7 # Multi-level database module for storing hash trees, arrays and simple
8 # key/value pairs into FTP-able, cross-platform binary database files.
10 # Type `perldoc DBM::Deep` for complete documentation.
14 # tie %db, 'DBM::Deep', 'my_database.db'; # standard tie() method
16 # my $db = new DBM::Deep( 'my_database.db' ); # preferred OO method
18 # $db->{my_scalar} = 'hello world';
19 # $db->{my_hash} = { larry => 'genius', hashes => 'fast' };
20 # $db->{my_array} = [ 1, 2, 3, time() ];
21 # $db->{my_complex} = [ 'hello', { perl => 'rules' }, 42, 99 ];
22 # push @{$db->{my_array}}, 'another value';
23 # my @key_list = keys %{$db->{my_hash}};
24 # print "This module " . $db->{my_complex}->[1]->{perl} . "!\n";
27 # (c) 2002-2006 Joseph Huckaby. All Rights Reserved.
28 # This program is free software; you can redistribute it and/or
29 # modify it under the same terms as Perl itself.
37 our $VERSION = q(0.99_03);
39 use Fcntl qw( :DEFAULT :flock :seek );
41 use FileHandle::Fmode ();
44 use DBM::Deep::Engine;
48 # Setup constants for users to pass to new()
50 sub TYPE_HASH () { DBM::Deep::Engine->SIG_HASH }
51 sub TYPE_ARRAY () { DBM::Deep::Engine->SIG_ARRAY }
59 $proto->_throw_error( "Odd number of parameters to " . (caller(1))[2] );
64 unless ( eval { local $SIG{'__DIE__'}; %{$_[0]} || 1 } ) {
65 $proto->_throw_error( "Not a hashref in args to " . (caller(1))[2] );
70 $args = { file => shift };
78 # Class constructor method for Perl OO interface.
79 # Calls tie() and returns blessed reference to tied hash or array,
80 # providing a hybrid OO/tie interface.
83 my $args = $class->_get_args( @_ );
86 # Check if we want a tied hash or array.
89 if (defined($args->{type}) && $args->{type} eq TYPE_ARRAY) {
90 $class = 'DBM::Deep::Array';
91 require DBM::Deep::Array;
92 tie @$self, $class, %$args;
95 $class = 'DBM::Deep::Hash';
96 require DBM::Deep::Hash;
97 tie %$self, $class, %$args;
100 return bless $self, $class;
103 # This initializer is called from the various TIE* methods. new() calls tie(),
104 # which allows for a single point of entry.
109 $args->{fileobj} = DBM::Deep::File->new( $args )
110 unless exists $args->{fileobj};
112 # locking implicitly enables autoflush
113 if ($args->{locking}) { $args->{autoflush} = 1; }
115 # These are the defaults to be optionally overridden below
118 base_offset => undef,
125 $self->{engine} = DBM::Deep::Engine->new( { %{$args}, obj => $self } );
127 # Grab the parameters we want to use
128 foreach my $param ( keys %$self ) {
129 next unless exists $args->{$param};
130 $self->{$param} = $args->{$param};
133 $self->_engine->setup_fh( $self );
135 $self->{fileobj}->set_db( $self );
142 require DBM::Deep::Hash;
143 return DBM::Deep::Hash->TIEHASH( @_ );
148 require DBM::Deep::Array;
149 return DBM::Deep::Array->TIEARRAY( @_ );
153 my $self = shift->_get_self;
154 return $self->_fileobj->lock( $self, @_ );
158 my $self = shift->_get_self;
159 return $self->_fileobj->unlock( $self, @_ );
163 my $self = shift->_get_self;
164 my ($spot, $value) = @_;
169 elsif ( eval { local $SIG{__DIE__}; $value->isa( 'DBM::Deep' ) } ) {
170 ${$spot} = $value->_repr;
171 $value->_copy_node( ${$spot} );
174 my $r = Scalar::Util::reftype( $value );
175 my $c = Scalar::Util::blessed( $value );
176 if ( $r eq 'ARRAY' ) {
177 ${$spot} = [ @{$value} ];
180 ${$spot} = { %{$value} };
182 ${$spot} = bless ${$spot}, $c
190 die "Must be implemented in a child class\n";
194 die "Must be implemented in a child class\n";
199 # Recursively export into standard Perl hashes and arrays.
