1 package DBIx::Class::Storage::DBI;
2 # -*- mode: cperl; cperl-indent-level: 2 -*-
7 use base 'DBIx::Class::Storage';
10 use Carp::Clan qw/^DBIx::Class/;
12 use DBIx::Class::Storage::DBI::Cursor;
13 use DBIx::Class::Storage::Statistics;
17 __PACKAGE__->mk_group_accessors('simple' =>
18 qw/_connect_info _dbi_connect_info _dbh _sql_maker _sql_maker_opts _conn_pid
19 _conn_tid transaction_depth _dbh_autocommit _driver_determined savepoints/
22 # the values for these accessors are picked out (and deleted) from
23 # the attribute hashref passed to connect_info
24 my @storage_options = qw/
25 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
26 disable_sth_caching unsafe auto_savepoint
28 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
31 # default cursor class, overridable in connect_info attributes
32 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
34 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class/);
35 __PACKAGE__->sql_maker_class('DBIx::Class::SQLAHacks');
40 DBIx::Class::Storage::DBI - DBI storage handler
44 my $schema = MySchema->connect('dbi:SQLite:my.db');
46 $schema->storage->debug(1);
47 $schema->dbh_do("DROP TABLE authors");
49 $schema->resultset('Book')->search({
50 written_on => $schema->storage->datetime_parser(DateTime->now)
55 This class represents the connection to an RDBMS via L<DBI>. See
56 L<DBIx::Class::Storage> for general information. This pod only
57 documents DBI-specific methods and behaviors.
64 my $new = shift->next::method(@_);
66 $new->transaction_depth(0);
67 $new->_sql_maker_opts({});
68 $new->{savepoints} = [];
69 $new->{_in_dbh_do} = 0;
77 This method is normally called by L<DBIx::Class::Schema/connection>, which
78 encapsulates its argument list in an arrayref before passing them here.
80 The argument list may contain:
86 The same 4-element argument set one would normally pass to
87 L<DBI/connect>, optionally followed by
88 L<extra attributes|/DBIx::Class specific connection attributes>
89 recognized by DBIx::Class:
91 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
95 A single code reference which returns a connected
96 L<DBI database handle|DBI/connect> optionally followed by
97 L<extra attributes|/DBIx::Class specific connection attributes> recognized
100 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
104 A single hashref with all the attributes and the dsn/user/password
107 $connect_info_args = [{
115 This is particularly useful for L<Catalyst> based applications, allowing the
116 following config (L<Config::General> style):
121 dsn dbi:mysql:database=test
128 The C<dsn>/C<user>/C<password> combination can be substituted by the
129 C<dbh_maker> key whose value is a coderef that returns the C<$dbh>.
133 Please note that the L<DBI> docs recommend that you always explicitly
134 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
135 recommends that it be set to I<1>, and that you perform transactions
136 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
137 to I<1> if you do not do explicitly set it to zero. This is the default
138 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
140 =head3 DBIx::Class specific connection attributes
142 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
143 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
144 the following connection options. These options can be mixed in with your other
145 L<DBI> connection attributes, or placed in a seperate hashref
146 (C<\%extra_attributes>) as shown above.
148 Every time C<connect_info> is invoked, any previous settings for
149 these options will be cleared before setting the new ones, regardless of
150 whether any options are specified in the new C<connect_info>.
157 Specifies things to do immediately after connecting or re-connecting to
158 the database. Its value may contain:
164 This contains one SQL statement to execute.
166 =item an array reference
168 This contains SQL statements to execute in order. Each element contains
169 a string or a code reference that returns a string.
171 =item a code reference
173 This contains some code to execute. Unlike code references within an
174 array reference, its return value is ignored.
178 =item on_disconnect_do
180 Takes arguments in the same form as L</on_connect_do> and executes them
181 immediately before disconnecting from the database.
183 Note, this only runs if you explicitly call L</disconnect> on the
186 =item on_connect_call
188 A more generalized form of L</on_connect_do> that calls the specified
189 C<connect_call_METHOD> methods in your storage driver.
191 on_connect_do => 'select 1'
195 on_connect_call => [ [ do_sql => 'select 1' ] ]
197 Its values may contain:
203 Will call the C<connect_call_METHOD> method.
205 =item a code reference
207 Will execute C<< $code->($storage) >>
209 =item an array reference
211 Each value can be a method name or code reference.
213 =item an array of arrays
215 For each array, the first item is taken to be the C<connect_call_> method name
216 or code reference, and the rest are parameters to it.
220 Some predefined storage methods you may use:
226 Executes a SQL string or a code reference that returns a SQL string. This is
227 what L</on_connect_do> and L</on_disconnect_do> use.
235 Will execute the scalar as SQL.
239 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
240 attributes hashref and bind values.
242 =item a code reference
244 Will execute C<< $code->($storage) >> and execute the return array refs as
251 Execute any statements necessary to initialize the database session to return
252 and accept datetime/timestamp values used with
253 L<DBIx::Class::InflateColumn::DateTime>.
255 Only necessary for some databases, see your specific storage driver for
256 implementation details.
260 =item on_disconnect_call
262 Takes arguments in the same form as L</on_connect_call> and executes them
263 immediately before disconnecting from the database.
265 Calls the C<disconnect_call_METHOD> methods as opposed to the
266 C<connect_call_METHOD> methods called by L</on_connect_call>.
268 Note, this only runs if you explicitly call L</disconnect> on the
271 =item disable_sth_caching
273 If set to a true value, this option will disable the caching of
274 statement handles via L<DBI/prepare_cached>.
278 Sets the limit dialect. This is useful for JDBC-bridge among others
279 where the remote SQL-dialect cannot be determined by the name of the
280 driver alone. See also L<SQL::Abstract::Limit>.
284 Specifies what characters to use to quote table and column names. If
285 you use this you will want to specify L</name_sep> as well.
287 C<quote_char> expects either a single character, in which case is it
288 is placed on either side of the table/column name, or an arrayref of length
289 2 in which case the table/column name is placed between the elements.
291 For example under MySQL you should use C<< quote_char => '`' >>, and for
292 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
296 This only needs to be used in conjunction with C<quote_char>, and is used to
297 specify the charecter that seperates elements (schemas, tables, columns) from
298 each other. In most cases this is simply a C<.>.
300 The consequences of not supplying this value is that L<SQL::Abstract>
301 will assume DBIx::Class' uses of aliases to be complete column
302 names. The output will look like I<"me.name"> when it should actually
307 This Storage driver normally installs its own C<HandleError>, sets
308 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
309 all database handles, including those supplied by a coderef. It does this
310 so that it can have consistent and useful error behavior.
312 If you set this option to a true value, Storage will not do its usual
313 modifications to the database handle's attributes, and instead relies on
314 the settings in your connect_info DBI options (or the values you set in
315 your connection coderef, in the case that you are connecting via coderef).
317 Note that your custom settings can cause Storage to malfunction,
318 especially if you set a C<HandleError> handler that suppresses exceptions
319 and/or disable C<RaiseError>.
323 If this option is true, L<DBIx::Class> will use savepoints when nesting
324 transactions, making it possible to recover from failure in the inner
325 transaction without having to abort all outer transactions.
