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 $connect_info_args = [{
116 dbh_maker => sub { DBI->connect (...) },
121 This is particularly useful for L<Catalyst> based applications, allowing the
122 following config (L<Config::General> style):
127 dsn dbi:mysql:database=test
134 The C<dsn>/C<user>/C<password> combination can be substituted by the
135 C<dbh_maker> key whose value is a coderef that returns a connected
136 L<DBI database handle|DBI/connect>
140 Please note that the L<DBI> docs recommend that you always explicitly
141 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
142 recommends that it be set to I<1>, and that you perform transactions
143 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
144 to I<1> if you do not do explicitly set it to zero. This is the default
145 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
147 =head3 DBIx::Class specific connection attributes
149 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
150 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
151 the following connection options. These options can be mixed in with your other
152 L<DBI> connection attributes, or placed in a seperate hashref
153 (C<\%extra_attributes>) as shown above.
155 Every time C<connect_info> is invoked, any previous settings for
156 these options will be cleared before setting the new ones, regardless of
157 whether any options are specified in the new C<connect_info>.
164 Specifies things to do immediately after connecting or re-connecting to
165 the database. Its value may contain:
171 This contains one SQL statement to execute.
173 =item an array reference
175 This contains SQL statements to execute in order. Each element contains
176 a string or a code reference that returns a string.
178 =item a code reference
180 This contains some code to execute. Unlike code references within an
181 array reference, its return value is ignored.
185 =item on_disconnect_do
187 Takes arguments in the same form as L</on_connect_do> and executes them
188 immediately before disconnecting from the database.
190 Note, this only runs if you explicitly call L</disconnect> on the
193 =item on_connect_call
195 A more generalized form of L</on_connect_do> that calls the specified
196 C<connect_call_METHOD> methods in your storage driver.
198 on_connect_do => 'select 1'
202 on_connect_call => [ [ do_sql => 'select 1' ] ]
204 Its values may contain:
210 Will call the C<connect_call_METHOD> method.
212 =item a code reference
214 Will execute C<< $code->($storage) >>
216 =item an array reference
218 Each value can be a method name or code reference.
220 =item an array of arrays
222 For each array, the first item is taken to be the C<connect_call_> method name
223 or code reference, and the rest are parameters to it.
227 Some predefined storage methods you may use:
233 Executes a SQL string or a code reference that returns a SQL string. This is
234 what L</on_connect_do> and L</on_disconnect_do> use.
242 Will execute the scalar as SQL.
246 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
247 attributes hashref and bind values.
249 =item a code reference
251 Will execute C<< $code->($storage) >> and execute the return array refs as
258 Execute any statements necessary to initialize the database session to return
259 and accept datetime/timestamp values used with
260 L<DBIx::Class::InflateColumn::DateTime>.
262 Only necessary for some databases, see your specific storage driver for
263 implementation details.
267 =item on_disconnect_call
269 Takes arguments in the same form as L</on_connect_call> and executes them
270 immediately before disconnecting from the database.
272 Calls the C<disconnect_call_METHOD> methods as opposed to the
273 C<connect_call_METHOD> methods called by L</on_connect_call>.
275 Note, this only runs if you explicitly call L</disconnect> on the
278 =item disable_sth_caching
280 If set to a true value, this option will disable the caching of
281 statement handles via L<DBI/prepare_cached>.
285 Sets the limit dialect. This is useful for JDBC-bridge among others
286 where the remote SQL-dialect cannot be determined by the name of the
287 driver alone. See also L<SQL::Abstract::Limit>.
291 Specifies what characters to use to quote table and column names. If
292 you use this you will want to specify L</name_sep> as well.
294 C<quote_char> expects either a single character, in which case is it
295 is placed on either side of the table/column name, or an arrayref of length
296 2 in which case the table/column name is placed between the elements.
298 For example under MySQL you should use C<< quote_char => '`' >>, and for
299 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
303 This only needs to be used in conjunction with C<quote_char>, and is used to
304 specify the charecter that seperates elements (schemas, tables, columns) from
305 each other. In most cases this is simply a C<.>.
307 The consequences of not supplying this value is that L<SQL::Abstract>
308 will assume DBIx::Class' uses of aliases to be complete column
309 names. The output will look like I<"me.name"> when it should actually
314 This Storage driver normally installs its own C<HandleError>, sets
315 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
316 all database handles, including those supplied by a coderef. It does this
317 so that it can have consistent and useful error behavior.
319 If you set this option to a true value, Storage will not do its usual
320 modifications to the database handle's attributes, and instead relies on
321 the settings in your connect_info DBI options (or the values you set in
322 your connection coderef, in the case that you are connecting via coderef).
324 Note that your custom settings can cause Storage to malfunction,
325 especially if you set a C<HandleError> handler that suppresses exceptions
326 and/or disable C<RaiseError>.
330 If this option is true, L<DBIx::Class> will use savepoints when nesting
331 transactions, making it possible to recover from failure in the inner
332 transaction without having to abort all outer transactions.
336 Use this argument to supply a cursor class other than the default
337 L<DBIx::Class::Storage::DBI::Cursor>.
341 Some real-life examples of arguments to L</connect_info> and
342 L<DBIx::Class::Schema/connect>
344 # Simple SQLite connection
345 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
348 ->connect_info([ sub { DBI->connect(...) } ]);
350 # Connect via subref in hashref
352 dbh_maker => sub { DBI->connect(...) },
353 on_connect_do => 'alter session ...',
356 # A bit more complicated
363 { quote_char => q{"}, name_sep => q{.} },
367 # Equivalent to the previous example
373 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
377 # Same, but with hashref as argument
378 # See parse_connect_info for explanation
381 dsn => 'dbi:Pg:dbname=foo',
383 password => 'my_pg_password',
390 # Subref + DBIx::Class-specific connection options
393 sub { DBI->connect(...) },
397 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
398 disable_sth_caching => 1,
408 my ($self, $info_arg) = @_;
410 return $self->_connect_info if !$info_arg;
412 my @args = @$info_arg; # take a shallow copy for further mutilation
413 $self->_connect_info([@args]); # copy for _connect_info
416 # combine/pre-parse arguments depending on invocation style
419 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
420 %attrs = %{ $args[1] || {} };
423 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
424 %attrs = %{$args[0]};
426 if (my $code = delete $attrs{dbh_maker}) {
429 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
432 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
433 . "to the result of 'dbh_maker'",
435 join (', ', map { "'$_'" } (@ignored) ),
440 @args = delete @attrs{qw/dsn user password/};
443 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
445 % { $args[3] || {} },
446 % { $args[4] || {} },
448 @args = @args[0,1,2];
451 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
452 # the new set of options
453 $self->_sql_maker(undef);
454 $self->_sql_maker_opts({});
457 for my $storage_opt (@storage_options, 'cursor_class') { # @storage_options is declared at the top of the module
458 if(my $value = delete $attrs{$storage_opt}) {
459 $self->$storage_opt($value);
462 for my $sql_maker_opt (qw/limit_dialect quote_char name_sep/) {
463 if(my $opt_val = delete $attrs{$sql_maker_opt}) {
464 $self->_sql_maker_opts->{$sql_maker_opt} = $opt_val;
469 if (ref $args[0] eq 'CODE') {
470 # _connect() never looks past $args[0] in this case
474 %{ $self->_default_dbi_connect_attributes || {} },
479 $self->_dbi_connect_info([@args, keys %attrs ? \%attrs : ()]);
480 $self->_connect_info;
