPDATE function and a list
482 my $table = $self->_table(shift);
483 my $data = shift || return;
486 my $sql = $self->_sqlcase('update') . " $table " . $self->_sqlcase('set ');
487 my(@sqlf, @sqlv) = ();
489 puke "Unsupported data type specified to \$sql->update"
490 unless ref $data eq 'HASH';
492 for my $k (sort keys %$data) {
495 my $label = $self->_quote($k);
497 # SQL included for values
499 my $sql = shift @bind;
500 push @sqlf, "$label = $sql";
501 push @sqlv, $self->_bindtype($k, @bind);
502 } elsif ($r eq 'SCALAR') {
503 # embedded literal SQL
504 push @sqlf, "$label = $$v";
506 push @sqlf, "$label = ?";
507 push @sqlv, $self->_bindtype($k, $v);
511 $sql .= join ', ', @sqlf;
514 my($wsql, @wval) = $self->where($where);
519 return wantarray ? ($sql, @sqlv) : $sql;
522 =head2 select($table, \@fields, \%where, \@order)
524 This takes a table, arrayref of fields (or '*'), optional hashref
525 WHERE clause, and optional arrayref order by, and returns the
526 corresponding SQL SELECT statement and list of bind values.
532 my $table = $self->_table(shift);
533 my $fields = shift || '*';
537 my $f = (ref $fields eq 'ARRAY') ? join ', ', map { $self->_quote($_) } @$fields : $fields;
538 my $sql = join ' ', $self->_sqlcase('select'), $f, $self->_sqlcase('from'), $table;
540 my(@sqlf, @sqlv) = ();
541 my($wsql, @wval) = $self->where($where, $order);
545 return wantarray ? ($sql, @sqlv) : $sql;
548 =head2 delete($table, \%where)
550 This takes a table name and optional hashref WHERE clause.
551 It returns an SQL DELETE statement and list of bind values.
557 my $table = $self->_table(shift);
560 my $sql = $self->_sqlcase('delete from') . " $table";
561 my(@sqlf, @sqlv) = ();
564 my($wsql, @wval) = $self->where($where);
569 return wantarray ? ($sql, @sqlv) : $sql;
572 =head2 where(\%where, \@order)
574 This is used to generate just the WHERE clause. For example,
575 if you have an arbitrary data structure and know what the
576 rest of your SQL is going to look like, but want an easy way
577 to produce a WHERE clause, use this. It returns an SQL WHERE
578 clause and list of bind values.
582 # Finally, a separate routine just to handle WHERE clauses
588 # Need a separate routine to properly wrap w/ "where"
590 my @ret = $self->_recurse_where($where);
593 $sql .= $self->_sqlcase(' where ') . $wh if $wh;
598 $sql .= $self->_order_by($order);
601 return wantarray ? ($sql, @ret) : $sql;
606 local $^W = 0; # really, you've gotta be fucking kidding me
608 my $where = _anoncopy(shift); # prevent destroying original
609 my $ref = ref $where || '';
610 my $join = shift || $self->{logic} ||
611 ($ref eq 'ARRAY' ? $self->_sqlcase('or') : $self->_sqlcase('and'));
613 # For assembling SQL fields and values
614 my(@sqlf, @sqlv) = ();
616 # If an arrayref, then we join each element
617 if ($ref eq 'ARRAY') {
618 # need to use while() so can shift() for arrays
620 while (my $el = shift @$where) {
622 # skip empty elements, otherwise get invalid trailing AND stuff
623 if (my $ref2 = ref $el) {
624 if ($ref2 eq 'ARRAY') {
626 } elsif ($ref2 eq 'HASH') {
628 $subjoin ||= $self->_sqlcase('and');
629 } elsif ($ref2 eq 'SCALAR') {
634 $self->_debug("$ref2(*top) means join with $subjoin");
636 # top-level arrayref with scalars, recurse in pairs
637 $self->_debug("NOREF(*top) means join with $subjoin");
638 $el = {$el => shift(@$where)};
640 my @ret = $self->_recurse_where($el, $subjoin);
641 push @sqlf, shift @ret;
645 elsif ($ref eq 'HASH') {
646 # Note: during recursion, the last element will always be a hashref,
647 # since it needs to point a column => value. So this be the end.