329 Use this argument to supply a cursor class other than the default
330 L<DBIx::Class::Storage::DBI::Cursor>.
334 Some real-life examples of arguments to L</connect_info> and
335 L<DBIx::Class::Schema/connect>
337 # Simple SQLite connection
338 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
341 ->connect_info([ sub { DBI->connect(...) } ]);
343 # Connect via subref in hashref
345 dbh_maker => sub { DBI->connect(...) },
346 on_connect_do => 'alter session ...',
349 # A bit more complicated
356 { quote_char => q{"}, name_sep => q{.} },
360 # Equivalent to the previous example
366 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
370 # Same, but with hashref as argument
371 # See parse_connect_info for explanation
374 dsn => 'dbi:Pg:dbname=foo',
376 password => 'my_pg_password',
383 # Subref + DBIx::Class-specific connection options
386 sub { DBI->connect(...) },
390 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
391 disable_sth_caching => 1,
401 my ($self, $info_arg) = @_;
403 return $self->_connect_info if !$info_arg;
405 my @args = @$info_arg; # take a shallow copy for further mutilation
406 $self->_connect_info([@args]); # copy for _connect_info
409 # combine/pre-parse arguments depending on invocation style
412 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
413 %attrs = %{ $args[1] || {} };
416 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
417 %attrs = %{$args[0]};
419 if (my $code = delete $attrs{dbh_maker}) {
421 if (delete @attrs{qw/dsn user password/}) {
422 warn 'dsn/user/password ignored when dbh_maker coderef used in ' .
427 @args = delete @attrs{qw/dsn user password/};
430 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
432 % { $args[3] || {} },
433 % { $args[4] || {} },
435 @args = @args[0,1,2];
438 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
439 # the new set of options
440 $self->_sql_maker(undef);
441 $self->_sql_maker_opts({});
444 for my $storage_opt (@storage_options, 'cursor_class') { # @storage_options is declared at the top of the module
445 if(my $value = delete $attrs{$storage_opt}) {
446 $self->$storage_opt($value);
449 for my $sql_maker_opt (qw/limit_dialect quote_char name_sep/) {
450 if(my $opt_val = delete $attrs{$sql_maker_opt}) {
451 $self->_sql_maker_opts->{$sql_maker_opt} = $opt_val;
456 if (ref $args[0] eq 'CODE') {
457 # _connect() never looks past $args[0] in this case
461 %{ $self->_default_dbi_connect_attributes || {} },
466 $self->_dbi_connect_info([@args, keys %attrs ? \%attrs : ()]);
467 $self->_connect_info;
470 sub _default_dbi_connect_attributes {
480 This method is deprecated in favour of setting via L</connect_info>.
484 =head2 on_disconnect_do
486 This method is deprecated in favour of setting via L</connect_info>.
490 sub _parse_connect_do {
491 my ($self, $type) = @_;
493 my $val = $self->$type;
494 return () if not defined $val;
499 push @res, [ 'do_sql', $val ];
500 } elsif (ref($val) eq 'CODE') {
502 } elsif (ref($val) eq 'ARRAY') {
503 push @res, map { [ 'do_sql', $_ ] } @$val;
505 $self->throw_exception("Invalid type for $type: ".ref($val));
513 Arguments: ($subref | $method_name), @extra_coderef_args?
515 Execute the given $subref or $method_name using the new exception-based
516 connection management.
518 The first two arguments will be the storage object that C<dbh_do> was called
519 on and a database handle to use. Any additional arguments will be passed
520 verbatim to the called subref as arguments 2 and onwards.
522 Using this (instead of $self->_dbh or $self->dbh) ensures correct
523 exception handling and reconnection (or failover in future subclasses).
525 Your subref should have no side-effects outside of the database, as
526 there is the potential for your subref to be partially double-executed
527 if the database connection was stale/dysfunctional.
531 my @stuff = $schema->storage->dbh_do(
533 my ($storage, $dbh, @cols) = @_;
534 my $cols = join(q{, }, @cols);
535 $dbh->selectrow_array("SELECT $cols FROM foo");
546 my $dbh = $self->_dbh;
548 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
549 || $self->{transaction_depth};
551 local $self->{_in_dbh_do} = 1;
554 my $want_array = wantarray;
557 $self->_verify_pid if $dbh;
559 $self->_populate_dbh;
564 @result = $self->$code($dbh, @_);
566 elsif(defined $want_array) {
567 $result[0] = $self->$code($dbh, @_);
570 $self->$code($dbh, @_);
574 # ->connected might unset $@ - copy
576 if(!$exception) { return $want_array ? @result : $result[0] }
578 $self->throw_exception($exception) if $self->connected;
580 # We were not connected - reconnect and retry, but let any
581 # exception fall right through this time
582 carp "Retrying $code after catching disconnected exception: $exception"
583 if $ENV{DBIC_DBIRETRY_DEBUG};
584 $self->_populate_dbh;
585 $self->$code($self->_dbh, @_);
588 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
589 # It also informs dbh_do to bypass itself while under the direction of txn_do,
590 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
595 ref $coderef eq 'CODE' or $self->throw_exception
596 ('$coderef must be a CODE reference');
598 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
600 local $self->{_in_dbh_do} = 1;
603 my $want_array = wantarray;
608 $self->_verify_pid if $self->_dbh;
609 $self->_populate_dbh if !$self->_dbh;
613 @result = $coderef->(@_);
615 elsif(defined $want_array) {
616 $result[0] = $coderef->(@_);
624 # ->connected might unset $@ - copy
626 if(!$exception) { return $want_array ? @result : $result[0] }
628 if($tried++ || $self->connected) {
629 eval { $self->txn_rollback };
630 my $rollback_exception = $@;
631 if($rollback_exception) {
632 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
633 $self->throw_exception($exception) # propagate nested rollback
634 if $rollback_exception =~ /$exception_class/;
636 $self->throw_exception(
637 "Transaction aborted: ${exception}. "
638 . "Rollback failed: ${rollback_exception}"
641 $self->throw_exception($exception)
644 # We were not connected, and was first try - reconnect and retry
646 carp "Retrying $coderef after catching disconnected exception: $exception"
647 if $ENV{DBIC_DBIRETRY_DEBUG};
648 $self->_populate_dbh;
654 Our C<disconnect> method also performs a rollback first if the
655 database is not in C<AutoCommit> mode.
665 push @actions, ( $self->on_disconnect_call || () );
666 push @actions, $self->_parse_connect_do ('on_disconnect_do');
668 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
670 $self->_dbh->rollback unless $self->_dbh_autocommit;
671 $self->_dbh->disconnect;
677 =head2 with_deferred_fk_checks
681 =item Arguments: C<$coderef>
683 =item Return Value: The return value of $coderef
687 Storage specific method to run the code ref with FK checks deferred or
688 in MySQL's case disabled entirely.
692 # Storage subclasses should override this
693 sub with_deferred_fk_checks {
694 my ($self, $sub) = @_;
703 =item Arguments: none
705 =item Return Value: 1|0
709 Verifies that the the current database handle is active and ready to execute
710 an SQL statement (i.e. the connection did not get stale, server is still
711 answering, etc.) This method is used internally by L</dbh>.