483 sub _default_dbi_connect_attributes {
493 This method is deprecated in favour of setting via L</connect_info>.
497 =head2 on_disconnect_do
499 This method is deprecated in favour of setting via L</connect_info>.
503 sub _parse_connect_do {
504 my ($self, $type) = @_;
506 my $val = $self->$type;
507 return () if not defined $val;
512 push @res, [ 'do_sql', $val ];
513 } elsif (ref($val) eq 'CODE') {
515 } elsif (ref($val) eq 'ARRAY') {
516 push @res, map { [ 'do_sql', $_ ] } @$val;
518 $self->throw_exception("Invalid type for $type: ".ref($val));
526 Arguments: ($subref | $method_name), @extra_coderef_args?
528 Execute the given $subref or $method_name using the new exception-based
529 connection management.
531 The first two arguments will be the storage object that C<dbh_do> was called
532 on and a database handle to use. Any additional arguments will be passed
533 verbatim to the called subref as arguments 2 and onwards.
535 Using this (instead of $self->_dbh or $self->dbh) ensures correct
536 exception handling and reconnection (or failover in future subclasses).
538 Your subref should have no side-effects outside of the database, as
539 there is the potential for your subref to be partially double-executed
540 if the database connection was stale/dysfunctional.
544 my @stuff = $schema->storage->dbh_do(
546 my ($storage, $dbh, @cols) = @_;
547 my $cols = join(q{, }, @cols);
548 $dbh->selectrow_array("SELECT $cols FROM foo");
559 my $dbh = $self->_dbh;
561 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
562 || $self->{transaction_depth};
564 local $self->{_in_dbh_do} = 1;
567 my $want_array = wantarray;
570 $self->_verify_pid if $dbh;
572 $self->_populate_dbh;
577 @result = $self->$code($dbh, @_);
579 elsif(defined $want_array) {
580 $result[0] = $self->$code($dbh, @_);
583 $self->$code($dbh, @_);
587 # ->connected might unset $@ - copy
589 if(!$exception) { return $want_array ? @result : $result[0] }
591 $self->throw_exception($exception) if $self->connected;
593 # We were not connected - reconnect and retry, but let any
594 # exception fall right through this time
595 carp "Retrying $code after catching disconnected exception: $exception"
596 if $ENV{DBIC_DBIRETRY_DEBUG};
597 $self->_populate_dbh;
598 $self->$code($self->_dbh, @_);
601 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
602 # It also informs dbh_do to bypass itself while under the direction of txn_do,
603 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
608 ref $coderef eq 'CODE' or $self->throw_exception
609 ('$coderef must be a CODE reference');
611 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
613 local $self->{_in_dbh_do} = 1;
616 my $want_array = wantarray;
621 $self->_verify_pid if $self->_dbh;
622 $self->_populate_dbh if !$self->_dbh;
626 @result = $coderef->(@_);
628 elsif(defined $want_array) {
629 $result[0] = $coderef->(@_);
637 # ->connected might unset $@ - copy
639 if(!$exception) { return $want_array ? @result : $result[0] }
641 if($tried++ || $self->connected) {
642 eval { $self->txn_rollback };
643 my $rollback_exception = $@;
644 if($rollback_exception) {
645 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
646 $self->throw_exception($exception) # propagate nested rollback
647 if $rollback_exception =~ /$exception_class/;
649 $self->throw_exception(
650 "Transaction aborted: ${exception}. "
651 . "Rollback failed: ${rollback_exception}"
654 $self->throw_exception($exception)
657 # We were not connected, and was first try - reconnect and retry
659 carp "Retrying $coderef after catching disconnected exception: $exception"
660 if $ENV{DBIC_DBIRETRY_DEBUG};
661 $self->_populate_dbh;
667 Our C<disconnect> method also performs a rollback first if the
668 database is not in C<AutoCommit> mode.
678 push @actions, ( $self->on_disconnect_call || () );
679 push @actions, $self->_parse_connect_do ('on_disconnect_do');
681 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
683 $self->_dbh->rollback unless $self->_dbh_autocommit;
684 $self->_dbh->disconnect;
690 =head2 with_deferred_fk_checks
694 =item Arguments: C<$coderef>
696 =item Return Value: The return value of $coderef
700 Storage specific method to run the code ref with FK checks deferred or
701 in MySQL's case disabled entirely.
705 # Storage subclasses should override this
706 sub with_deferred_fk_checks {
707 my ($self, $sub) = @_;
716 =item Arguments: none
718 =item Return Value: 1|0
722 Verifies that the the current database handle is active and ready to execute
723 an SQL statement (i.e. the connection did not get stale, server is still
724 answering, etc.) This method is used internally by L</dbh>.
730 return 0 unless $self->_seems_connected;
733 local $self->_dbh->{RaiseError} = 1;
738 sub _seems_connected {
741 my $dbh = $self->_dbh
744 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
751 return 0 if !$self->_dbh;
754 return $dbh->FETCH('Active');
760 my $dbh = $self->_dbh or return 0;
765 # handle pid changes correctly
766 # NOTE: assumes $self->_dbh is a valid $dbh
770 return if defined $self->_conn_pid && $self->_conn_pid == $$;
772 $self->_dbh->{InactiveDestroy} = 1;
779 sub ensure_connected {
782 unless ($self->connected) {
783 $self->_populate_dbh;
789 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
790 is guaranteed to be healthy by implicitly calling L</connected>, and if
791 necessary performing a reconnection before returning. Keep in mind that this
792 is very B<expensive> on some database engines. Consider using L<dbh_do>
800 if (not $self->_dbh) {
801 $self->_populate_dbh;
803 $self->ensure_connected;
808 # this is the internal "get dbh or connect (don't check)" method
811 $self->_populate_dbh unless $self->_dbh;
815 sub _sql_maker_args {
820 array_datatypes => 1,
821 limit_dialect => $self->_get_dbh,
822 %{$self->_sql_maker_opts}
828 unless ($self->_sql_maker) {
829 my $sql_maker_class = $self->sql_maker_class;
830 $self->ensure_class_loaded ($sql_maker_class);
831 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
833 return $self->_sql_maker;
841 my @info = @{$self->_dbi_connect_info || []};
842 $self->_dbh(undef); # in case ->connected failed we might get sent here
843 $self->_dbh($self->_connect(@info));
845 $self->_conn_pid($$);
846 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
848 $self->_determine_driver;
850 # Always set the transaction depth on connect, since
851 # there is no transaction in progress by definition
852 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
854 $self->_run_connection_actions unless $self->{_in_determine_driver};
857 sub _run_connection_actions {
861 push @actions, ( $self->on_connect_call || () );
862 push @actions, $self->_parse_connect_do ('on_connect_do');
864 $self->_do_connection_actions(connect_call_ => $_) for @actions;