648 for my $k (sort keys %$where) {
649 my $v = $where->{$k};
650 my $label = $self->_quote($k);
651 if ($k =~ /^-(\D+)/) {
652 # special nesting, like -and, -or, -nest, so shift over
653 my $subjoin = $self->_modlogic($1);
654 $self->_debug("OP(-$1) means special logic ($subjoin), recursing...");
655 my @ret = $self->_recurse_where($v, $subjoin);
656 push @sqlf, shift @ret;
658 } elsif (! defined($v)) {
660 $self->_debug("UNDEF($k) means IS NULL");
661 push @sqlf, $label . $self->_sqlcase(' is null');
662 } elsif (ref $v eq 'ARRAY') {
665 # multiple elements: multiple options
666 $self->_debug("ARRAY($k) means multiple elements: [ @v ]");
668 # special nesting, like -and, -or, -nest, so shift over
669 my $subjoin = $self->_sqlcase('or');
670 if ($v[0] =~ /^-(\D+)/) {
671 $subjoin = $self->_modlogic($1); # override subjoin
672 $self->_debug("OP(-$1) means special logic ($subjoin), shifting...");
676 # map into an array of hashrefs and recurse
677 my @ret = $self->_recurse_where([map { {$k => $_} } @v], $subjoin);
679 # push results into our structure
680 push @sqlf, shift @ret;
682 } elsif (ref $v eq 'HASH') {
683 # modified operator { '!=', 'completed' }
684 for my $f (sort keys %$v) {
686 $self->_debug("HASH($k) means modified operator: { $f }");
688 # check for the operator being "IN" or "BETWEEN" or whatever
689 if (ref $x eq 'ARRAY') {
690 if ($f =~ /^-?\s*(not[\s_]+)?(in|between)\s*$/i) {
691 my $u = $self->_modlogic($1 . $2);
692 $self->_debug("HASH($f => $x) uses special operator: [ $u ]");
693 if ($u =~ /between/i) {
695 push @sqlf, join ' ', $self->_convert($label), $u, $self->_convert('?'),
696 $self->_sqlcase('and'), $self->_convert('?');
698 push @sqlf, join ' ', $self->_convert($label), $u, '(',
699 join(', ', map { $self->_convert('?') } @$x),
702 push @sqlv, $self->_bindtype($k, @$x);
704 # multiple elements: multiple options
705 $self->_debug("ARRAY($x) means multiple elements: [ @$x ]");
707 # map into an array of hashrefs and recurse
708 my @ret = $self->_recurse_where([map { {$k => {$f, $_}} } @$x]);
710 # push results into our structure
711 push @sqlf, shift @ret;
714 } elsif (! defined($x)) {
716 my $not = ($f eq '!=' || $f eq 'not like') ? ' not' : '';
717 push @sqlf, $label . $self->_sqlcase(" is$not null");
720 $f =~ s/^-//; # strip leading -like =>
721 $f =~ s/_/ /; # _ => " "
722 push @sqlf, join ' ', $self->_convert($label), $self->_sqlcase($f), $self->_convert('?');
723 push @sqlv, $self->_bindtype($k, $x);
726 } elsif (ref $v eq 'SCALAR') {
728 $self->_debug("SCALAR($k) means literal SQL: $$v");
729 push @sqlf, "$label $$v";
731 # standard key => val
732 $self->_debug("NOREF($k) means simple key=val: $k $self->{cmp} $v");
733 push @sqlf, join ' ', $self->_convert($label), $self->_sqlcase($self->{cmp}), $self->_convert('?');
734 push @sqlv, $self->_bindtype($k, $v);
738 elsif ($ref eq 'SCALAR') {
740 $self->_debug("SCALAR(*top) means literal SQL: $$where");
743 elsif (defined $where) {
745 $self->_debug("NOREF(*top) means literal SQL: $where");
749 # assemble and return sql
750 my $wsql = @sqlf ? '( ' . join(" $join ", @sqlf) . ' )' : '';
751 return wantarray ? ($wsql, @sqlv) : $wsql;
758 my @vals = $ref eq 'ARRAY' ? @{$_[0]} :
759 $ref eq 'SCALAR' ? ${$_[0]} :
761 puke "Unsupported data struct $ref for ORDER BY";
763 my $val = join ', ', map { $self->_quote($_) } @vals;
764 return $val ? $self->_sqlcase(' order by')." $val" : '';
767 =head2 values(\%data)
769 This just returns the values from the hash C<%data>, in the same
770 order that would be returned from any of the other above queries.
771 Using this allows you to markedly speed up your queries if you
772 are affecting lots of rows. See below under the L</"PERFORMANCE"> section.