717 return 0 unless $self->_seems_connected;
720 local $self->_dbh->{RaiseError} = 1;
725 sub _seems_connected {
728 my $dbh = $self->_dbh
731 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
738 return 0 if !$self->_dbh;
741 return $dbh->FETCH('Active');
747 my $dbh = $self->_dbh or return 0;
752 # handle pid changes correctly
753 # NOTE: assumes $self->_dbh is a valid $dbh
757 return if defined $self->_conn_pid && $self->_conn_pid == $$;
759 $self->_dbh->{InactiveDestroy} = 1;
766 sub ensure_connected {
769 unless ($self->connected) {
770 $self->_populate_dbh;
776 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
777 is guaranteed to be healthy by implicitly calling L</connected>, and if
778 necessary performing a reconnection before returning. Keep in mind that this
779 is very B<expensive> on some database engines. Consider using L<dbh_do>
787 if (not $self->_dbh) {
788 $self->_populate_dbh;
790 $self->ensure_connected;
795 # this is the internal "get dbh or connect (don't check)" method
798 $self->_populate_dbh unless $self->_dbh;
802 sub _sql_maker_args {
807 array_datatypes => 1,
808 limit_dialect => $self->_get_dbh,
809 %{$self->_sql_maker_opts}
815 unless ($self->_sql_maker) {
816 my $sql_maker_class = $self->sql_maker_class;
817 $self->ensure_class_loaded ($sql_maker_class);
818 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
820 return $self->_sql_maker;
828 my @info = @{$self->_dbi_connect_info || []};
829 $self->_dbh(undef); # in case ->connected failed we might get sent here
830 $self->_dbh($self->_connect(@info));
832 $self->_conn_pid($$);
833 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
835 $self->_determine_driver;
837 # Always set the transaction depth on connect, since
838 # there is no transaction in progress by definition
839 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
841 $self->_run_connection_actions unless $self->{_in_determine_driver};
844 sub _run_connection_actions {
848 push @actions, ( $self->on_connect_call || () );
849 push @actions, $self->_parse_connect_do ('on_connect_do');
851 $self->_do_connection_actions(connect_call_ => $_) for @actions;
854 sub _determine_driver {
857 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
858 my $started_unconnected = 0;
859 local $self->{_in_determine_driver} = 1;
861 if (ref($self) eq __PACKAGE__) {
863 if ($self->_dbh) { # we are connected
864 $driver = $self->_dbh->{Driver}{Name};
866 # try to use dsn to not require being connected, the driver may still
867 # force a connection in _rebless to determine version
868 ($driver) = $self->_dbi_connect_info->[0] =~ /dbi:([^:]+):/i;
869 $started_unconnected = 1;
872 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
873 if ($self->load_optional_class($storage_class)) {
874 mro::set_mro($storage_class, 'c3');
875 bless $self, $storage_class;
880 $self->_driver_determined(1);
882 $self->_run_connection_actions
883 if $started_unconnected && defined $self->_dbh;
887 sub _do_connection_actions {
889 my $method_prefix = shift;
892 if (not ref($call)) {
893 my $method = $method_prefix . $call;
895 } elsif (ref($call) eq 'CODE') {
897 } elsif (ref($call) eq 'ARRAY') {
898 if (ref($call->[0]) ne 'ARRAY') {
899 $self->_do_connection_actions($method_prefix, $_) for @$call;
901 $self->_do_connection_actions($method_prefix, @$_) for @$call;
904 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
910 sub connect_call_do_sql {
912 $self->_do_query(@_);
915 sub disconnect_call_do_sql {
917 $self->_do_query(@_);
920 # override in db-specific backend when necessary
921 sub connect_call_datetime_setup { 1 }
924 my ($self, $action) = @_;
926 if (ref $action eq 'CODE') {
927 $action = $action->($self);
928 $self->_do_query($_) foreach @$action;
931 # Most debuggers expect ($sql, @bind), so we need to exclude
932 # the attribute hash which is the second argument to $dbh->do
933 # furthermore the bind values are usually to be presented
934 # as named arrayref pairs, so wrap those here too
935 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
936 my $sql = shift @do_args;
937 my $attrs = shift @do_args;
938 my @bind = map { [ undef, $_ ] } @do_args;
940 $self->_query_start($sql, @bind);
941 $self->_dbh->do($sql, $attrs, @do_args);
942 $self->_query_end($sql, @bind);
949 my ($self, @info) = @_;
951 $self->throw_exception("You failed to provide any connection info")
954 my ($old_connect_via, $dbh);
956 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
957 $old_connect_via = $DBI::connect_via;
958 $DBI::connect_via = 'connect';
962 if(ref $info[0] eq 'CODE') {
966 $dbh = DBI->connect(@info);
969 if($dbh && !$self->unsafe) {
970 my $weak_self = $self;
971 Scalar::Util::weaken($weak_self);
972 $dbh->{HandleError} = sub {
974 $weak_self->throw_exception("DBI Exception: $_[0]");
977 croak ("DBI Exception: $_[0]");
980 $dbh->{ShowErrorStatement} = 1;
981 $dbh->{RaiseError} = 1;
982 $dbh->{PrintError} = 0;
986 $DBI::connect_via = $old_connect_via if $old_connect_via;
988 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
991 $self->_dbh_autocommit($dbh->{AutoCommit});
997 my ($self, $name) = @_;
999 $name = $self->_svp_generate_name
1000 unless defined $name;
1002 $self->throw_exception ("You can't use savepoints outside a transaction")
1003 if $self->{transaction_depth} == 0;
1005 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1006 unless $self->can('_svp_begin');
1008 push @{ $self->{savepoints} }, $name;
1010 $self->debugobj->svp_begin($name) if $self->debug;
1012 return $self->_svp_begin($name);
1016 my ($self, $name) = @_;
1018 $self->throw_exception ("You can't use savepoints outside a transaction")
1019 if $self->{transaction_depth} == 0;
1021 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1022 unless $self->can('_svp_release');
1024 if (defined $name) {
1025 $self->throw_exception ("Savepoint '$name' does not exist")
1026 unless grep { $_ eq $name } @{ $self->{savepoints} };
1028 # Dig through the stack until we find the one we are releasing. This keeps
1029 # the stack up to date.
1032 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1034 $name = pop @{ $self->{savepoints} };
1037 $self->debugobj->svp_release($name) if $self->debug;
1039 return $self->_svp_release($name);
1043 my ($self, $name) = @_;
1045 $self->throw_exception ("You can't use savepoints outside a transaction")
1046 if $self->{transaction_depth} == 0;
1048 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1049 unless $self->can('_svp_rollback');
1051 if (defined $name) {
1052 # If they passed us a name, verify that it exists in the stack
1053 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1054 $self->throw_exception("Savepoint '$name' does not exist!");
1057 # Dig through the stack until we find the one we are releasing. This keeps
1058 # the stack up to date.
1059 while(my $s = pop(@{ $self->{savepoints} })) {
1060 last if($s eq $name);
1062 # Add the savepoint back to the stack, as a rollback doesn't remove the
1063 # named savepoint, only everything after it.