867 sub _determine_driver {
870 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
871 my $started_unconnected = 0;
872 local $self->{_in_determine_driver} = 1;
874 if (ref($self) eq __PACKAGE__) {
876 if ($self->_dbh) { # we are connected
877 $driver = $self->_dbh->{Driver}{Name};
879 # try to use dsn to not require being connected, the driver may still
880 # force a connection in _rebless to determine version
881 ($driver) = $self->_dbi_connect_info->[0] =~ /dbi:([^:]+):/i;
882 $started_unconnected = 1;
885 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
886 if ($self->load_optional_class($storage_class)) {
887 mro::set_mro($storage_class, 'c3');
888 bless $self, $storage_class;
893 $self->_driver_determined(1);
895 $self->_run_connection_actions
896 if $started_unconnected && defined $self->_dbh;
900 sub _do_connection_actions {
902 my $method_prefix = shift;
905 if (not ref($call)) {
906 my $method = $method_prefix . $call;
908 } elsif (ref($call) eq 'CODE') {
910 } elsif (ref($call) eq 'ARRAY') {
911 if (ref($call->[0]) ne 'ARRAY') {
912 $self->_do_connection_actions($method_prefix, $_) for @$call;
914 $self->_do_connection_actions($method_prefix, @$_) for @$call;
917 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
923 sub connect_call_do_sql {
925 $self->_do_query(@_);
928 sub disconnect_call_do_sql {
930 $self->_do_query(@_);
933 # override in db-specific backend when necessary
934 sub connect_call_datetime_setup { 1 }
937 my ($self, $action) = @_;
939 if (ref $action eq 'CODE') {
940 $action = $action->($self);
941 $self->_do_query($_) foreach @$action;
944 # Most debuggers expect ($sql, @bind), so we need to exclude
945 # the attribute hash which is the second argument to $dbh->do
946 # furthermore the bind values are usually to be presented
947 # as named arrayref pairs, so wrap those here too
948 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
949 my $sql = shift @do_args;
950 my $attrs = shift @do_args;
951 my @bind = map { [ undef, $_ ] } @do_args;
953 $self->_query_start($sql, @bind);
954 $self->_dbh->do($sql, $attrs, @do_args);
955 $self->_query_end($sql, @bind);
962 my ($self, @info) = @_;
964 $self->throw_exception("You failed to provide any connection info")
967 my ($old_connect_via, $dbh);
969 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
970 $old_connect_via = $DBI::connect_via;
971 $DBI::connect_via = 'connect';
975 if(ref $info[0] eq 'CODE') {
979 $dbh = DBI->connect(@info);
982 if($dbh && !$self->unsafe) {
983 my $weak_self = $self;
984 Scalar::Util::weaken($weak_self);
985 $dbh->{HandleError} = sub {
987 $weak_self->throw_exception("DBI Exception: $_[0]");
990 croak ("DBI Exception: $_[0]");
993 $dbh->{ShowErrorStatement} = 1;
994 $dbh->{RaiseError} = 1;
995 $dbh->{PrintError} = 0;
999 $DBI::connect_via = $old_connect_via if $old_connect_via;
1001 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
1004 $self->_dbh_autocommit($dbh->{AutoCommit});
1010 my ($self, $name) = @_;
1012 $name = $self->_svp_generate_name
1013 unless defined $name;
1015 $self->throw_exception ("You can't use savepoints outside a transaction")
1016 if $self->{transaction_depth} == 0;
1018 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1019 unless $self->can('_svp_begin');
1021 push @{ $self->{savepoints} }, $name;
1023 $self->debugobj->svp_begin($name) if $self->debug;
1025 return $self->_svp_begin($name);
1029 my ($self, $name) = @_;
1031 $self->throw_exception ("You can't use savepoints outside a transaction")
1032 if $self->{transaction_depth} == 0;
1034 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1035 unless $self->can('_svp_release');
1037 if (defined $name) {
1038 $self->throw_exception ("Savepoint '$name' does not exist")
1039 unless grep { $_ eq $name } @{ $self->{savepoints} };
1041 # Dig through the stack until we find the one we are releasing. This keeps
1042 # the stack up to date.
1045 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1047 $name = pop @{ $self->{savepoints} };
1050 $self->debugobj->svp_release($name) if $self->debug;
1052 return $self->_svp_release($name);
1056 my ($self, $name) = @_;
1058 $self->throw_exception ("You can't use savepoints outside a transaction")
1059 if $self->{transaction_depth} == 0;
1061 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1062 unless $self->can('_svp_rollback');
1064 if (defined $name) {
1065 # If they passed us a name, verify that it exists in the stack
1066 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1067 $self->throw_exception("Savepoint '$name' does not exist!");
1070 # Dig through the stack until we find the one we are releasing. This keeps
1071 # the stack up to date.
1072 while(my $s = pop(@{ $self->{savepoints} })) {
1073 last if($s eq $name);
1075 # Add the savepoint back to the stack, as a rollback doesn't remove the
1076 # named savepoint, only everything after it.
1077 push(@{ $self->{savepoints} }, $name);
1079 # We'll assume they want to rollback to the last savepoint
1080 $name = $self->{savepoints}->[-1];
1083 $self->debugobj->svp_rollback($name) if $self->debug;
1085 return $self->_svp_rollback($name);
1088 sub _svp_generate_name {
1091 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1096 if($self->{transaction_depth} == 0) {
1097 $self->debugobj->txn_begin()
1100 # being here implies we have AutoCommit => 1
1101 # if the user is utilizing txn_do - good for
1102 # him, otherwise we need to ensure that the
1103 # $dbh is healthy on BEGIN
1104 my $dbh_method = $self->{_in_dbh_do} ? '_dbh' : 'dbh';
1105 $self->$dbh_method->begin_work;
1107 } elsif ($self->auto_savepoint) {
1110 $self->{transaction_depth}++;
1115 if ($self->{transaction_depth} == 1) {
1116 my $dbh = $self->_dbh;
1117 $self->debugobj->txn_commit()
1120 $self->{transaction_depth} = 0
1121 if $self->_dbh_autocommit;
1123 elsif($self->{transaction_depth} > 1) {
1124 $self->{transaction_depth}--;
1126 if $self->auto_savepoint;
1132 my $dbh = $self->_dbh;
1134 if ($self->{transaction_depth} == 1) {
1135 $self->debugobj->txn_rollback()
1137 $self->{transaction_depth} = 0
1138 if $self->_dbh_autocommit;
1141 elsif($self->{transaction_depth} > 1) {
1142 $self->{transaction_depth}--;
1143 if ($self->auto_savepoint) {
1144 $self->svp_rollback;
1149 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1154 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1155 $error =~ /$exception_class/ and $self->throw_exception($error);
1156 # ensure that a failed rollback resets the transaction depth
1157 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1158 $self->throw_exception($error);
1162 # This used to be the top-half of _execute. It was split out to make it
1163 # easier to override in NoBindVars without duping the rest. It takes up
1164 # all of _execute's args, and emits $sql, @bind.