778 my $data = shift || return;
779 puke "Argument to ", __PACKAGE__, "->values must be a \\%hash"
780 unless ref $data eq 'HASH';
781 return map { $self->_bindtype($_, $data->{$_}) } sort keys %$data;
784 =head2 generate($any, 'number', $of, \@data, $struct, \%types)
786 Warning: This is an experimental method and subject to change.
788 This returns arbitrarily generated SQL. It's a really basic shortcut.
789 It will return two different things, depending on return context:
791 my($stmt, @bind) = $sql->generate('create table', \$table, \@fields);
792 my $stmt_and_val = $sql->generate('create table', \$table, \@fields);
794 These would return the following:
797 $stmt = "CREATE TABLE test (?, ?)";
798 @bind = (field1, field2);
800 # Second calling form
801 $stmt_and_val = "CREATE TABLE test (field1, field2)";
803 Depending on what you're trying to do, it's up to you to choose the correct
804 format. In this example, the second form is what you would want.
808 $sql->generate('alter session', { nls_date_format => 'MM/YY' });
812 ALTER SESSION SET nls_date_format = 'MM/YY'
814 You get the idea. Strings get their case twiddled, but everything
815 else remains verbatim.
822 my(@sql, @sqlq, @sqlv);
826 if ($ref eq 'HASH') {
827 for my $k (sort keys %$_) {
830 my $label = $self->_quote($k);
832 # SQL included for values
834 my $sql = shift @bind;
835 push @sqlq, "$label = $sql";
836 push @sqlv, $self->_bindtype($k, @bind);
837 } elsif ($r eq 'SCALAR') {
838 # embedded literal SQL
839 push @sqlq, "$label = $$v";
841 push @sqlq, "$label = ?";
842 push @sqlv, $self->_bindtype($k, $v);
845 push @sql, $self->_sqlcase('set'), join ', ', @sqlq;
846 } elsif ($ref eq 'ARRAY') {
847 # unlike insert(), assume these are ONLY the column names, i.e. for SQL
852 push @sqlq, shift @val;
854 } elsif ($r eq 'SCALAR') {
855 # embedded literal SQL
862 push @sql, '(' . join(', ', @sqlq) . ')';
863 } elsif ($ref eq 'SCALAR') {
867 # strings get case twiddled
868 push @sql, $self->_sqlcase($_);
872 my $sql = join ' ', @sql;
874 # this is pretty tricky
875 # if ask for an array, return ($stmt, @bind)
876 # otherwise, s/?/shift @sqlv/ to put it inline
878 return ($sql, @sqlv);
880 1 while $sql =~ s/\?/my $d = shift(@sqlv);
881 ref $d ? $d->[1] : $d/e;
888 # This allows us to check for a local, then _form, attr
890 my($name) = $AUTOLOAD =~ /.*::(.+)/;
891 return $self->generate($name, @_);
900 This module uses a variation on the idea from L<DBIx::Abstract>. It
901 is B<NOT>, repeat I<not> 100% compatible. B<The main logic of this
902 module is that things in arrays are OR'ed, and things in hashes
905 The easiest way to explain is to show lots of examples. After
906 each C<%where> hash shown, it is assumed you used:
908 my($stmt, @bind) = $sql->where(\%where);
910 However, note that the C<%where> hash can be used directly in any
911 of the other functions as well, as described above.
913 So, let's get started. To begin, a simple hash:
917 status => 'completed'
920 Is converted to SQL C<key = val> statements:
922 $stmt = "WHERE user = ? AND status = ?";
923 @bind = ('nwiger', 'completed');
925 One common thing I end up doing is having a list of values that
926 a field can be in. To do this, simply specify a list inside of
931 status => ['assigned', 'in-progress', 'pending'];
934 This simple code will create the following:
936 $stmt = "WHERE user = ? AND ( status = ? OR status = ? OR status = ? )";
937 @bind = ('nwiger', 'assigned', 'in-progress', 'pending');
939 If you want to specify a different type of operator for your comparison,
940 you can use a hashref for a given column:
944 status => { '!=', 'completed' }
947 Which would generate:
949 $stmt = "WHERE user = ? AND status != ?";
950 @bind = ('nwiger', 'completed');
952 To test against multiple values, just enclose the values in an arrayref:
954 status => { '!=', ['assigned', 'in-progress', 'pending'] };
956 Which would give you:
958 "WHERE status != ? OR status != ? OR status != ?"