1064 push(@{ $self->{savepoints} }, $name);
1066 # We'll assume they want to rollback to the last savepoint
1067 $name = $self->{savepoints}->[-1];
1070 $self->debugobj->svp_rollback($name) if $self->debug;
1072 return $self->_svp_rollback($name);
1075 sub _svp_generate_name {
1078 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1083 if($self->{transaction_depth} == 0) {
1084 $self->debugobj->txn_begin()
1087 # being here implies we have AutoCommit => 1
1088 # if the user is utilizing txn_do - good for
1089 # him, otherwise we need to ensure that the
1090 # $dbh is healthy on BEGIN
1091 my $dbh_method = $self->{_in_dbh_do} ? '_dbh' : 'dbh';
1092 $self->$dbh_method->begin_work;
1094 } elsif ($self->auto_savepoint) {
1097 $self->{transaction_depth}++;
1102 if ($self->{transaction_depth} == 1) {
1103 my $dbh = $self->_dbh;
1104 $self->debugobj->txn_commit()
1107 $self->{transaction_depth} = 0
1108 if $self->_dbh_autocommit;
1110 elsif($self->{transaction_depth} > 1) {
1111 $self->{transaction_depth}--;
1113 if $self->auto_savepoint;
1119 my $dbh = $self->_dbh;
1121 if ($self->{transaction_depth} == 1) {
1122 $self->debugobj->txn_rollback()
1124 $self->{transaction_depth} = 0
1125 if $self->_dbh_autocommit;
1128 elsif($self->{transaction_depth} > 1) {
1129 $self->{transaction_depth}--;
1130 if ($self->auto_savepoint) {
1131 $self->svp_rollback;
1136 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1141 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1142 $error =~ /$exception_class/ and $self->throw_exception($error);
1143 # ensure that a failed rollback resets the transaction depth
1144 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1145 $self->throw_exception($error);
1149 # This used to be the top-half of _execute. It was split out to make it
1150 # easier to override in NoBindVars without duping the rest. It takes up
1151 # all of _execute's args, and emits $sql, @bind.
1152 sub _prep_for_execute {
1153 my ($self, $op, $extra_bind, $ident, $args) = @_;
1155 if( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1156 $ident = $ident->from();
1159 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1162 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1164 return ($sql, \@bind);
1168 sub _fix_bind_params {
1169 my ($self, @bind) = @_;
1171 ### Turn @bind from something like this:
1172 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1174 ### ( "'1'", "'1'", "'3'" )
1177 if ( defined( $_ && $_->[1] ) ) {
1178 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1185 my ( $self, $sql, @bind ) = @_;
1187 if ( $self->debug ) {
1188 @bind = $self->_fix_bind_params(@bind);
1190 $self->debugobj->query_start( $sql, @bind );
1195 my ( $self, $sql, @bind ) = @_;
1197 if ( $self->debug ) {
1198 @bind = $self->_fix_bind_params(@bind);
1199 $self->debugobj->query_end( $sql, @bind );
1204 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1206 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1208 $self->_query_start( $sql, @$bind );
1210 my $sth = $self->sth($sql,$op);
1212 my $placeholder_index = 1;
1214 foreach my $bound (@$bind) {
1215 my $attributes = {};
1216 my($column_name, @data) = @$bound;
1218 if ($bind_attributes) {
1219 $attributes = $bind_attributes->{$column_name}
1220 if defined $bind_attributes->{$column_name};
1223 foreach my $data (@data) {
1224 my $ref = ref $data;
1225 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1227 $sth->bind_param($placeholder_index, $data, $attributes);
1228 $placeholder_index++;
1232 # Can this fail without throwing an exception anyways???
1233 my $rv = $sth->execute();
1234 $self->throw_exception($sth->errstr) if !$rv;
1236 $self->_query_end( $sql, @$bind );
1238 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1243 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1247 my ($self, $source, $to_insert) = @_;
1249 # redispatch to insert method of storage we reblessed into, if necessary
1250 if (not $self->_driver_determined) {
1251 $self->_determine_driver;
1252 goto $self->can('insert');
1255 my $ident = $source->from;
1256 my $bind_attributes = $self->source_bind_attributes($source);
1258 my $updated_cols = {};
1260 foreach my $col ( $source->columns ) {
1261 if ( !defined $to_insert->{$col} ) {
1262 my $col_info = $source->column_info($col);
1264 if ( $col_info->{auto_nextval} ) {
1265 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1267 $col_info->{sequence} ||
1268 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source)
1274 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
1276 return $updated_cols;
1279 ## Still not quite perfect, and EXPERIMENTAL
1280 ## Currently it is assumed that all values passed will be "normal", i.e. not
1281 ## scalar refs, or at least, all the same type as the first set, the statement is
1282 ## only prepped once.
1284 my ($self, $source, $cols, $data) = @_;
1286 my $table = $source->from;
1287 @colvalues{@$cols} = (0..$#$cols);
1288 my ($sql, @bind) = $self->sql_maker->insert($table, \%colvalues);
1290 $self->_determine_driver;
1292 $self->_query_start( $sql, @bind );
1293 my $sth = $self->sth($sql);
1295 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1297 ## This must be an arrayref, else nothing works!
1298 my $tuple_status = [];
1300 ## Get the bind_attributes, if any exist
1301 my $bind_attributes = $self->source_bind_attributes($source);
1303 ## Bind the values and execute
1304 my $placeholder_index = 1;
1306 foreach my $bound (@bind) {
1308 my $attributes = {};
1309 my ($column_name, $data_index) = @$bound;
1311 if( $bind_attributes ) {
1312 $attributes = $bind_attributes->{$column_name}
1313 if defined $bind_attributes->{$column_name};
1316 my @data = map { $_->[$data_index] } @$data;
1318 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1319 $placeholder_index++;
1321 my $rv = eval { $sth->execute_array({ArrayTupleStatus => $tuple_status}) };
1324 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1326 $self->throw_exception($sth->errstr || "Unexpected populate error: $err")
1327 if ($i > $#$tuple_status);
1329 require Data::Dumper;
1330 local $Data::Dumper::Terse = 1;
1331 local $Data::Dumper::Indent = 1;
1332 local $Data::Dumper::Useqq = 1;
1333 local $Data::Dumper::Quotekeys = 0;
1335 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1336 $tuple_status->[$i][1],
1337 Data::Dumper::Dumper(
1338 { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) }
1342 $self->throw_exception($sth->errstr) if !$rv;
1344 $self->_query_end( $sql, @bind );
1345 return (wantarray ? ($rv, $sth, @bind) : $rv);
1349 my $self = shift @_;
1350 my $source = shift @_;
1351 $self->_determine_driver;
1352 my $bind_attributes = $self->source_bind_attributes($source);
1354 return $self->_execute('update' => [], $source, $bind_attributes, @_);
1359 my $self = shift @_;
1360 my $source = shift @_;
1361 $self->_determine_driver;
1362 my $bind_attrs = $self->source_bind_attributes($source);
1364 return $self->_execute('delete' => [], $source, $bind_attrs, @_);
1367 # We were sent here because the $rs contains a complex search
1368 # which will require a subquery to select the correct rows
1369 # (i.e. joined or limited resultsets)
1371 # Genarating a single PK column subquery is trivial and supported
1372 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1373 # Look at _multipk_update_delete()
1374 sub _subq_update_delete {
1376 my ($rs, $op, $values) = @_;
1378 my $rsrc = $rs->result_source;
1380 # we already check this, but double check naively just in case. Should be removed soon
1381 my $sel = $rs->_resolved_attrs->{select};
1382 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1383 my @pcols = $rsrc->primary_columns;
1384 if (@$sel != @pcols) {
1385 $self->throw_exception (
1386 'Subquery update/delete can not be called on resultsets selecting a'
1387 .' number of columns different than the number of primary keys'
1394 $op eq 'update' ? $values : (),
1395 { $pcols[0] => { -in => $rs->as_query } },
1400 return $self->_multipk_update_delete (@_);
1404 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1405 # resultset update/delete involving subqueries. So by default resort
1406 # to simple (and inefficient) delete_all style per-row opearations,
1407 # while allowing specific storages to override this with a faster
1410 sub _multipk_update_delete {
1411 return shift->_per_row_update_delete (@_);
1414 # This is the default loop used to delete/update rows for multi PK
1415 # resultsets, and used by mysql exclusively (because it can't do anything
1418 # We do not use $row->$op style queries, because resultset update/delete
1419 # is not expected to cascade (this is what delete_all/update_all is for).