1165 sub _prep_for_execute {
1166 my ($self, $op, $extra_bind, $ident, $args) = @_;
1168 if( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1169 $ident = $ident->from();
1172 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1175 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1177 return ($sql, \@bind);
1181 sub _fix_bind_params {
1182 my ($self, @bind) = @_;
1184 ### Turn @bind from something like this:
1185 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1187 ### ( "'1'", "'1'", "'3'" )
1190 if ( defined( $_ && $_->[1] ) ) {
1191 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1198 my ( $self, $sql, @bind ) = @_;
1200 if ( $self->debug ) {
1201 @bind = $self->_fix_bind_params(@bind);
1203 $self->debugobj->query_start( $sql, @bind );
1208 my ( $self, $sql, @bind ) = @_;
1210 if ( $self->debug ) {
1211 @bind = $self->_fix_bind_params(@bind);
1212 $self->debugobj->query_end( $sql, @bind );
1217 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1219 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1221 $self->_query_start( $sql, @$bind );
1223 my $sth = $self->sth($sql,$op);
1225 my $placeholder_index = 1;
1227 foreach my $bound (@$bind) {
1228 my $attributes = {};
1229 my($column_name, @data) = @$bound;
1231 if ($bind_attributes) {
1232 $attributes = $bind_attributes->{$column_name}
1233 if defined $bind_attributes->{$column_name};
1236 foreach my $data (@data) {
1237 my $ref = ref $data;
1238 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1240 $sth->bind_param($placeholder_index, $data, $attributes);
1241 $placeholder_index++;
1245 # Can this fail without throwing an exception anyways???
1246 my $rv = $sth->execute();
1247 $self->throw_exception($sth->errstr) if !$rv;
1249 $self->_query_end( $sql, @$bind );
1251 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1256 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1260 my ($self, $source, $to_insert) = @_;
1262 # redispatch to insert method of storage we reblessed into, if necessary
1263 if (not $self->_driver_determined) {
1264 $self->_determine_driver;
1265 goto $self->can('insert');
1268 my $ident = $source->from;
1269 my $bind_attributes = $self->source_bind_attributes($source);
1271 my $updated_cols = {};
1273 foreach my $col ( $source->columns ) {
1274 if ( !defined $to_insert->{$col} ) {
1275 my $col_info = $source->column_info($col);
1277 if ( $col_info->{auto_nextval} ) {
1278 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1280 $col_info->{sequence} ||
1281 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source)
1287 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
1289 return $updated_cols;
1292 ## Still not quite perfect, and EXPERIMENTAL
1293 ## Currently it is assumed that all values passed will be "normal", i.e. not
1294 ## scalar refs, or at least, all the same type as the first set, the statement is
1295 ## only prepped once.
1297 my ($self, $source, $cols, $data) = @_;
1299 my $table = $source->from;
1300 @colvalues{@$cols} = (0..$#$cols);
1301 my ($sql, @bind) = $self->sql_maker->insert($table, \%colvalues);
1303 $self->_determine_driver;
1305 $self->_query_start( $sql, @bind );
1306 my $sth = $self->sth($sql);
1308 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1310 ## This must be an arrayref, else nothing works!
1311 my $tuple_status = [];
1313 ## Get the bind_attributes, if any exist
1314 my $bind_attributes = $self->source_bind_attributes($source);
1316 ## Bind the values and execute
1317 my $placeholder_index = 1;
1319 foreach my $bound (@bind) {
1321 my $attributes = {};
1322 my ($column_name, $data_index) = @$bound;
1324 if( $bind_attributes ) {
1325 $attributes = $bind_attributes->{$column_name}
1326 if defined $bind_attributes->{$column_name};
1329 my @data = map { $_->[$data_index] } @$data;
1331 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1332 $placeholder_index++;
1334 my $rv = eval { $sth->execute_array({ArrayTupleStatus => $tuple_status}) };
1337 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1339 $self->throw_exception($sth->errstr || "Unexpected populate error: $err")
1340 if ($i > $#$tuple_status);
1342 require Data::Dumper;
1343 local $Data::Dumper::Terse = 1;
1344 local $Data::Dumper::Indent = 1;
1345 local $Data::Dumper::Useqq = 1;
1346 local $Data::Dumper::Quotekeys = 0;
1348 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1349 $tuple_status->[$i][1],
1350 Data::Dumper::Dumper(
1351 { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) }
1355 $self->throw_exception($sth->errstr) if !$rv;
1357 $self->_query_end( $sql, @bind );
1358 return (wantarray ? ($rv, $sth, @bind) : $rv);
1362 my $self = shift @_;
1363 my $source = shift @_;
1364 $self->_determine_driver;
1365 my $bind_attributes = $self->source_bind_attributes($source);
1367 return $self->_execute('update' => [], $source, $bind_attributes, @_);
1372 my $self = shift @_;
1373 my $source = shift @_;
1374 $self->_determine_driver;
1375 my $bind_attrs = $self->source_bind_attributes($source);
1377 return $self->_execute('delete' => [], $source, $bind_attrs, @_);
1380 # We were sent here because the $rs contains a complex search
1381 # which will require a subquery to select the correct rows
1382 # (i.e. joined or limited resultsets)
1384 # Genarating a single PK column subquery is trivial and supported
1385 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1386 # Look at _multipk_update_delete()
1387 sub _subq_update_delete {
1389 my ($rs, $op, $values) = @_;
1391 my $rsrc = $rs->result_source;
1393 # we already check this, but double check naively just in case. Should be removed soon
1394 my $sel = $rs->_resolved_attrs->{select};
1395 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1396 my @pcols = $rsrc->primary_columns;
1397 if (@$sel != @pcols) {
1398 $self->throw_exception (
1399 'Subquery update/delete can not be called on resultsets selecting a'
1400 .' number of columns different than the number of primary keys'
1407 $op eq 'update' ? $values : (),
1408 { $pcols[0] => { -in => $rs->as_query } },
1413 return $self->_multipk_update_delete (@_);
1417 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1418 # resultset update/delete involving subqueries. So by default resort
1419 # to simple (and inefficient) delete_all style per-row opearations,
1420 # while allowing specific storages to override this with a faster
1423 sub _multipk_update_delete {
1424 return shift->_per_row_update_delete (@_);
1427 # This is the default loop used to delete/update rows for multi PK
1428 # resultsets, and used by mysql exclusively (because it can't do anything
1431 # We do not use $row->$op style queries, because resultset update/delete
1432 # is not expected to cascade (this is what delete_all/update_all is for).
