1 package DBIx::DBSchema;
4 use vars qw($VERSION $DEBUG $errstr);
6 use DBIx::DBSchema::_util qw(_load_driver _dbh);
7 use DBIx::DBSchema::Table 0.04;
8 use DBIx::DBSchema::Index;
9 use DBIx::DBSchema::Column;
10 use DBIx::DBSchema::ColGroup::Unique;
11 use DBIx::DBSchema::ColGroup::Index;
14 $VERSION = eval $VERSION; # modperlstyle: convert the string into a number
20 DBIx::DBSchema - Database-independent schema objects
26 $schema = new DBIx::DBSchema @dbix_dbschema_table_objects;
27 $schema = new_odbc DBIx::DBSchema $dbh;
28 $schema = new_odbc DBIx::DBSchema $dsn, $user, $pass;
29 $schema = new_native DBIx::DBSchema $dbh;
30 $schema = new_native DBIx::DBSchema $dsn, $user, $pass;
32 $schema->save("filename");
33 $schema = load DBIx::DBSchema "filename" or die $DBIx::DBSchema::errstr;
35 $schema->addtable($dbix_dbschema_table_object);
37 @table_names = $schema->tables;
39 $DBIx_DBSchema_table_object = $schema->table("table_name");
41 @sql = $schema->sql($dbh);
42 @sql = $schema->sql($dsn, $username, $password);
43 @sql = $schema->sql($dsn); #doesn't connect to database - less reliable
45 $perl_code = $schema->pretty_print;
46 %hash = eval $perl_code;
47 use DBI qw(:sql_types); $schema = pretty_read DBIx::DBSchema \%hash;
51 DBIx::DBSchema objects are collections of DBIx::DBSchema::Table objects and
52 represent a database schema.
54 This module implements an OO-interface to database schemas. Using this module,
55 you can create a database schema with an OO Perl interface. You can read the
56 schema from an existing database. You can save the schema to disk and restore
57 it in a different process. You can write SQL CREATE statements statements for
58 different databases from a single source. In recent versions, you can
59 transform one schema to another, adding any necessary new columns and tables
60 (and, as of 0.33, indices).
62 Currently supported databases are MySQL, PostgreSQL and SQLite. Sybase and
63 Oracle drivers are partially implemented. DBIx::DBSchema will attempt to use
64 generic SQL syntax for other databases. Assistance adding support for other
65 databases is welcomed. See L<DBIx::DBSchema::DBD>, "Driver Writer's Guide and
72 =item new TABLE_OBJECT, TABLE_OBJECT, ...
74 Creates a new DBIx::DBSchema object.
79 my($proto, @tables) = @_;
80 my %tables = map { $_->name, $_ } @tables; #check for duplicates?
82 my $class = ref($proto) || $proto;
87 bless ($self, $class);
91 =item new_odbc DATABASE_HANDLE | DATA_SOURCE USERNAME PASSWORD [ ATTR ]
93 Creates a new DBIx::DBSchema object from an existing data source, which can be
94 specified by passing an open DBI database handle, or by passing the DBI data
95 source name, username, and password. This uses the experimental DBI type_info
96 method to create a schema with standard (ODBC) SQL column types that most
97 closely correspond to any non-portable column types. Use this to import a
98 schema that you wish to use with many different database engines. Although
99 primary key and (unique) index information will only be read from databases
100 with DBIx::DBSchema::DBD drivers (currently MySQL and PostgreSQL), import of
101 column names and attributes *should* work for any database. Note that this
102 method only uses "ODBC" column types; it does not require or use an ODBC
108 my($proto, $dbh) = ( shift, _dbh(@_) );
110 map { new_odbc DBIx::DBSchema::Table $dbh, $_ } _tables_from_dbh($dbh)
114 =item new_native DATABASE_HANDLE | DATA_SOURCE USERNAME PASSWORD [ ATTR ]
116 Creates a new DBIx::DBSchema object from an existing data source, which can be
117 specified by passing an open DBI database handle, or by passing the DBI data
118 source name, username and password. This uses database-native methods to read
119 the schema, and will preserve any non-portable column types. The method is
120 only available if there is a DBIx::DBSchema::DBD for the corresponding database engine (currently, MySQL and PostgreSQL).
125 my($proto, $dbh) = (shift, _dbh(@_) );
127 map { new_native DBIx::DBSchema::Table ( $dbh, $_ ) } _tables_from_dbh($dbh)
133 Loads a DBIx::DBSchema object from a file. If there is an error, returns
134 false and puts an error message in $DBIx::DBSchema::errstr;
139 my($proto,$file)=@_; #use $proto ?