201 my $self = shift->_get_self;
203 my $temp = $self->_repr;
206 $self->_copy_node( $temp );
209 # This will always work because $self, after _get_self() is a HASH
210 if ( $self->{parent} ) {
211 my $c = Scalar::Util::blessed(
212 $self->{parent}->get($self->{parent_key})
214 if ( $c && !$c->isa( 'DBM::Deep' ) ) {
224 # Recursively import Perl hash/array structure
226 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
228 my $self = shift->_get_self;
231 # struct is not a reference, so just import based on our type
233 $struct = $self->_repr( @_ );
236 return $self->_import( $struct );
241 # Rebuild entire database into new file, then move
242 # it back on top of original.
244 my $self = shift->_get_self;
246 #XXX Need to create a new test for this
247 # if ($self->_fileobj->{links} > 1) {
248 # $self->_throw_error("Cannot optimize: reference count is greater than 1");
251 my $db_temp = DBM::Deep->new(
252 file => $self->_fileobj->{file} . '.tmp',
257 $self->_copy_node( $db_temp );
261 # Attempt to copy user, group and permissions over to new file
263 my @stats = stat($self->_fh);
264 my $perms = $stats[2] & 07777;
267 chown( $uid, $gid, $self->_fileobj->{file} . '.tmp' );
268 chmod( $perms, $self->_fileobj->{file} . '.tmp' );
270 # q.v. perlport for more information on this variable
271 if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) {
273 # Potential race condition when optmizing on Win32 with locking.
274 # The Windows filesystem requires that the filehandle be closed
275 # before it is overwritten with rename(). This could be redone
279 $self->_fileobj->close;
282 if (!rename $self->_fileobj->{file} . '.tmp', $self->_fileobj->{file}) {
283 unlink $self->_fileobj->{file} . '.tmp';
285 $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!");
289 $self->_fileobj->close;
290 $self->_fileobj->open;
291 $self->_engine->setup_fh( $self );
298 # Make copy of object and return
300 my $self = shift->_get_self;
302 return DBM::Deep->new(
303 type => $self->_type,
304 base_offset => $self->_base_offset,
305 fileobj => $self->_fileobj,
310 my %is_legal_filter = map {
313 store_key store_value
314 fetch_key fetch_value
319 # Setup filter function for storing or fetching the key or value
321 my $self = shift->_get_self;
325 if ( $is_legal_filter{$type} ) {
326 $self->_fileobj->{"filter_$type"} = $func;
335 my $self = shift->_get_self;
336 $self->_fileobj->begin_transaction;
341 my $self = shift->_get_self;
342 $self->_fileobj->end_transaction;
347 my $self = shift->_get_self;
348 $self->_fileobj->commit_transaction;
357 my $self = $_[0]->_get_self;
358 return $self->{engine};
362 my $self = $_[0]->_get_self;
363 return $self->{fileobj};
367 my $self = $_[0]->_get_self;
368 return $self->{type};
372 my $self = $_[0]->_get_self;
373 return $self->{base_offset};
377 my $self = $_[0]->_get_self;
378 return $self->_fileobj->{fh};
386 die "DBM::Deep: $_[1]\n";
393 #XXX This if() is redundant
394 if ( my $parent = $self->{parent} ) {
396 while ( $parent->{parent} ) {
398 $parent->_type eq TYPE_HASH
399 ? "\{q{$child->{parent_key}}\}"
400 : "\[$child->{parent_key}\]"
404 $parent = $parent->{parent};
407 $base = "\$db->get( q{$child->{parent_key}} )->" . $base;
410 $base = "\$db->get( q{$child->{parent_key}} )";
418 # Store single hash key/value or array element in database.