1434 # There should be no race conditions as the entire operation is rolled
1437 sub _per_row_update_delete {
1439 my ($rs, $op, $values) = @_;
1441 my $rsrc = $rs->result_source;
1442 my @pcols = $rsrc->primary_columns;
1444 my $guard = $self->txn_scope_guard;
1446 # emulate the return value of $sth->execute for non-selects
1447 my $row_cnt = '0E0';
1449 my $subrs_cur = $rs->cursor;
1450 while (my @pks = $subrs_cur->next) {
1453 for my $i (0.. $#pcols) {
1454 $cond->{$pcols[$i]} = $pks[$i];
1459 $op eq 'update' ? $values : (),
1474 # localization is neccessary as
1475 # 1) there is no infrastructure to pass this around before SQLA2
1476 # 2) _select_args sets it and _prep_for_execute consumes it
1477 my $sql_maker = $self->sql_maker;
1478 local $sql_maker->{_dbic_rs_attrs};
1480 return $self->_execute($self->_select_args(@_));
1483 sub _select_args_to_query {
1486 # localization is neccessary as
1487 # 1) there is no infrastructure to pass this around before SQLA2
1488 # 2) _select_args sets it and _prep_for_execute consumes it
1489 my $sql_maker = $self->sql_maker;
1490 local $sql_maker->{_dbic_rs_attrs};
1492 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $order, $rows, $offset)
1493 # = $self->_select_args($ident, $select, $cond, $attrs);
1494 my ($op, $bind, $ident, $bind_attrs, @args) =
1495 $self->_select_args(@_);
1497 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $order, $rows, $offset ]);
1498 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1499 $prepared_bind ||= [];
1502 ? ($sql, $prepared_bind, $bind_attrs)
1503 : \[ "($sql)", @$prepared_bind ]
1508 my ($self, $ident, $select, $where, $attrs) = @_;
1510 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1512 my $sql_maker = $self->sql_maker;
1513 $sql_maker->{_dbic_rs_attrs} = {
1519 ? ( _source_handle => $alias2source->{$rs_alias}->handle )
1524 # calculate bind_attrs before possible $ident mangling
1525 my $bind_attrs = {};
1526 for my $alias (keys %$alias2source) {
1527 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
1528 for my $col (keys %$bindtypes) {
1530 my $fqcn = join ('.', $alias, $col);
1531 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
1533 # Unqialified column names are nice, but at the same time can be
1534 # rather ambiguous. What we do here is basically go along with
1535 # the loop, adding an unqualified column slot to $bind_attrs,
1536 # alongside the fully qualified name. As soon as we encounter
1537 # another column by that name (which would imply another table)
1538 # we unset the unqualified slot and never add any info to it
1539 # to avoid erroneous type binding. If this happens the users
1540 # only choice will be to fully qualify his column name
1542 if (exists $bind_attrs->{$col}) {
1543 $bind_attrs->{$col} = {};
1546 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
1553 $attrs->{software_limit}
1555 $sql_maker->_default_limit_syntax eq "GenericSubQ"
1557 $attrs->{software_limit} = 1;
1560 $self->throw_exception("rows attribute must be positive if present")
1561 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1563 # MySQL actually recommends this approach. I cringe.
1564 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1569 # see if we need to tear the prefetch apart (either limited has_many or grouped prefetch)
1570 # otherwise delegate the limiting to the storage, unless software limit was requested
1572 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
1574 ( $attrs->{group_by} && @{$attrs->{group_by}} &&
1575 $attrs->{_prefetch_select} && @{$attrs->{_prefetch_select}} )
1577 ($ident, $select, $where, $attrs)
1578 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
1580 elsif (! $attrs->{software_limit} ) {
1581 push @limit, $attrs->{rows}, $attrs->{offset};
1585 # This would be the point to deflate anything found in $where
1586 # (and leave $attrs->{bind} intact). Problem is - inflators historically
1587 # expect a row object. And all we have is a resultsource (it is trivial
1588 # to extract deflator coderefs via $alias2source above).
1590 # I don't see a way forward other than changing the way deflators are
1591 # invoked, and that's just bad...
1595 { $attrs->{$_} ? ( $_ => $attrs->{$_} ) : () }
1596 (qw/order_by group_by having/ )
1599 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $order, @limit);
1603 # This is the code producing joined subqueries like:
1604 # SELECT me.*, other.* FROM ( SELECT me.* FROM ... ) JOIN other ON ...
1606 sub _adjust_select_args_for_complex_prefetch {
1607 my ($self, $from, $select, $where, $attrs) = @_;
1609 $self->throw_exception ('Nothing to prefetch... how did we get here?!')
1610 if not @{$attrs->{_prefetch_select}};
1612 $self->throw_exception ('Complex prefetches are not supported on resultsets with a custom from attribute')
1613 if (ref $from ne 'ARRAY' || ref $from->[0] ne 'HASH' || ref $from->[1] ne 'ARRAY');
1616 # generate inner/outer attribute lists, remove stuff that doesn't apply
1617 my $outer_attrs = { %$attrs };
1618 delete $outer_attrs->{$_} for qw/where bind rows offset group_by having/;
1620 my $inner_attrs = { %$attrs };
1621 delete $inner_attrs->{$_} for qw/for collapse _prefetch_select _collapse_order_by select as/;
1624 # bring over all non-collapse-induced order_by into the inner query (if any)
1625 # the outer one will have to keep them all
1626 delete $inner_attrs->{order_by};
1627 if (my $ord_cnt = @{$outer_attrs->{order_by}} - @{$outer_attrs->{_collapse_order_by}} ) {
1628 $inner_attrs->{order_by} = [
1629 @{$outer_attrs->{order_by}}[ 0 .. $ord_cnt - 1]
1634 # generate the inner/outer select lists
1635 # for inside we consider only stuff *not* brought in by the prefetch
1636 # on the outside we substitute any function for its alias
1637 my $outer_select = [ @$select ];
1638 my $inner_select = [];
1639 for my $i (0 .. ( @$outer_select - @{$outer_attrs->{_prefetch_select}} - 1) ) {
1640 my $sel = $outer_select->[$i];
1642 if (ref $sel eq 'HASH' ) {
1643 $sel->{-as} ||= $attrs->{as}[$i];
1644 $outer_select->[$i] = join ('.', $attrs->{alias}, ($sel->{-as} || "inner_column_$i") );
1647 push @$inner_select, $sel;
1650 # normalize a copy of $from, so it will be easier to work with further
1651 # down (i.e. promote the initial hashref to an AoH)
1653 $from->[0] = [ $from->[0] ];
1654 my %original_join_info = map { $_->[0]{-alias} => $_->[0] } (@$from);
1657 # decide which parts of the join will remain in either part of
1658 # the outer/inner query
1660 # First we compose a list of which aliases are used in restrictions
1661 # (i.e. conditions/order/grouping/etc). Since we do not have
1662 # introspectable SQLA, we fall back to ugly scanning of raw SQL for
1663 # WHERE, and for pieces of ORDER BY in order to determine which aliases
1664 # need to appear in the resulting sql.