1421 # There should be no race conditions as the entire operation is rolled
1424 sub _per_row_update_delete {
1426 my ($rs, $op, $values) = @_;
1428 my $rsrc = $rs->result_source;
1429 my @pcols = $rsrc->primary_columns;
1431 my $guard = $self->txn_scope_guard;
1433 # emulate the return value of $sth->execute for non-selects
1434 my $row_cnt = '0E0';
1436 my $subrs_cur = $rs->cursor;
1437 while (my @pks = $subrs_cur->next) {
1440 for my $i (0.. $#pcols) {
1441 $cond->{$pcols[$i]} = $pks[$i];
1446 $op eq 'update' ? $values : (),
1461 # localization is neccessary as
1462 # 1) there is no infrastructure to pass this around before SQLA2
1463 # 2) _select_args sets it and _prep_for_execute consumes it
1464 my $sql_maker = $self->sql_maker;
1465 local $sql_maker->{_dbic_rs_attrs};
1467 return $self->_execute($self->_select_args(@_));
1470 sub _select_args_to_query {
1473 # localization is neccessary as
1474 # 1) there is no infrastructure to pass this around before SQLA2
1475 # 2) _select_args sets it and _prep_for_execute consumes it
1476 my $sql_maker = $self->sql_maker;
1477 local $sql_maker->{_dbic_rs_attrs};
1479 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $order, $rows, $offset)
1480 # = $self->_select_args($ident, $select, $cond, $attrs);
1481 my ($op, $bind, $ident, $bind_attrs, @args) =
1482 $self->_select_args(@_);
1484 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $order, $rows, $offset ]);
1485 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1486 $prepared_bind ||= [];
1489 ? ($sql, $prepared_bind, $bind_attrs)
1490 : \[ "($sql)", @$prepared_bind ]
1495 my ($self, $ident, $select, $where, $attrs) = @_;
1497 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1499 my $sql_maker = $self->sql_maker;
1500 $sql_maker->{_dbic_rs_attrs} = {
1506 ? ( _source_handle => $alias2source->{$rs_alias}->handle )
1511 # calculate bind_attrs before possible $ident mangling
1512 my $bind_attrs = {};
1513 for my $alias (keys %$alias2source) {
1514 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
1515 for my $col (keys %$bindtypes) {
1517 my $fqcn = join ('.', $alias, $col);
1518 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
1520 # Unqialified column names are nice, but at the same time can be
1521 # rather ambiguous. What we do here is basically go along with
1522 # the loop, adding an unqualified column slot to $bind_attrs,
1523 # alongside the fully qualified name. As soon as we encounter
1524 # another column by that name (which would imply another table)
1525 # we unset the unqualified slot and never add any info to it
1526 # to avoid erroneous type binding. If this happens the users
1527 # only choice will be to fully qualify his column name
1529 if (exists $bind_attrs->{$col}) {
1530 $bind_attrs->{$col} = {};
1533 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
1540 $attrs->{software_limit}
1542 $sql_maker->_default_limit_syntax eq "GenericSubQ"
1544 $attrs->{software_limit} = 1;
1547 $self->throw_exception("rows attribute must be positive if present")
1548 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1550 # MySQL actually recommends this approach. I cringe.
1551 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1556 # see if we need to tear the prefetch apart (either limited has_many or grouped prefetch)
1557 # otherwise delegate the limiting to the storage, unless software limit was requested
1559 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
1561 ( $attrs->{group_by} && @{$attrs->{group_by}} &&
1562 $attrs->{_prefetch_select} && @{$attrs->{_prefetch_select}} )
1564 ($ident, $select, $where, $attrs)
1565 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
1567 elsif (! $attrs->{software_limit} ) {
1568 push @limit, $attrs->{rows}, $attrs->{offset};
1572 # This would be the point to deflate anything found in $where
1573 # (and leave $attrs->{bind} intact). Problem is - inflators historically
1574 # expect a row object. And all we have is a resultsource (it is trivial
1575 # to extract deflator coderefs via $alias2source above).
1577 # I don't see a way forward other than changing the way deflators are
1578 # invoked, and that's just bad...
1582 { $attrs->{$_} ? ( $_ => $attrs->{$_} ) : () }
1583 (qw/order_by group_by having/ )
1586 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $order, @limit);
1590 # This is the code producing joined subqueries like:
1591 # SELECT me.*, other.* FROM ( SELECT me.* FROM ... ) JOIN other ON ...
1593 sub _adjust_select_args_for_complex_prefetch {
1594 my ($self, $from, $select, $where, $attrs) = @_;
1596 $self->throw_exception ('Nothing to prefetch... how did we get here?!')
1597 if not @{$attrs->{_prefetch_select}};
1599 $self->throw_exception ('Complex prefetches are not supported on resultsets with a custom from attribute')
1600 if (ref $from ne 'ARRAY' || ref $from->[0] ne 'HASH' || ref $from->[1] ne 'ARRAY');
1603 # generate inner/outer attribute lists, remove stuff that doesn't apply
1604 my $outer_attrs = { %$attrs };
1605 delete $outer_attrs->{$_} for qw/where bind rows offset group_by having/;
1607 my $inner_attrs = { %$attrs };
1608 delete $inner_attrs->{$_} for qw/for collapse _prefetch_select _collapse_order_by select as/;
1611 # bring over all non-collapse-induced order_by into the inner query (if any)
1612 # the outer one will have to keep them all
1613 delete $inner_attrs->{order_by};
1614 if (my $ord_cnt = @{$outer_attrs->{order_by}} - @{$outer_attrs->{_collapse_order_by}} ) {
1615 $inner_attrs->{order_by} = [
1616 @{$outer_attrs->{order_by}}[ 0 .. $ord_cnt - 1]
1621 # generate the inner/outer select lists
1622 # for inside we consider only stuff *not* brought in by the prefetch
1623 # on the outside we substitute any function for its alias
1624 my $outer_select = [ @$select ];
1625 my $inner_select = [];
1626 for my $i (0 .. ( @$outer_select - @{$outer_attrs->{_prefetch_select}} - 1) ) {
1627 my $sel = $outer_select->[$i];
1629 if (ref $sel eq 'HASH' ) {
1630 $sel->{-as} ||= $attrs->{as}[$i];
1631 $outer_select->[$i] = join ('.', $attrs->{alias}, ($sel->{-as} || "inner_column_$i") );
1634 push @$inner_select, $sel;
1637 # normalize a copy of $from, so it will be easier to work with further
1638 # down (i.e. promote the initial hashref to an AoH)
1640 $from->[0] = [ $from->[0] ];
1641 my %original_join_info = map { $_->[0]{-alias} => $_->[0] } (@$from);
1644 # decide which parts of the join will remain in either part of
1645 # the outer/inner query
1647 # First we compose a list of which aliases are used in restrictions
1648 # (i.e. conditions/order/grouping/etc). Since we do not have
1649 # introspectable SQLA, we fall back to ugly scanning of raw SQL for
1650 # WHERE, and for pieces of ORDER BY in order to determine which aliases
1651 # need to appear in the resulting sql.