420 my $self = shift->_get_self;
421 my ($key, $value, $orig_key) = @_;
424 if ( !FileHandle::Fmode::is_W( $self->_fh ) ) {
425 $self->_throw_error( 'Cannot write to a readonly filehandle' );
428 #XXX The second condition needs to disappear
429 if ( defined $orig_key && !( $self->_type eq TYPE_ARRAY && $orig_key eq 'length') ) {
432 my $r = Scalar::Util::reftype( $value ) || '';
433 if ( $r eq 'HASH' ) {
436 elsif ( $r eq 'ARRAY' ) {
439 elsif ( defined $value ) {
446 if ( my $c = Scalar::Util::blessed( $value ) ) {
447 $rhs = "bless $rhs, '$c'";
450 my $lhs = $self->_find_parent;
452 if ( $self->_type eq TYPE_HASH ) {
453 $lhs .= "->\{q{$orig_key}\}";
456 $lhs .= "->\[$orig_key\]";
462 $lhs = "\$db->put(q{$orig_key},$rhs);";
465 $self->_fileobj->audit($lhs);
469 # Request exclusive lock for writing
471 $self->lock( LOCK_EX );
473 my $md5 = $self->_engine->{digest}->($key);
475 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5, { create => 1 } );
477 # User may be storing a hash, in which case we do not want it run
478 # through the filtering system
479 if ( !ref($value) && $self->_fileobj->{filter_store_value} ) {
480 $value = $self->_fileobj->{filter_store_value}->( $value );
484 # Add key/value to bucket list
486 $self->_engine->add_bucket( $tag, $md5, $key, $value, undef, $orig_key );
495 # Fetch single value or element given plain key or array index
497 my $self = shift->_get_self;
498 my ($key, $orig_key) = @_;
500 my $md5 = $self->_engine->{digest}->($key);
503 # Request shared lock for reading
505 $self->lock( LOCK_SH );
507 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );#, { create => 1 } );
508 #XXX This needs to autovivify
515 # Get value from bucket list
517 my $result = $self->_engine->get_bucket_value( $tag, $md5, $orig_key );
521 # Filters only apply to scalar values, so the ref check is making
522 # sure the fetched bucket is a scalar, not a child hash or array.
523 return ($result && !ref($result) && $self->_fileobj->{filter_fetch_value})
524 ? $self->_fileobj->{filter_fetch_value}->($result)
530 # Delete single key/value pair or element given plain key or array index
532 my $self = shift->_get_self;
533 my ($key, $orig_key) = @_;
535 if ( !FileHandle::Fmode::is_W( $self->_fh ) ) {
536 $self->_throw_error( 'Cannot write to a readonly filehandle' );
539 if ( defined $orig_key ) {
540 my $lhs = $self->_find_parent;
542 $self->_fileobj->audit( "delete $lhs;" );
545 $self->_fileobj->audit( "\$db->delete('$orig_key');" );
550 # Request exclusive lock for writing
552 $self->lock( LOCK_EX );
554 my $md5 = $self->_engine->{digest}->($key);
556 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );
565 my $value = $self->_engine->get_bucket_value( $tag, $md5 );
567 if (defined $value && !ref($value) && $self->_fileobj->{filter_fetch_value}) {
568 $value = $self->_fileobj->{filter_fetch_value}->($value);
571 my $result = $self->_engine->delete_bucket( $tag, $md5, $orig_key );
574 # If this object is an array and the key deleted was on the end of the stack,
575 # decrement the length variable.
585 # Check if a single key or element exists given plain key or array index
587 my $self = shift->_get_self;
590 my $md5 = $self->_engine->{digest}->($key);
593 # Request shared lock for reading
595 $self->lock( LOCK_SH );
597 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );
602 # For some reason, the built-in exists() function returns '' for false
608 # Check if bucket exists and return 1 or ''
610 my $result = $self->_engine->bucket_exists( $tag, $md5 ) || '';
619 # Clear all keys from hash, or all elements from array.
621 my $self = shift->_get_self;
623 if ( !FileHandle::Fmode::is_W( $self->_fh ) ) {
624 $self->_throw_error( 'Cannot write to a readonly filehandle' );
628 my $lhs = $self->_find_parent;
630 if ( $self->_type eq TYPE_HASH ) {
631 $lhs = '%{' . $lhs . '}';
634 $lhs = '@{' . $lhs . '}';
637 $self->_fileobj->audit( "$lhs = ();" );
641 # Request exclusive lock for writing
643 $self->lock( LOCK_EX );
645 if ( $self->_type eq TYPE_HASH ) {
646 my $key = $self->first_key;
648 my $next_key = $self->next_key( $key );
649 my $md5 = $self->_engine->{digest}->($key);
650 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );
651 $self->_engine->delete_bucket( $tag, $md5, $key );
656 my $size = $self->FETCHSIZE;
657 for my $key ( map { pack ( $self->_engine->{long_pack}, $_ ) } 0 .. $size - 1 ) {
658 my $md5 = $self->_engine->{digest}->($key);
659 my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );
660 $self->_engine->delete_bucket( $tag, $md5, $key );
662 $self->STORESIZE( 0 );
664 #XXX This needs updating to use _release_space
665 # $self->_engine->write_tag(
666 # $self->_base_offset, $self->_type,
667 # chr(0)x$self->_engine->{index_size},
676 # Public method aliases
678 sub put { (shift)->STORE( @_ ) }
679 sub store { (shift)->STORE( @_ ) }
680 sub get { (shift)->FETCH( @_ ) }
681 sub fetch { (shift)->FETCH( @_ ) }
682 sub delete { (shift)->DELETE( @_ ) }
683 sub exists { (shift)->EXISTS( @_ ) }
684 sub clear { (shift)->CLEAR( @_ ) }
691 DBM::Deep - A pure perl multi-level hash/array DBM
696 my $db = DBM::Deep->new( "foo.db" );
698 $db->{key} = 'value';
701 $db->put('key' => 'value');
702 print $db->get('key');
704 # true multi-level support
705 $db->{my_complex} = [
706 'hello', { perl => 'rules' },
710 tie my %db, 'DBM::Deep', 'foo.db';
714 tied(%db)->put('key' => 'value');
715 print tied(%db)->get('key');
719 A unique flat-file database module, written in pure perl. True multi-level
720 hash/array support (unlike MLDBM, which is faked), hybrid OO / tie()
721 interface, cross-platform FTPable files, ACID transactions, and is quite fast.