1665 # It may not be very efficient, but it's a reasonable stop-gap
1666 # Also unqualified column names will not be considered, but more often
1667 # than not this is actually ok
1669 # In the same loop we enumerate part of the selection aliases, as
1670 # it requires the same sqla hack for the time being
1671 my ($restrict_aliases, $select_aliases, $prefetch_aliases);
1673 # produce stuff unquoted, so it can be scanned
1674 my $sql_maker = $self->sql_maker;
1675 local $sql_maker->{quote_char};
1676 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
1679 my $non_prefetch_select_sql = $sql_maker->_recurse_fields ($inner_select);
1680 my $prefetch_select_sql = $sql_maker->_recurse_fields ($outer_attrs->{_prefetch_select});
1681 my $where_sql = $sql_maker->where ($where);
1682 my $group_by_sql = $sql_maker->_order_by({
1683 map { $_ => $inner_attrs->{$_} } qw/group_by having/
1685 my @non_prefetch_order_by_chunks = (map
1686 { ref $_ ? $_->[0] : $_ }
1687 $sql_maker->_order_by_chunks ($inner_attrs->{order_by})
1691 for my $alias (keys %original_join_info) {
1692 my $seen_re = qr/\b $alias $sep/x;
1694 for my $piece ($where_sql, $group_by_sql, @non_prefetch_order_by_chunks ) {
1695 if ($piece =~ $seen_re) {
1696 $restrict_aliases->{$alias} = 1;
1700 if ($non_prefetch_select_sql =~ $seen_re) {
1701 $select_aliases->{$alias} = 1;
1704 if ($prefetch_select_sql =~ $seen_re) {
1705 $prefetch_aliases->{$alias} = 1;
1711 # Add any non-left joins to the restriction list (such joins are indeed restrictions)
1712 for my $j (values %original_join_info) {
1713 my $alias = $j->{-alias} or next;
1714 $restrict_aliases->{$alias} = 1 if (
1715 (not $j->{-join_type})
1717 ($j->{-join_type} !~ /^left (?: \s+ outer)? $/xi)
1721 # mark all join parents as mentioned
1722 # (e.g. join => { cds => 'tracks' } - tracks will need to bring cds too )
1723 for my $collection ($restrict_aliases, $select_aliases) {
1724 for my $alias (keys %$collection) {
1725 $collection->{$_} = 1
1726 for (@{ $original_join_info{$alias}{-join_path} || [] });
1730 # construct the inner $from for the subquery
1731 my %inner_joins = (map { %{$_ || {}} } ($restrict_aliases, $select_aliases) );
1733 for my $j (@$from) {
1734 push @inner_from, $j if $inner_joins{$j->[0]{-alias}};
1737 # if a multi-type join was needed in the subquery ("multi" is indicated by
1738 # presence in {collapse}) - add a group_by to simulate the collapse in the subq
1739 unless ($inner_attrs->{group_by}) {
1740 for my $alias (keys %inner_joins) {
1742 # the dot comes from some weirdness in collapse
1743 # remove after the rewrite
1744 if ($attrs->{collapse}{".$alias"}) {
1745 $inner_attrs->{group_by} ||= $inner_select;
1751 # demote the inner_from head
1752 $inner_from[0] = $inner_from[0][0];
1754 # generate the subquery
1755 my $subq = $self->_select_args_to_query (
1762 my $subq_joinspec = {
1763 -alias => $attrs->{alias},
1764 -source_handle => $inner_from[0]{-source_handle},
1765 $attrs->{alias} => $subq,
1768 # Generate the outer from - this is relatively easy (really just replace
1769 # the join slot with the subquery), with a major caveat - we can not
1770 # join anything that is non-selecting (not part of the prefetch), but at
1771 # the same time is a multi-type relationship, as it will explode the result.
1773 # There are two possibilities here
1774 # - either the join is non-restricting, in which case we simply throw it away
1775 # - it is part of the restrictions, in which case we need to collapse the outer
1776 # result by tackling yet another group_by to the outside of the query
1778 # so first generate the outer_from, up to the substitution point
1780 while (my $j = shift @$from) {
1781 if ($j->[0]{-alias} eq $attrs->{alias}) { # time to swap
1786 last; # we'll take care of what's left in $from below
1789 push @outer_from, $j;
1793 # see what's left - throw away if not selecting/restricting
1794 # also throw in a group_by if restricting to guard against
1795 # cross-join explosions
1797 while (my $j = shift @$from) {
1798 my $alias = $j->[0]{-alias};
1800 if ($select_aliases->{$alias} || $prefetch_aliases->{$alias}) {
1801 push @outer_from, $j;
1803 elsif ($restrict_aliases->{$alias}) {
1804 push @outer_from, $j;
1806 # FIXME - this should be obviated by SQLA2, as I'll be able to
1807 # have restrict_inner and restrict_outer... or something to that
1808 # effect... I think...
1810 # FIXME2 - I can't find a clean way to determine if a particular join
1811 # is a multi - instead I am just treating everything as a potential
1812 # explosive join (ribasushi)
1814 # if (my $handle = $j->[0]{-source_handle}) {
1815 # my $rsrc = $handle->resolve;
1816 # ... need to bail out of the following if this is not a multi,
1817 # as it will be much easier on the db ...
1819 $outer_attrs->{group_by} ||= $outer_select;
1824 # demote the outer_from head
1825 $outer_from[0] = $outer_from[0][0];
1827 # This is totally horrific - the $where ends up in both the inner and outer query
1828 # Unfortunately not much can be done until SQLA2 introspection arrives, and even
1829 # then if where conditions apply to the *right* side of the prefetch, you may have
1830 # to both filter the inner select (e.g. to apply a limit) and then have to re-filter
1831 # the outer select to exclude joins you didin't want in the first place
1833 # OTOH it can be seen as a plus: <ash> (notes that this query would make a DBA cry ;)
1834 return (\@outer_from, $outer_select, $where, $outer_attrs);
1837 sub _resolve_ident_sources {
1838 my ($self, $ident) = @_;
1840 my $alias2source = {};
1843 # the reason this is so contrived is that $ident may be a {from}
1844 # structure, specifying multiple tables to join
1845 if ( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1846 # this is compat mode for insert/update/delete which do not deal with aliases
1847 $alias2source->{me} = $ident;
1850 elsif (ref $ident eq 'ARRAY') {
1854 if (ref $_ eq 'HASH') {
1856 $rs_alias = $tabinfo->{-alias};
1858 if (ref $_ eq 'ARRAY' and ref $_->[0] eq 'HASH') {
1862 $alias2source->{$tabinfo->{-alias}} = $tabinfo->{-source_handle}->resolve
1863 if ($tabinfo->{-source_handle});
1867 return ($alias2source, $rs_alias);
1870 # Takes $ident, \@column_names
1872 # returns { $column_name => \%column_info, ... }
1873 # also note: this adds -result_source => $rsrc to the column info
1876 # my $col_sources = $self->_resolve_column_info($ident, @column_names);
1877 sub _resolve_column_info {
1878 my ($self, $ident, $colnames) = @_;
1879 my ($alias2src, $root_alias) = $self->_resolve_ident_sources($ident);
1881 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
1884 my (%return, %seen_cols);
1886 # compile a global list of column names, to be able to properly
1887 # disambiguate unqualified column names (if at all possible)
1888 for my $alias (keys %$alias2src) {
1889 my $rsrc = $alias2src->{$alias};
1890 for my $colname ($rsrc->columns) {
1891 push @{$seen_cols{$colname}}, $alias;
1896 foreach my $col (@$colnames) {
1897 my ($alias, $colname) = $col =~ m/^ (?: ([^$sep]+) $sep)? (.+) $/x;
1900 # see if the column was seen exactly once (so we know which rsrc it came from)
1901 if ($seen_cols{$colname} and @{$seen_cols{$colname}} == 1) {
1902 $alias = $seen_cols{$colname}[0];
1909 my $rsrc = $alias2src->{$alias};
1910 $return{$col} = $rsrc && {
1911 %{$rsrc->column_info($colname)},
1912 -result_source => $rsrc,
1913 -source_alias => $alias,
1920 # Returns a counting SELECT for a simple count
1921 # query. Abstracted so that a storage could override
1922 # this to { count => 'firstcol' } or whatever makes
1923 # sense as a performance optimization
1925 #my ($self, $source, $rs_attrs) = @_;
1926 return { count => '*' };
1929 # Returns a SELECT which will end up in the subselect
1930 # There may or may not be a group_by, as the subquery
1931 # might have been called to accomodate a limit
1933 # Most databases would be happy with whatever ends up
1934 # here, but some choke in various ways.