960 But, this is probably not what you want in this case (look at it). So
961 the hashref can also contain multiple pairs, in which case it is expanded
962 into an C<AND> of its elements:
966 status => { '!=', 'completed', -not_like => 'pending%' }
969 # Or more dynamically, like from a form
970 $where{user} = 'nwiger';
971 $where{status}{'!='} = 'completed';
972 $where{status}{'-not_like'} = 'pending%';
975 $stmt = "WHERE user = ? AND status != ? AND status NOT LIKE ?";
976 @bind = ('nwiger', 'completed', 'pending%');
978 To get an OR instead, you can combine it with the arrayref idea:
982 priority => [ {'=', 2}, {'!=', 1} ]
985 Which would generate:
987 $stmt = "WHERE user = ? AND priority = ? OR priority != ?";
988 @bind = ('nwiger', '2', '1');
990 However, there is a subtle trap if you want to say something like
991 this (notice the C<AND>):
993 WHERE priority != ? AND priority != ?
995 Because, in Perl you I<can't> do this:
997 priority => { '!=', 2, '!=', 1 }
999 As the second C<!=> key will obliterate the first. The solution
1000 is to use the special C<-modifier> form inside an arrayref:
1002 priority => [ -and => {'!=', 2}, {'!=', 1} ]
1004 Normally, these would be joined by C<OR>, but the modifier tells it
1005 to use C<AND> instead. (Hint: You can use this in conjunction with the
1006 C<logic> option to C<new()> in order to change the way your queries
1007 work by default.) B<Important:> Note that the C<-modifier> goes
1008 B<INSIDE> the arrayref, as an extra first element. This will
1009 B<NOT> do what you think it might:
1011 priority => -and => [{'!=', 2}, {'!=', 1}] # WRONG!
1013 Here is a quick list of equivalencies, since there is some overlap:
1016 status => {'!=', 'completed', 'not like', 'pending%' }
1017 status => [ -and => {'!=', 'completed'}, {'not like', 'pending%'}]
1020 status => {'=', ['assigned', 'in-progress']}
1021 status => [ -or => {'=', 'assigned'}, {'=', 'in-progress'}]
1022 status => [ {'=', 'assigned'}, {'=', 'in-progress'} ]
1024 In addition to C<-and> and C<-or>, there is also a special C<-nest>
1025 operator which adds an additional set of parens, to create a subquery.
1026 For example, to get something like this:
1028 $stmt = WHERE user = ? AND ( workhrs > ? OR geo = ? )
1029 @bind = ('nwiger', '20', 'ASIA');
1035 -nest => [ workhrs => {'>', 20}, geo => 'ASIA' ],
1038 You can also use the hashref format to compare a list of fields using the
1039 C<IN> comparison operator, by specifying the list as an arrayref:
1042 status => 'completed',
1043 reportid => { -in => [567, 2335, 2] }
1046 Which would generate:
1048 $stmt = "WHERE status = ? AND reportid IN (?,?,?)";
1049 @bind = ('completed', '567', '2335', '2');
1051 You can use this same format to use other grouping functions, such
1052 as C<BETWEEN>, C<SOME>, and so forth. For example:
1056 completion_date => {
1057 -not_between => ['2002-10-01', '2003-02-06']
1063 WHERE user = ? AND completion_date NOT BETWEEN ( ? AND ? )
1065 So far, we've seen how multiple conditions are joined with a top-level
1066 C<AND>. We can change this by putting the different conditions we want in
1067 hashes and then putting those hashes in an array. For example:
1072 status => { -like => ['pending%', 'dispatched'] },
1076 status => 'unassigned',
1080 This data structure would create the following:
1082 $stmt = "WHERE ( user = ? AND ( status LIKE ? OR status LIKE ? ) )
1083 OR ( user = ? AND status = ? ) )";
1084 @bind = ('nwiger', 'pending', 'dispatched', 'robot', 'unassigned');
1086 This can be combined with the C<-nest> operator to properly group
1093 -and => [workhrs => {'>', 20}, geo => 'ASIA' ],
1094 -and => [workhrs => {'<', 50}, geo => 'EURO' ]
1101 WHERE ( user = ? AND
1102 ( ( workhrs > ? AND geo = ? )
1103 OR ( workhrs < ? AND geo = ? ) ) )
1105 Finally, sometimes only literal SQL will do. If you want to include
1106 literal SQL verbatim, you can specify it as a scalar reference, namely:
1108 my $inn = 'is Not Null';
1110 priority => { '<', 2 },
1116 $stmt = "WHERE priority < ? AND requestor is Not Null";
1119 Note that in this example, you only get one bind parameter back, since
1120 the verbatim SQL is passed as part of the statement.