144 eval { $self = Storable::retrieve($file); };
146 if ( $@ && $@ =~ /not.*storable/i ) { #then try FreezeThaw
149 eval "use FreezeThaw;";
154 or do { $errstr = "Can't open $file: $!"; return ''; };
155 my $string = join('',<FILE>);
157 or do { $errstr = "Can't close $file: $!"; return ''; };
158 ($self) = FreezeThaw::thaw($string);
172 Saves a DBIx::DBSchema object to a file.
177 #my($self, $file) = @_;
178 Storable::nstore(@_);
181 =item addtable TABLE_OBJECT
183 Adds the given DBIx::DBSchema::Table object to this DBIx::DBSchema.
189 $self->{'tables'}->{$table->name} = $table; #check for dupliates?
194 Returns a list of the names of all tables.
200 keys %{$self->{'tables'}};
203 =item table TABLENAME
205 Returns the specified DBIx::DBSchema::Table object.
211 $self->{'tables'}->{$table};
214 =item sql [ DATABASE_HANDLE | DATA_SOURCE [ USERNAME PASSWORD [ ATTR ] ] ]
216 Returns a list of SQL `CREATE' statements for this schema.
218 The data source can be specified by passing an open DBI database handle, or by
219 passing the DBI data source name, username and password.
221 Although the username and password are optional, it is best to call this method
222 with a database handle or data source including a valid username and password -
223 a DBI connection will be opened and the quoting and type mapping will be more
226 If passed a DBI data source (or handle) such as `DBI:mysql:database' or
227 `DBI:Pg:dbname=database', will use syntax specific to that database engine.
228 Currently supported databases are MySQL and PostgreSQL.
230 If not passed a data source (or handle), or if there is no driver for the
231 specified database, will attempt to use generic SQL syntax.
236 my($self, $dbh) = ( shift, _dbh(@_) );
237 map { $self->table($_)->sql_create_table($dbh); } $self->tables;
240 =item sql_update_schema PROTOTYPE_SCHEMA [ DATABASE_HANDLE | DATA_SOURCE [ USERNAME PASSWORD [ ATTR ] ] ]
242 Returns a list of SQL statements to update this schema so that it is idential
243 to the provided prototype schema, also a DBIx::DBSchema object.
245 #Optionally, the data source can be specified by passing an open DBI database
246 #handle, or by passing the DBI data source name, username and password.
248 #If passed a DBI data source (or handle) such as `DBI:mysql:database' or
249 #`DBI:Pg:dbname=database', will use syntax specific to that database engine.
250 #Currently supported databases are MySQL and PostgreSQL.
252 #If not passed a data source (or handle), or if there is no driver for the
253 #specified database, will attempt to use generic SQL syntax.
255 Right now this method knows how to add new tables and alter existing tables.
256 It doesn't know how to drop tables yet.
258 See L<DBIx::DBSchema::Table/sql_alter_table>,
259 L<DBIx::DBSchema::Column/sql_add_coumn> and
260 L<DBIx::DBSchema::Column/sql_alter_column> for additional specifics and
265 #gosh, false laziness w/DBSchema::Table::sql_alter_schema
267 sub sql_update_schema {
268 my($self, $new, $dbh) = ( shift, shift, _dbh(@_) );
272 foreach my $table ( $new->tables ) {
274 if ( $self->table($table) ) {
276 warn "$table exists\n" if $DEBUG > 1;
279 $self->table($table)->sql_alter_table( $new->table($table), $dbh );
283 warn "table $table does not exist.\n" if $DEBUG;
286 $new->table($table)->sql_create_table( $dbh );
292 # should eventually drop tables not in $new
294 warn join("\n", @r). "\n"
301 =item update_schema PROTOTYPE_SCHEMA, DATABASE_HANDLE | DATA_SOURCE [ USERNAME PASSWORD [ ATTR ] ]
303 Same as sql_update_schema, except actually runs the SQL commands to update
304 the schema. Throws a fatal error if any statement fails.
309 my($self, $new, $dbh) = ( shift, shift, _dbh(@_) );
311 foreach my $statement ( $self->sql_update_schema( $new, $dbh ) ) {
312 $dbh->do( $statement )
313 or die "Error: ". $dbh->errstr. "\n executing: $statement";
320 Returns the data in this schema as Perl source, suitable for assigning to a
331 my $table = $self->table($tablename);
332 my %indices = $table->indices;
334 "'$tablename' => {\n".
337 #cant because -w complains about , in qw()
338 # (also biiiig problems with empty lengths)
340 #$table->column($_)->type. " ".
341 #( $table->column($_)->null ? 'NULL' : 0 ). " ".
342 #$table->column($_)->length. " ),\n"
344 "'". $table->column($_)->type. "', ".
345 "'". $table->column($_)->null. "', ".
346 "'". $table->column($_)->length. "', ".
347 "'". $table->column($_)->default. "', ".
348 "'". $table->column($_)->local. "',\n"
352 " 'primary_key' => '". $table->primary_key. "',\n".
354 #old style index representation..