1652 # It may not be very efficient, but it's a reasonable stop-gap
1653 # Also unqualified column names will not be considered, but more often
1654 # than not this is actually ok
1656 # In the same loop we enumerate part of the selection aliases, as
1657 # it requires the same sqla hack for the time being
1658 my ($restrict_aliases, $select_aliases, $prefetch_aliases);
1660 # produce stuff unquoted, so it can be scanned
1661 my $sql_maker = $self->sql_maker;
1662 local $sql_maker->{quote_char};
1663 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
1666 my $non_prefetch_select_sql = $sql_maker->_recurse_fields ($inner_select);
1667 my $prefetch_select_sql = $sql_maker->_recurse_fields ($outer_attrs->{_prefetch_select});
1668 my $where_sql = $sql_maker->where ($where);
1669 my $group_by_sql = $sql_maker->_order_by({
1670 map { $_ => $inner_attrs->{$_} } qw/group_by having/
1672 my @non_prefetch_order_by_chunks = (map
1673 { ref $_ ? $_->[0] : $_ }
1674 $sql_maker->_order_by_chunks ($inner_attrs->{order_by})
1678 for my $alias (keys %original_join_info) {
1679 my $seen_re = qr/\b $alias $sep/x;
1681 for my $piece ($where_sql, $group_by_sql, @non_prefetch_order_by_chunks ) {
1682 if ($piece =~ $seen_re) {
1683 $restrict_aliases->{$alias} = 1;
1687 if ($non_prefetch_select_sql =~ $seen_re) {
1688 $select_aliases->{$alias} = 1;
1691 if ($prefetch_select_sql =~ $seen_re) {
1692 $prefetch_aliases->{$alias} = 1;
1698 # Add any non-left joins to the restriction list (such joins are indeed restrictions)
1699 for my $j (values %original_join_info) {
1700 my $alias = $j->{-alias} or next;
1701 $restrict_aliases->{$alias} = 1 if (
1702 (not $j->{-join_type})
1704 ($j->{-join_type} !~ /^left (?: \s+ outer)? $/xi)
1708 # mark all join parents as mentioned
1709 # (e.g. join => { cds => 'tracks' } - tracks will need to bring cds too )
1710 for my $collection ($restrict_aliases, $select_aliases) {
1711 for my $alias (keys %$collection) {
1712 $collection->{$_} = 1
1713 for (@{ $original_join_info{$alias}{-join_path} || [] });
1717 # construct the inner $from for the subquery
1718 my %inner_joins = (map { %{$_ || {}} } ($restrict_aliases, $select_aliases) );
1720 for my $j (@$from) {
1721 push @inner_from, $j if $inner_joins{$j->[0]{-alias}};
1724 # if a multi-type join was needed in the subquery ("multi" is indicated by
1725 # presence in {collapse}) - add a group_by to simulate the collapse in the subq
1726 unless ($inner_attrs->{group_by}) {
1727 for my $alias (keys %inner_joins) {
1729 # the dot comes from some weirdness in collapse
1730 # remove after the rewrite
1731 if ($attrs->{collapse}{".$alias"}) {
1732 $inner_attrs->{group_by} ||= $inner_select;
1738 # demote the inner_from head
1739 $inner_from[0] = $inner_from[0][0];
1741 # generate the subquery
1742 my $subq = $self->_select_args_to_query (
1749 my $subq_joinspec = {
1750 -alias => $attrs->{alias},
1751 -source_handle => $inner_from[0]{-source_handle},
1752 $attrs->{alias} => $subq,
1755 # Generate the outer from - this is relatively easy (really just replace
1756 # the join slot with the subquery), with a major caveat - we can not
1757 # join anything that is non-selecting (not part of the prefetch), but at
1758 # the same time is a multi-type relationship, as it will explode the result.
1760 # There are two possibilities here
1761 # - either the join is non-restricting, in which case we simply throw it away
1762 # - it is part of the restrictions, in which case we need to collapse the outer
1763 # result by tackling yet another group_by to the outside of the query
1765 # so first generate the outer_from, up to the substitution point
1767 while (my $j = shift @$from) {
1768 if ($j->[0]{-alias} eq $attrs->{alias}) { # time to swap
1773 last; # we'll take care of what's left in $from below
1776 push @outer_from, $j;
1780 # see what's left - throw away if not selecting/restricting
1781 # also throw in a group_by if restricting to guard against
1782 # cross-join explosions
1784 while (my $j = shift @$from) {
1785 my $alias = $j->[0]{-alias};
1787 if ($select_aliases->{$alias} || $prefetch_aliases->{$alias}) {
1788 push @outer_from, $j;
1790 elsif ($restrict_aliases->{$alias}) {
1791 push @outer_from, $j;
1793 # FIXME - this should be obviated by SQLA2, as I'll be able to
1794 # have restrict_inner and restrict_outer... or something to that
1795 # effect... I think...
1797 # FIXME2 - I can't find a clean way to determine if a particular join
1798 # is a multi - instead I am just treating everything as a potential
1799 # explosive join (ribasushi)
1801 # if (my $handle = $j->[0]{-source_handle}) {
1802 # my $rsrc = $handle->resolve;
1803 # ... need to bail out of the following if this is not a multi,
1804 # as it will be much easier on the db ...
1806 $outer_attrs->{group_by} ||= $outer_select;
1811 # demote the outer_from head
1812 $outer_from[0] = $outer_from[0][0];
1814 # This is totally horrific - the $where ends up in both the inner and outer query
1815 # Unfortunately not much can be done until SQLA2 introspection arrives, and even
1816 # then if where conditions apply to the *right* side of the prefetch, you may have
1817 # to both filter the inner select (e.g. to apply a limit) and then have to re-filter
1818 # the outer select to exclude joins you didin't want in the first place
1820 # OTOH it can be seen as a plus: <ash> (notes that this query would make a DBA cry ;)
1821 return (\@outer_from, $outer_select, $where, $outer_attrs);
1824 sub _resolve_ident_sources {
1825 my ($self, $ident) = @_;
1827 my $alias2source = {};
1830 # the reason this is so contrived is that $ident may be a {from}
1831 # structure, specifying multiple tables to join
1832 if ( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1833 # this is compat mode for insert/update/delete which do not deal with aliases
1834 $alias2source->{me} = $ident;
1837 elsif (ref $ident eq 'ARRAY') {
1841 if (ref $_ eq 'HASH') {
1843 $rs_alias = $tabinfo->{-alias};
1845 if (ref $_ eq 'ARRAY' and ref $_->[0] eq 'HASH') {
1849 $alias2source->{$tabinfo->{-alias}} = $tabinfo->{-source_handle}->resolve
1850 if ($tabinfo->{-source_handle});
1854 return ($alias2source, $rs_alias);
1857 # Takes $ident, \@column_names
1859 # returns { $column_name => \%column_info, ... }
1860 # also note: this adds -result_source => $rsrc to the column info
1863 # my $col_sources = $self->_resolve_column_info($ident, @column_names);
1864 sub _resolve_column_info {
1865 my ($self, $ident, $colnames) = @_;
1866 my ($alias2src, $root_alias) = $self->_resolve_ident_sources($ident);
1868 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
1871 my (%return, %seen_cols);
1873 # compile a global list of column names, to be able to properly
1874 # disambiguate unqualified column names (if at all possible)
1875 for my $alias (keys %$alias2src) {
1876 my $rsrc = $alias2src->{$alias};
1877 for my $colname ($rsrc->columns) {
1878 push @{$seen_cols{$colname}}, $alias;
1883 foreach my $col (@$colnames) {
1884 my ($alias, $colname) = $col =~ m/^ (?: ([^$sep]+) $sep)? (.+) $/x;
1887 # see if the column was seen exactly once (so we know which rsrc it came from)
1888 if ($seen_cols{$colname} and @{$seen_cols{$colname}} == 1) {
1889 $alias = $seen_cols{$colname}[0];
1896 my $rsrc = $alias2src->{$alias};
1897 $return{$col} = $rsrc && {
1898 %{$rsrc->column_info($colname)},
1899 -result_source => $rsrc,
1900 -source_alias => $alias,
1907 # Returns a counting SELECT for a simple count
1908 # query. Abstracted so that a storage could override
1909 # this to { count => 'firstcol' } or whatever makes
1910 # sense as a performance optimization
1912 #my ($self, $source, $rs_attrs) = @_;
1913 return { count => '*' };
1916 # Returns a SELECT which will end up in the subselect
1917 # There may or may not be a group_by, as the subquery
1918 # might have been called to accomodate a limit
1920 # Most databases would be happy with whatever ends up
1921 # here, but some choke in various ways.