1936 sub _subq_count_select {
1937 my ($self, $source, $rs_attrs) = @_;
1938 return $rs_attrs->{group_by} if $rs_attrs->{group_by};
1940 my @pcols = map { join '.', $rs_attrs->{alias}, $_ } ($source->primary_columns);
1941 return @pcols ? \@pcols : [ 1 ];
1945 sub source_bind_attributes {
1946 my ($self, $source) = @_;
1948 my $bind_attributes;
1949 foreach my $column ($source->columns) {
1951 my $data_type = $source->column_info($column)->{data_type} || '';
1952 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
1956 return $bind_attributes;
1963 =item Arguments: $ident, $select, $condition, $attrs
1967 Handle a SQL select statement.
1973 my ($ident, $select, $condition, $attrs) = @_;
1974 return $self->cursor_class->new($self, \@_, $attrs);
1979 my ($rv, $sth, @bind) = $self->_select(@_);
1980 my @row = $sth->fetchrow_array;
1981 my @nextrow = $sth->fetchrow_array if @row;
1982 if(@row && @nextrow) {
1983 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
1985 # Need to call finish() to work round broken DBDs
1994 =item Arguments: $sql
1998 Returns a L<DBI> sth (statement handle) for the supplied SQL.
2003 my ($self, $dbh, $sql) = @_;
2005 # 3 is the if_active parameter which avoids active sth re-use
2006 my $sth = $self->disable_sth_caching
2007 ? $dbh->prepare($sql)
2008 : $dbh->prepare_cached($sql, {}, 3);
2010 # XXX You would think RaiseError would make this impossible,
2011 # but apparently that's not true :(
2012 $self->throw_exception($dbh->errstr) if !$sth;
2018 my ($self, $sql) = @_;
2019 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2022 sub _dbh_columns_info_for {
2023 my ($self, $dbh, $table) = @_;
2025 if ($dbh->can('column_info')) {
2028 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2029 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2031 while ( my $info = $sth->fetchrow_hashref() ){
2033 $column_info{data_type} = $info->{TYPE_NAME};
2034 $column_info{size} = $info->{COLUMN_SIZE};
2035 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2036 $column_info{default_value} = $info->{COLUMN_DEF};
2037 my $col_name = $info->{COLUMN_NAME};
2038 $col_name =~ s/^\"(.*)\"$/$1/;
2040 $result{$col_name} = \%column_info;
2043 return \%result if !$@ && scalar keys %result;
2047 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2049 my @columns = @{$sth->{NAME_lc}};
2050 for my $i ( 0 .. $#columns ){
2052 $column_info{data_type} = $sth->{TYPE}->[$i];
2053 $column_info{size} = $sth->{PRECISION}->[$i];
2054 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2056 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2057 $column_info{data_type} = $1;
2058 $column_info{size} = $2;
2061 $result{$columns[$i]} = \%column_info;
2065 foreach my $col (keys %result) {
2066 my $colinfo = $result{$col};
2067 my $type_num = $colinfo->{data_type};
2069 if(defined $type_num && $dbh->can('type_info')) {
2070 my $type_info = $dbh->type_info($type_num);
2071 $type_name = $type_info->{TYPE_NAME} if $type_info;
2072 $colinfo->{data_type} = $type_name if $type_name;
2079 sub columns_info_for {
2080 my ($self, $table) = @_;
2081 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2084 =head2 last_insert_id
2086 Return the row id of the last insert.
2090 sub _dbh_last_insert_id {
2091 # All Storage's need to register their own _dbh_last_insert_id
2092 # the old SQLite-based method was highly inappropriate
2095 my $class = ref $self;
2096 $self->throw_exception (<<EOE);
2098 No _dbh_last_insert_id() method found in $class.
2099 Since the method of obtaining the autoincrement id of the last insert
2100 operation varies greatly between different databases, this method must be
2101 individually implemented for every storage class.
2105 sub last_insert_id {
2107 $self->_dbh_last_insert_id ($self->_dbh, @_);
2110 =head2 _native_data_type
2114 =item Arguments: $type_name
2118 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2119 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2120 L<::Sybase|DBIx::Class::Storage::DBI::Sybase>.
2122 The default implementation returns C<undef>, implement in your Storage driver if
2123 you need this functionality.
2125 Should map types from other databases to the native RDBMS type, for example
2126 C<VARCHAR2> to C<VARCHAR>.
2128 Types with modifiers should map to the underlying data type. For example,
2129 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2131 Composite types should map to the container type, for example
2132 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2136 sub _native_data_type {
2137 #my ($self, $data_type) = @_;
2143 Returns the database driver name.
2150 if (not $self->_driver_determined) {
2151 $self->_determine_driver;
2152 goto $self->can ('sqlt_type');
2155 $self->_get_dbh->{Driver}->{Name};
2158 =head2 bind_attribute_by_data_type
2160 Given a datatype from column info, returns a database specific bind
2161 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2162 let the database planner just handle it.
2164 Generally only needed for special case column types, like bytea in postgres.
2168 sub bind_attribute_by_data_type {
2172 =head2 is_datatype_numeric
2174 Given a datatype from column_info, returns a boolean value indicating if
2175 the current RDBMS considers it a numeric value. This controls how
2176 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2177 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2178 be performed instead of the usual C<eq>.
2182 sub is_datatype_numeric {
2183 my ($self, $dt) = @_;
2185 return 0 unless $dt;
2187 return $dt =~ /^ (?:
2188 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2193 =head2 create_ddl_dir (EXPERIMENTAL)
2197 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2201 Creates a SQL file based on the Schema, for each of the specified
2202 database engines in C<\@databases> in the given directory.
2203 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2205 Given a previous version number, this will also create a file containing
2206 the ALTER TABLE statements to transform the previous schema into the
2207 current one. Note that these statements may contain C<DROP TABLE> or
2208 C<DROP COLUMN> statements that can potentially destroy data.
2210 The file names are created using the C<ddl_filename> method below, please
2211 override this method in your schema if you would like a different file
2212 name format. For the ALTER file, the same format is used, replacing
2213 $version in the name with "$preversion-$version".
2215 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2216 The most common value for this would be C<< { add_drop_table => 1 } >>
2217 to have the SQL produced include a C<DROP TABLE> statement for each table
2218 created. For quoting purposes supply C<quote_table_names> and
2219 C<quote_field_names>.
2221 If no arguments are passed, then the following default values are assumed:
2225 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2227 =item version - $schema->schema_version
2229 =item directory - './'
2231 =item preversion - <none>
2235 By default, C<\%sqlt_args> will have
2237 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2239 merged with the hash passed in. To disable any of those features, pass in a
2240 hashref like the following
2242 { ignore_constraint_names => 0, # ... other options }
2245 Note that this feature is currently EXPERIMENTAL and may not work correctly
2246 across all databases, or fully handle complex relationships.