722 Can handle millions of keys and unlimited levels without significant
723 slow-down. Written from the ground-up in pure perl -- this is NOT a wrapper
724 around a C-based DBM. Out-of-the-box compatibility with Unix, Mac OS X and
727 =head1 VERSION DIFFERENCES
729 B<NOTE>: 0.99_01 and above have significant file format differences from 0.983 and
730 before. There will be a backwards-compatibility layer in 1.00, but that is
731 slated for a later 0.99_x release. This version is B<NOT> backwards compatible
732 with 0.983 and before.
736 Construction can be done OO-style (which is the recommended way), or using
737 Perl's tie() function. Both are examined here.
739 =head2 OO CONSTRUCTION
741 The recommended way to construct a DBM::Deep object is to use the new()
742 method, which gets you a blessed I<and> tied hash (or array) reference.
744 my $db = DBM::Deep->new( "foo.db" );
746 This opens a new database handle, mapped to the file "foo.db". If this
747 file does not exist, it will automatically be created. DB files are
748 opened in "r+" (read/write) mode, and the type of object returned is a
749 hash, unless otherwise specified (see L<OPTIONS> below).
751 You can pass a number of options to the constructor to specify things like
752 locking, autoflush, etc. This is done by passing an inline hash (or hashref):
754 my $db = DBM::Deep->new(
760 Notice that the filename is now specified I<inside> the hash with
761 the "file" parameter, as opposed to being the sole argument to the
762 constructor. This is required if any options are specified.
763 See L<OPTIONS> below for the complete list.
765 You can also start with an array instead of a hash. For this, you must
766 specify the C<type> parameter:
768 my $db = DBM::Deep->new(
770 type => DBM::Deep->TYPE_ARRAY
773 B<Note:> Specifing the C<type> parameter only takes effect when beginning
774 a new DB file. If you create a DBM::Deep object with an existing file, the
775 C<type> will be loaded from the file header, and an error will be thrown if
776 the wrong type is passed in.
778 =head2 TIE CONSTRUCTION
780 Alternately, you can create a DBM::Deep handle by using Perl's built-in
781 tie() function. The object returned from tie() can be used to call methods,
782 such as lock() and unlock(). (That object can be retrieved from the tied
783 variable at any time using tied() - please see L<perltie/> for more info.
786 my $db = tie %hash, "DBM::Deep", "foo.db";
789 my $db = tie @array, "DBM::Deep", "bar.db";
791 As with the OO constructor, you can replace the DB filename parameter with
792 a hash containing one or more options (see L<OPTIONS> just below for the
795 tie %hash, "DBM::Deep", {
803 There are a number of options that can be passed in when constructing your
804 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
810 Filename of the DB file to link the handle to. You can pass a full absolute
811 filesystem path, partial path, or a plain filename if the file is in the
812 current working directory. This is a required parameter (though q.v. fh).
816 If you want, you can pass in the fh instead of the file. This is most useful for doing
819 my $db = DBM::Deep->new( { fh => \*DATA } );
821 You are responsible for making sure that the fh has been opened appropriately for your
822 needs. If you open it read-only and attempt to write, an exception will be thrown. If you
823 open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
824 needs to read from the fh.
826 =item * audit_file / audit_fh
828 These are just like file/fh, except for auditing. Please see L</AUDITING> for
833 This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
834 not need to set this. However, it's there if you want it.