357 $table->{'unique'} # $table->unique
358 ? " 'unique' => [ ". join(', ',
359 map { "[ '". join("', '", @{$_}). "' ]" }
360 @{$table->unique->lol_ref}
365 ( $table->{'index'} # $table->index
366 ? " 'index' => [ ". join(', ',
367 map { "[ '". join("', '", @{$_}). "' ]" }
368 @{$table->index->lol_ref}
374 " 'indices' => { ". join( ",\n ",
376 map { my $iname = $_;
377 my $index = $indices{$iname};
380 ? " 'using' => '". $index->using ."',\n"
383 " 'unique' => ". $index->unique .",\n".
385 join("', '", @{$index->columns} ).
399 =item pretty_read HASHREF
401 Creates a schema as specified by a data structure such as that created by
402 B<pretty_print> method.
407 my($proto, $href) = @_;
409 my $schema = $proto->new( map {
412 my $info = $href->{$tablename};
415 while ( @{$info->{'columns'}} ) {
416 push @columns, DBIx::DBSchema::Column->new(
417 splice @{$info->{'columns'}}, 0, 6
421 DBIx::DBSchema::Table->new({
422 'name' => $tablename,
423 'primary_key' => $info->{'primary_key'},
424 'columns' => \@columns,
427 'unique' => DBIx::DBSchema::ColGroup::Unique->new($info->{'unique'}),
428 'index' => DBIx::DBSchema::ColGroup::Index->new($info->{'index'}),
431 'indices' => [ map { my $idx_info = $info->{'indices'}{$_};
432 DBIx::DBSchema::Index->new({
435 'unique' => $idx_info->{'unique'},
436 'columns' => $idx_info->{'columns'},
439 keys %{ $info->{'indices'} }
447 # private subroutines
449 sub _tables_from_dbh {
451 my $driver = _load_driver($dbh);
453 scalar(eval "DBIx::DBSchema::DBD::$driver->default_db_catalog");
455 scalar(eval "DBIx::DBSchema::DBD::$driver->default_db_schema");
456 my $sth = $dbh->table_info($db_catalog, $db_schema, '', 'TABLE')
458 #map { $_->{TABLE_NAME} } grep { $_->{TABLE_TYPE} eq 'TABLE' }
459 # @{ $sth->fetchall_arrayref({ TABLE_NAME=>1, TABLE_TYPE=>1}) };
460 map { $_->[0] } grep { $_->[1] =~ /^TABLE$/i }
461 @{ $sth->fetchall_arrayref([2,3]) };
468 Ivan Kohler <ivan-dbix-dbschema@420.am>
470 Charles Shapiro <charles.shapiro@numethods.com> and Mitchell Friedman
471 <mitchell.friedman@numethods.com> contributed the start of a Sybase driver.
473 Daniel Hanks <hanksdc@about-inc.com> contributed the Oracle driver.
475 Jesse Vincent contributed the SQLite driver.
479 Contributions are welcome! I'm especially keen on any interest in the top
480 items/projects below under BUGS.
484 Copyright (c) 2000-2007 Ivan Kohler
485 Copyright (c) 2000 Mail Abuse Prevention System LLC
486 Copyright (c) 2007 Freeside Internet Services, Inc.
488 This program is free software; you can redistribute it and/or modify it under
489 the same terms as Perl itself.
493 Multiple primary keys are not yet supported.
495 Foreign keys and other constraints are not yet supported.
497 Eventually it would be nice to have additional transformations (deleted,
498 modified columns, deleted tables). sql_update_schema doesn't drop tables
499 or deal with deleted or modified columns yet.
501 Need to port and test with additional databases
503 On schema updates, index changes are not as efficent as they could be,
504 especially with large data sets. Specifically, we don't currently recognize
505 existing indices with different/"wrong" names that we could use "ALTER INDEX
506 name RENAME TO new_name" on, and instead drop the "wrongly named" index and
507 re-build a new one. Since these are indices and not columns, its not a huge
508 deal, everything turns out right in the end, though inefficient.
510 Each DBIx::DBSchema object should have a name which corresponds to its name
511 within the SQL database engine (DBI data source).
513 pretty_print is actually pretty ugly.
515 pretty_print isn't so good about quoting values... save/load is a much better
516 alternative to using pretty_print/pretty_read
518 pretty_read is pretty ugly too.
520 pretty_read should *not* create and pass in old-style unique/index indices
521 when nothing is given in the read.
523 Perhaps pretty_read should eval column types so that we can use DBI
524 qw(:sql_types) here instead of externally.
526 Need to support "using" index attribute in pretty_read and in reverse
529 perhaps we should just get rid of pretty_read entirely. pretty_print is useful
530 for debugging, but pretty_read is pretty bunk.
532 sql CREATE TABLE output should convert integers
533 (i.e. use DBI qw(:sql_types);) to local types using DBI->type_info plus a hash
538 L<DBIx::DBSchema::Table>, L<DBIx::DBSchema::Index>,
539 L<DBIx::DBSchema::Column>, L<DBIx::DBSchema::DBD>,
540 L<DBIx::DBSchema::DBD::mysql>, L<DBIx::DBSchema::DBD::Pg>, L<FS::Record>,