1923 sub _subq_count_select {
1924 my ($self, $source, $rs_attrs) = @_;
1925 return $rs_attrs->{group_by} if $rs_attrs->{group_by};
1927 my @pcols = map { join '.', $rs_attrs->{alias}, $_ } ($source->primary_columns);
1928 return @pcols ? \@pcols : [ 1 ];
1932 sub source_bind_attributes {
1933 my ($self, $source) = @_;
1935 my $bind_attributes;
1936 foreach my $column ($source->columns) {
1938 my $data_type = $source->column_info($column)->{data_type} || '';
1939 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
1943 return $bind_attributes;
1950 =item Arguments: $ident, $select, $condition, $attrs
1954 Handle a SQL select statement.
1960 my ($ident, $select, $condition, $attrs) = @_;
1961 return $self->cursor_class->new($self, \@_, $attrs);
1966 my ($rv, $sth, @bind) = $self->_select(@_);
1967 my @row = $sth->fetchrow_array;
1968 my @nextrow = $sth->fetchrow_array if @row;
1969 if(@row && @nextrow) {
1970 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
1972 # Need to call finish() to work round broken DBDs
1981 =item Arguments: $sql
1985 Returns a L<DBI> sth (statement handle) for the supplied SQL.
1990 my ($self, $dbh, $sql) = @_;
1992 # 3 is the if_active parameter which avoids active sth re-use
1993 my $sth = $self->disable_sth_caching
1994 ? $dbh->prepare($sql)
1995 : $dbh->prepare_cached($sql, {}, 3);
1997 # XXX You would think RaiseError would make this impossible,
1998 # but apparently that's not true :(
1999 $self->throw_exception($dbh->errstr) if !$sth;
2005 my ($self, $sql) = @_;
2006 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2009 sub _dbh_columns_info_for {
2010 my ($self, $dbh, $table) = @_;
2012 if ($dbh->can('column_info')) {
2015 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2016 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2018 while ( my $info = $sth->fetchrow_hashref() ){
2020 $column_info{data_type} = $info->{TYPE_NAME};
2021 $column_info{size} = $info->{COLUMN_SIZE};
2022 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2023 $column_info{default_value} = $info->{COLUMN_DEF};
2024 my $col_name = $info->{COLUMN_NAME};
2025 $col_name =~ s/^\"(.*)\"$/$1/;
2027 $result{$col_name} = \%column_info;
2030 return \%result if !$@ && scalar keys %result;
2034 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2036 my @columns = @{$sth->{NAME_lc}};
2037 for my $i ( 0 .. $#columns ){
2039 $column_info{data_type} = $sth->{TYPE}->[$i];
2040 $column_info{size} = $sth->{PRECISION}->[$i];
2041 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2043 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2044 $column_info{data_type} = $1;
2045 $column_info{size} = $2;
2048 $result{$columns[$i]} = \%column_info;
2052 foreach my $col (keys %result) {
2053 my $colinfo = $result{$col};
2054 my $type_num = $colinfo->{data_type};
2056 if(defined $type_num && $dbh->can('type_info')) {
2057 my $type_info = $dbh->type_info($type_num);
2058 $type_name = $type_info->{TYPE_NAME} if $type_info;
2059 $colinfo->{data_type} = $type_name if $type_name;
2066 sub columns_info_for {
2067 my ($self, $table) = @_;
2068 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2071 =head2 last_insert_id
2073 Return the row id of the last insert.
2077 sub _dbh_last_insert_id {
2078 # All Storage's need to register their own _dbh_last_insert_id
2079 # the old SQLite-based method was highly inappropriate
2082 my $class = ref $self;
2083 $self->throw_exception (<<EOE);
2085 No _dbh_last_insert_id() method found in $class.
2086 Since the method of obtaining the autoincrement id of the last insert
2087 operation varies greatly between different databases, this method must be
2088 individually implemented for every storage class.
2092 sub last_insert_id {
2094 $self->_dbh_last_insert_id ($self->_dbh, @_);
2097 =head2 _native_data_type
2101 =item Arguments: $type_name
2105 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2106 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2107 L<::Sybase|DBIx::Class::Storage::DBI::Sybase>.
2109 The default implementation returns C<undef>, implement in your Storage driver if
2110 you need this functionality.
2112 Should map types from other databases to the native RDBMS type, for example
2113 C<VARCHAR2> to C<VARCHAR>.
2115 Types with modifiers should map to the underlying data type. For example,
2116 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2118 Composite types should map to the container type, for example
2119 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2123 sub _native_data_type {
2124 #my ($self, $data_type) = @_;
2130 Returns the database driver name.
2137 if (not $self->_driver_determined) {
2138 $self->_determine_driver;
2139 goto $self->can ('sqlt_type');
2142 $self->_get_dbh->{Driver}->{Name};
2145 =head2 bind_attribute_by_data_type
2147 Given a datatype from column info, returns a database specific bind
2148 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2149 let the database planner just handle it.
2151 Generally only needed for special case column types, like bytea in postgres.
2155 sub bind_attribute_by_data_type {
2159 =head2 is_datatype_numeric
2161 Given a datatype from column_info, returns a boolean value indicating if
2162 the current RDBMS considers it a numeric value. This controls how
2163 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2164 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2165 be performed instead of the usual C<eq>.
2169 sub is_datatype_numeric {
2170 my ($self, $dt) = @_;
2172 return 0 unless $dt;
2174 return $dt =~ /^ (?:
2175 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2180 =head2 create_ddl_dir (EXPERIMENTAL)
2184 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2188 Creates a SQL file based on the Schema, for each of the specified
2189 database engines in C<\@databases> in the given directory.
2190 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2192 Given a previous version number, this will also create a file containing
2193 the ALTER TABLE statements to transform the previous schema into the
2194 current one. Note that these statements may contain C<DROP TABLE> or
2195 C<DROP COLUMN> statements that can potentially destroy data.
2197 The file names are created using the C<ddl_filename> method below, please
2198 override this method in your schema if you would like a different file
2199 name format. For the ALTER file, the same format is used, replacing
2200 $version in the name with "$preversion-$version".
2202 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2203 The most common value for this would be C<< { add_drop_table => 1 } >>
2204 to have the SQL produced include a C<DROP TABLE> statement for each table
2205 created. For quoting purposes supply C<quote_table_names> and
2206 C<quote_field_names>.
2208 If no arguments are passed, then the following default values are assumed:
2212 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2214 =item version - $schema->schema_version
2216 =item directory - './'
2218 =item preversion - <none>
2222 By default, C<\%sqlt_args> will have
2224 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2226 merged with the hash passed in. To disable any of those features, pass in a
2227 hashref like the following
2229 { ignore_constraint_names => 0, # ... other options }
2232 Note that this feature is currently EXPERIMENTAL and may not work correctly
2233 across all databases, or fully handle complex relationships.