836 If you pass in fh and do not set this, it will be set appropriately.
840 This parameter specifies what type of object to create, a hash or array. Use
841 one of these two constants:
845 =item * C<DBM::Deep-E<gt>TYPE_HASH>
847 =item * C<DBM::Deep-E<gt>TYPE_ARRAY>.
851 This only takes effect when beginning a new file. This is an optional
852 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
856 Specifies whether locking is to be enabled. DBM::Deep uses Perl's flock()
857 function to lock the database in exclusive mode for writes, and shared mode
858 for reads. Pass any true value to enable. This affects the base DB handle
859 I<and any child hashes or arrays> that use the same DB file. This is an
860 optional parameter, and defaults to 0 (disabled). See L<LOCKING> below for
865 Specifies whether autoflush is to be enabled on the underlying filehandle.
866 This obviously slows down write operations, but is required if you may have
867 multiple processes accessing the same DB file (also consider enable I<locking>).
868 Pass any true value to enable. This is an optional parameter, and defaults to 0
873 If I<autobless> mode is enabled, DBM::Deep will preserve the class something
874 is blessed into, and restores it when fetched. This is an optional parameter, and defaults to 1 (enabled).
876 B<Note:> If you use the OO-interface, you will not be able to call any methods
877 of DBM::Deep on the blessed item. This is considered to be a feature.
881 See L</FILTERS> below.
887 With DBM::Deep you can access your databases using Perl's standard hash/array
888 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
889 treat them as such. DBM::Deep will intercept all reads/writes and direct them
890 to the right place -- the DB file. This has nothing to do with the
891 L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
892 using regular hashes and arrays, rather than calling functions like C<get()>
893 and C<put()> (although those work too). It is entirely up to you how to want
894 to access your databases.
898 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
899 or even nested hashes (or arrays) using standard Perl syntax:
901 my $db = DBM::Deep->new( "foo.db" );
903 $db->{mykey} = "myvalue";
905 $db->{myhash}->{subkey} = "subvalue";
907 print $db->{myhash}->{subkey} . "\n";
909 You can even step through hash keys using the normal Perl C<keys()> function:
911 foreach my $key (keys %$db) {
912 print "$key: " . $db->{$key} . "\n";
915 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
916 pushes them onto an array, all before the loop even begins. If you have an
917 extremely large hash, this may exhaust Perl's memory. Instead, consider using
918 Perl's C<each()> function, which pulls keys/values one at a time, using very
921 while (my ($key, $value) = each %$db) {
922 print "$key: $value\n";
925 Please note that when using C<each()>, you should always pass a direct
926 hash reference, not a lookup. Meaning, you should B<never> do this:
929 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
931 This causes an infinite loop, because for each iteration, Perl is calling
932 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
933 it effectively keeps returning the first key over and over again. Instead,
934 assign a temporary variable to C<$db->{foo}>, then pass that to each().
938 As with hashes, you can treat any DBM::Deep object like a normal Perl array
939 reference. This includes inserting, removing and manipulating elements,
940 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
941 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
942 or simply be a nested array reference inside a hash. Example:
944 my $db = DBM::Deep->new(
945 file => "foo-array.db",
946 type => DBM::Deep->TYPE_ARRAY
950 push @$db, "bar", "baz";
953 my $last_elem = pop @$db; # baz
954 my $first_elem = shift @$db; # bah
955 my $second_elem = $db->[1]; # bar
957 my $num_elements = scalar @$db;
961 In addition to the I<tie()> interface, you can also use a standard OO interface
962 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
963 array) has its own methods, but both types share the following common methods:
964 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>. C<fetch()> and
965 C<store(> are aliases to C<put()> and C<get()>, respectively.
969 =item * new() / clone()
971 These are the constructor and copy-functions.
973 =item * put() / store()
975 Stores a new hash key/value pair, or sets an array element value. Takes two
976 arguments, the hash key or array index, and the new value. The value can be
977 a scalar, hash ref or array ref. Returns true on success, false on failure.
979 $db->put("foo", "bar"); # for hashes
980 $db->put(1, "bar"); # for arrays
982 =item * get() / fetch()
984 Fetches the value of a hash key or array element. Takes one argument: the hash
985 key or array index. Returns a scalar, hash ref or array ref, depending on the
988 my $value = $db->get("foo"); # for hashes
989 my $value = $db->get(1); # for arrays
993 Checks if a hash key or array index exists. Takes one argument: the hash key
994 or array index. Returns true if it exists, false if not.