1122 Of course, just to prove a point, the above can also be accomplished
1126 priority => { '<', 2 },
1127 requestor => { '!=', undef },
1132 These pages could go on for a while, since the nesting of the data
1133 structures this module can handle are pretty much unlimited (the
1134 module implements the C<WHERE> expansion as a recursive function
1135 internally). Your best bet is to "play around" with the module a
1136 little to see how the data structures behave, and choose the best
1137 format for your data based on that.
1139 And of course, all the values above will probably be replaced with
1140 variables gotten from forms or the command line. After all, if you
1141 knew everything ahead of time, you wouldn't have to worry about
1142 dynamically-generating SQL and could just hardwire it into your
1147 Thanks to some benchmarking by Mark Stosberg, it turns out that
1148 this module is many orders of magnitude faster than using C<DBIx::Abstract>.
1149 I must admit this wasn't an intentional design issue, but it's a
1150 byproduct of the fact that you get to control your C<DBI> handles
1153 To maximize performance, use a code snippet like the following:
1155 # prepare a statement handle using the first row
1156 # and then reuse it for the rest of the rows
1158 for my $href (@array_of_hashrefs) {
1159 $stmt ||= $sql->insert('table', $href);
1160 $sth ||= $dbh->prepare($stmt);
1161 $sth->execute($sql->values($href));
1164 The reason this works is because the keys in your C<$href> are sorted
1165 internally by B<SQL::Abstract>. Thus, as long as your data retains
1166 the same structure, you only have to generate the SQL the first time
1167 around. On subsequent queries, simply use the C<values> function provided
1168 by this module to return your values in the correct order.
1172 If you use my C<CGI::FormBuilder> module at all, you'll hopefully
1173 really like this part (I do, at least). Building up a complex query
1174 can be as simple as the following:
1178 use CGI::FormBuilder;
1181 my $form = CGI::FormBuilder->new(...);
1182 my $sql = SQL::Abstract->new;
1184 if ($form->submitted) {
1185 my $field = $form->field;
1186 my $id = delete $field->{id};
1187 my($stmt, @bind) = $sql->update('table', $field, {id => $id});
1190 Of course, you would still have to connect using C<DBI> to run the
1191 query, but the point is that if you make your form look like your
1192 table, the actual query script can be extremely simplistic.
1194 If you're B<REALLY> lazy (I am), check out C<HTML::QuickTable> for
1195 a fast interface to returning and formatting data. I frequently
1196 use these three modules together to write complex database query
1197 apps in under 50 lines.
1201 There is not (yet) any explicit support for SQL compound logic
1202 statements like "AND NOT". Instead, just do the de Morgan's
1203 law transformations yourself. For example, this:
1205 "lname LIKE '%son%' AND NOT ( age < 10 OR age > 20 )"
1209 "lname LIKE '%son%' AND ( age >= 10 AND age <= 20 )"
1211 With the corresponding C<%where> hash:
1214 lname => {like => '%son%'},
1215 age => [-and => {'>=', 10}, {'<=', 20}],
1218 Again, remember that the C<-and> goes I<inside> the arrayref.
1220 =head1 ACKNOWLEDGEMENTS
1222 There are a number of individuals that have really helped out with
1223 this module. Unfortunately, most of them submitted bugs via CPAN
1224 so I have no idea who they are! But the people I do know are:
1226 Mark Stosberg (benchmarking)
1227 Chas Owens (initial "IN" operator support)
1228 Philip Collins (per-field SQL functions)
1229 Eric Kolve (hashref "AND" support)
1230 Mike Fragassi (enhancements to "BETWEEN" and "LIKE")
1231 Dan Kubb (support for "quote_char" and "name_sep")
1237 If found, please DO NOT submit anything via C<rt.cpan.org> - that
1238 just causes me a ton of work. Email me a patch (or script demonstrating
1239 the problem) to the below address, and include the VERSION string you'll
1244 L<DBIx::Abstract>, L<DBI|DBI>, L<CGI::FormBuilder>, L<HTML::QuickTable>
1252 Copyright (c) 2001-2006 Nathan Wiger <nate@wiger.org>. All Rights Reserved.
1254 This module is free software; you may copy this under the terms of
1255 the GNU General Public License, or the Artistic License, copies of
1256 which should have accompanied your Perl kit.