2248 WARNING: Please check all SQL files created, before applying them.
2252 sub create_ddl_dir {
2253 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2255 if(!$dir || !-d $dir) {
2256 carp "No directory given, using ./\n";
2259 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2260 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2262 my $schema_version = $schema->schema_version || '1.x';
2263 $version ||= $schema_version;
2266 add_drop_table => 1,
2267 ignore_constraint_names => 1,
2268 ignore_index_names => 1,
2272 $self->throw_exception(q{Can't create a ddl file without SQL::Translator 0.09003: '}
2273 . $self->_check_sqlt_message . q{'})
2274 if !$self->_check_sqlt_version;
2276 my $sqlt = SQL::Translator->new( $sqltargs );
2278 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2279 my $sqlt_schema = $sqlt->translate({ data => $schema })
2280 or $self->throw_exception ($sqlt->error);
2282 foreach my $db (@$databases) {
2284 $sqlt->{schema} = $sqlt_schema;
2285 $sqlt->producer($db);
2288 my $filename = $schema->ddl_filename($db, $version, $dir);
2289 if (-e $filename && ($version eq $schema_version )) {
2290 # if we are dumping the current version, overwrite the DDL
2291 carp "Overwriting existing DDL file - $filename";
2295 my $output = $sqlt->translate;
2297 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2300 if(!open($file, ">$filename")) {
2301 $self->throw_exception("Can't open $filename for writing ($!)");
2304 print $file $output;
2307 next unless ($preversion);
2309 require SQL::Translator::Diff;
2311 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2312 if(!-e $prefilename) {
2313 carp("No previous schema file found ($prefilename)");
2317 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2319 carp("Overwriting existing diff file - $difffile");
2325 my $t = SQL::Translator->new($sqltargs);
2330 or $self->throw_exception ($t->error);
2332 my $out = $t->translate( $prefilename )
2333 or $self->throw_exception ($t->error);
2335 $source_schema = $t->schema;
2337 $source_schema->name( $prefilename )
2338 unless ( $source_schema->name );
2341 # The "new" style of producers have sane normalization and can support
2342 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2343 # And we have to diff parsed SQL against parsed SQL.
2344 my $dest_schema = $sqlt_schema;
2346 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2347 my $t = SQL::Translator->new($sqltargs);
2352 or $self->throw_exception ($t->error);
2354 my $out = $t->translate( $filename )
2355 or $self->throw_exception ($t->error);
2357 $dest_schema = $t->schema;
2359 $dest_schema->name( $filename )
2360 unless $dest_schema->name;
2363 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2367 if(!open $file, ">$difffile") {
2368 $self->throw_exception("Can't write to $difffile ($!)");
2376 =head2 deployment_statements
2380 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2384 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2386 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2387 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2389 C<$directory> is used to return statements from files in a previously created
2390 L</create_ddl_dir> directory and is optional. The filenames are constructed
2391 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2393 If no C<$directory> is specified then the statements are constructed on the
2394 fly using L<SQL::Translator> and C<$version> is ignored.
2396 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2400 sub deployment_statements {
2401 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2402 $type ||= $self->sqlt_type;
2403 $version ||= $schema->schema_version || '1.x';
2405 my $filename = $schema->ddl_filename($type, $version, $dir);
2409 open($file, "<$filename")
2410 or $self->throw_exception("Can't open $filename ($!)");
2413 return join('', @rows);
2416 $self->throw_exception(q{Can't deploy without SQL::Translator 0.09003: '}
2417 . $self->_check_sqlt_message . q{'})
2418 if !$self->_check_sqlt_version;
2420 # sources needs to be a parser arg, but for simplicty allow at top level
2422 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2423 if exists $sqltargs->{sources};
2425 my $tr = SQL::Translator->new(
2426 producer => "SQL::Translator::Producer::${type}",
2428 parser => 'SQL::Translator::Parser::DBIx::Class',
2431 return $tr->translate;
2435 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2438 return if($line =~ /^--/);
2440 # next if($line =~ /^DROP/m);
2441 return if($line =~ /^BEGIN TRANSACTION/m);
2442 return if($line =~ /^COMMIT/m);
2443 return if $line =~ /^\s+$/; # skip whitespace only
2444 $self->_query_start($line);
2446 # do a dbh_do cycle here, as we need some error checking in
2447 # place (even though we will ignore errors)
2448 $self->dbh_do (sub { $_[1]->do($line) });
2451 carp qq{$@ (running "${line}")};
2453 $self->_query_end($line);
2455 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2456 if (@statements > 1) {
2457 foreach my $statement (@statements) {
2458 $deploy->( $statement );
2461 elsif (@statements == 1) {
2462 foreach my $line ( split(";\n", $statements[0])) {
2468 =head2 datetime_parser
2470 Returns the datetime parser class
2474 sub datetime_parser {
2476 return $self->{datetime_parser} ||= do {
2477 $self->_populate_dbh unless $self->_dbh;
2478 $self->build_datetime_parser(@_);
2482 =head2 datetime_parser_type
2484 Defines (returns) the datetime parser class - currently hardwired to
2485 L<DateTime::Format::MySQL>
2489 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2491 =head2 build_datetime_parser
2493 See L</datetime_parser>
2497 sub build_datetime_parser {
2499 my $type = $self->datetime_parser_type(@_);
2501 $self->throw_exception("Couldn't load ${type}: $@") if $@;
2506 my $_check_sqlt_version; # private
2507 my $_check_sqlt_message; # private
2508 sub _check_sqlt_version {
2509 return $_check_sqlt_version if defined $_check_sqlt_version;
2510 eval 'use SQL::Translator "0.09003"';
2511 $_check_sqlt_message = $@ || '';
2512 $_check_sqlt_version = !$@;
2515 sub _check_sqlt_message {
2516 _check_sqlt_version if !defined $_check_sqlt_message;
2517 $_check_sqlt_message;
2521 =head2 is_replicating
2523 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2524 replicate from a master database. Default is undef, which is the result
2525 returned by databases that don't support replication.
2529 sub is_replicating {
2534 =head2 lag_behind_master
2536 Returns a number that represents a certain amount of lag behind a master db
2537 when a given storage is replicating. The number is database dependent, but
2538 starts at zero and increases with the amount of lag. Default in undef
2542 sub lag_behind_master {
2548 $self->_verify_pid if $self->_dbh;
2550 # some databases need this to stop spewing warnings
2551 if (my $dbh = $self->_dbh) {
2552 eval { $dbh->disconnect };
2562 =head2 DBIx::Class and AutoCommit
2564 DBIx::Class can do some wonderful magic with handling exceptions,
2565 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2566 (the default) combined with C<txn_do> for transaction support.
2568 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2569 in an assumed transaction between commits, and you're telling us you'd
2570 like to manage that manually. A lot of the magic protections offered by
2571 this module will go away. We can't protect you from exceptions due to database
2572 disconnects because we don't know anything about how to restart your
2573 transactions. You're on your own for handling all sorts of exceptional
2574 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2580 Matt S. Trout <mst@shadowcatsystems.co.uk>
2582 Andy Grundman <andy@hybridized.org>
2586 You may distribute this code under the same terms as Perl itself.