2235 WARNING: Please check all SQL files created, before applying them.
2239 sub create_ddl_dir {
2240 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2242 if(!$dir || !-d $dir) {
2243 carp "No directory given, using ./\n";
2246 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2247 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2249 my $schema_version = $schema->schema_version || '1.x';
2250 $version ||= $schema_version;
2253 add_drop_table => 1,
2254 ignore_constraint_names => 1,
2255 ignore_index_names => 1,
2259 $self->throw_exception(q{Can't create a ddl file without SQL::Translator 0.09003: '}
2260 . $self->_check_sqlt_message . q{'})
2261 if !$self->_check_sqlt_version;
2263 my $sqlt = SQL::Translator->new( $sqltargs );
2265 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2266 my $sqlt_schema = $sqlt->translate({ data => $schema })
2267 or $self->throw_exception ($sqlt->error);
2269 foreach my $db (@$databases) {
2271 $sqlt->{schema} = $sqlt_schema;
2272 $sqlt->producer($db);
2275 my $filename = $schema->ddl_filename($db, $version, $dir);
2276 if (-e $filename && ($version eq $schema_version )) {
2277 # if we are dumping the current version, overwrite the DDL
2278 carp "Overwriting existing DDL file - $filename";
2282 my $output = $sqlt->translate;
2284 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2287 if(!open($file, ">$filename")) {
2288 $self->throw_exception("Can't open $filename for writing ($!)");
2291 print $file $output;
2294 next unless ($preversion);
2296 require SQL::Translator::Diff;
2298 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2299 if(!-e $prefilename) {
2300 carp("No previous schema file found ($prefilename)");
2304 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2306 carp("Overwriting existing diff file - $difffile");
2312 my $t = SQL::Translator->new($sqltargs);
2317 or $self->throw_exception ($t->error);
2319 my $out = $t->translate( $prefilename )
2320 or $self->throw_exception ($t->error);
2322 $source_schema = $t->schema;
2324 $source_schema->name( $prefilename )
2325 unless ( $source_schema->name );
2328 # The "new" style of producers have sane normalization and can support
2329 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2330 # And we have to diff parsed SQL against parsed SQL.
2331 my $dest_schema = $sqlt_schema;
2333 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2334 my $t = SQL::Translator->new($sqltargs);
2339 or $self->throw_exception ($t->error);
2341 my $out = $t->translate( $filename )
2342 or $self->throw_exception ($t->error);
2344 $dest_schema = $t->schema;
2346 $dest_schema->name( $filename )
2347 unless $dest_schema->name;
2350 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2354 if(!open $file, ">$difffile") {
2355 $self->throw_exception("Can't write to $difffile ($!)");
2363 =head2 deployment_statements
2367 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2371 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2373 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2374 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2376 C<$directory> is used to return statements from files in a previously created
2377 L</create_ddl_dir> directory and is optional. The filenames are constructed
2378 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2380 If no C<$directory> is specified then the statements are constructed on the
2381 fly using L<SQL::Translator> and C<$version> is ignored.
2383 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2387 sub deployment_statements {
2388 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2389 $type ||= $self->sqlt_type;
2390 $version ||= $schema->schema_version || '1.x';
2392 my $filename = $schema->ddl_filename($type, $version, $dir);
2396 open($file, "<$filename")
2397 or $self->throw_exception("Can't open $filename ($!)");
2400 return join('', @rows);
2403 $self->throw_exception(q{Can't deploy without SQL::Translator 0.09003: '}
2404 . $self->_check_sqlt_message . q{'})
2405 if !$self->_check_sqlt_version;
2407 # sources needs to be a parser arg, but for simplicty allow at top level
2409 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2410 if exists $sqltargs->{sources};
2412 my $tr = SQL::Translator->new(
2413 producer => "SQL::Translator::Producer::${type}",
2415 parser => 'SQL::Translator::Parser::DBIx::Class',
2418 return $tr->translate;
2422 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2425 return if($line =~ /^--/);
2427 # next if($line =~ /^DROP/m);
2428 return if($line =~ /^BEGIN TRANSACTION/m);
2429 return if($line =~ /^COMMIT/m);
2430 return if $line =~ /^\s+$/; # skip whitespace only
2431 $self->_query_start($line);
2433 # do a dbh_do cycle here, as we need some error checking in
2434 # place (even though we will ignore errors)
2435 $self->dbh_do (sub { $_[1]->do($line) });
2438 carp qq{$@ (running "${line}")};
2440 $self->_query_end($line);
2442 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2443 if (@statements > 1) {
2444 foreach my $statement (@statements) {
2445 $deploy->( $statement );
2448 elsif (@statements == 1) {
2449 foreach my $line ( split(";\n", $statements[0])) {
2455 =head2 datetime_parser
2457 Returns the datetime parser class
2461 sub datetime_parser {
2463 return $self->{datetime_parser} ||= do {
2464 $self->_populate_dbh unless $self->_dbh;
2465 $self->build_datetime_parser(@_);
2469 =head2 datetime_parser_type
2471 Defines (returns) the datetime parser class - currently hardwired to
2472 L<DateTime::Format::MySQL>
2476 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2478 =head2 build_datetime_parser
2480 See L</datetime_parser>
2484 sub build_datetime_parser {
2486 my $type = $self->datetime_parser_type(@_);
2488 $self->throw_exception("Couldn't load ${type}: $@") if $@;
2493 my $_check_sqlt_version; # private
2494 my $_check_sqlt_message; # private
2495 sub _check_sqlt_version {
2496 return $_check_sqlt_version if defined $_check_sqlt_version;
2497 eval 'use SQL::Translator "0.09003"';
2498 $_check_sqlt_message = $@ || '';
2499 $_check_sqlt_version = !$@;
2502 sub _check_sqlt_message {
2503 _check_sqlt_version if !defined $_check_sqlt_message;
2504 $_check_sqlt_message;
2508 =head2 is_replicating
2510 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2511 replicate from a master database. Default is undef, which is the result
2512 returned by databases that don't support replication.
2516 sub is_replicating {
2521 =head2 lag_behind_master
2523 Returns a number that represents a certain amount of lag behind a master db
2524 when a given storage is replicating. The number is database dependent, but
2525 starts at zero and increases with the amount of lag. Default in undef
2529 sub lag_behind_master {
2535 $self->_verify_pid if $self->_dbh;
2537 # some databases need this to stop spewing warnings
2538 if (my $dbh = $self->_dbh) {
2539 eval { $dbh->disconnect };
2549 =head2 DBIx::Class and AutoCommit
2551 DBIx::Class can do some wonderful magic with handling exceptions,
2552 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2553 (the default) combined with C<txn_do> for transaction support.
2555 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2556 in an assumed transaction between commits, and you're telling us you'd
2557 like to manage that manually. A lot of the magic protections offered by
2558 this module will go away. We can't protect you from exceptions due to database
2559 disconnects because we don't know anything about how to restart your
2560 transactions. You're on your own for handling all sorts of exceptional
2561 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2567 Matt S. Trout <mst@shadowcatsystems.co.uk>
2569 Andy Grundman <andy@hybridized.org>
2573 You may distribute this code under the same terms as Perl itself.