996 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
997 if ($db->exists(1)) { print "yay!\n"; } # for arrays
1001 Deletes one hash key/value pair or array element. Takes one argument: the hash
1002 key or array index. Returns true on success, false if not found. For arrays,
1003 the remaining elements located after the deleted element are NOT moved over.
1004 The deleted element is essentially just undefined, which is exactly how Perl's
1005 internal arrays work. Please note that the space occupied by the deleted
1006 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
1007 below for details and workarounds.
1009 $db->delete("foo"); # for hashes
1010 $db->delete(1); # for arrays
1014 Deletes B<all> hash keys or array elements. Takes no arguments. No return
1015 value. Please note that the space occupied by the deleted keys/values or
1016 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
1017 details and workarounds.
1019 $db->clear(); # hashes or arrays
1021 =item * lock() / unlock()
1027 Recover lost disk space. This is important to do, especially if you use
1030 =item * import() / export()
1032 Data going in and out.
1038 For hashes, DBM::Deep supports all the common methods described above, and the
1039 following additional methods: C<first_key()> and C<next_key()>.
1045 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
1046 fetched in an undefined order (which appears random). Takes no arguments,
1047 returns the key as a scalar value.
1049 my $key = $db->first_key();
1053 Returns the "next" key in the hash, given the previous one as the sole argument.
1054 Returns undef if there are no more keys to be fetched.
1056 $key = $db->next_key($key);
1060 Here are some examples of using hashes:
1062 my $db = DBM::Deep->new( "foo.db" );
1064 $db->put("foo", "bar");
1065 print "foo: " . $db->get("foo") . "\n";
1067 $db->put("baz", {}); # new child hash ref
1068 $db->get("baz")->put("buz", "biz");
1069 print "buz: " . $db->get("baz")->get("buz") . "\n";
1071 my $key = $db->first_key();
1073 print "$key: " . $db->get($key) . "\n";
1074 $key = $db->next_key($key);
1077 if ($db->exists("foo")) { $db->delete("foo"); }
1081 For arrays, DBM::Deep supports all the common methods described above, and the
1082 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
1083 C<unshift()> and C<splice()>.
1089 Returns the number of elements in the array. Takes no arguments.
1091 my $len = $db->length();
1095 Adds one or more elements onto the end of the array. Accepts scalars, hash
1096 refs or array refs. No return value.
1098 $db->push("foo", "bar", {});
1102 Fetches the last element in the array, and deletes it. Takes no arguments.
1103 Returns undef if array is empty. Returns the element value.
1105 my $elem = $db->pop();
1109 Fetches the first element in the array, deletes it, then shifts all the
1110 remaining elements over to take up the space. Returns the element value. This
1111 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
1114 my $elem = $db->shift();
1118 Inserts one or more elements onto the beginning of the array, shifting all
1119 existing elements over to make room. Accepts scalars, hash refs or array refs.
1120 No return value. This method is not recommended with large arrays -- see
1121 <LARGE ARRAYS> below for details.
1123 $db->unshift("foo", "bar", {});
1127 Performs exactly like Perl's built-in function of the same name. See L<perldoc
1128 -f splice> for usage -- it is too complicated to document here. This method is
1129 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
1133 Here are some examples of using arrays:
1135 my $db = DBM::Deep->new(
1137 type => DBM::Deep->TYPE_ARRAY
1140 $db->push("bar", "baz");
1141 $db->unshift("foo");
1144 my $len = $db->length();
1145 print "length: $len\n"; # 4
1147 for (my $k=0; $k<$len; $k++) {
1148 print "$k: " . $db->get($k) . "\n";
1151 $db->splice(1, 2, "biz", "baf");
1153 while (my $elem = shift @$db) {
1154 print "shifted: $elem\n";
1159 Enable automatic file locking by passing a true value to the C<locking>
1160 parameter when constructing your DBM::Deep object (see L<SETUP> above).
1162 my $db = DBM::Deep->new(
1167 This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
1168 mode for writes, and shared mode for reads. This is required if you have
1169 multiple processes accessing the same database file, to avoid file corruption.
1170 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
1171 NFS> below for more.
1173 =head2 EXPLICIT LOCKING
1175 You can explicitly lock a database, so it remains locked for multiple
1176 transactions. This is done by calling the C<lock()> method, and passing an
1177 optional lock mode argument (defaults to exclusive mode). This is particularly
1178 useful for things like counters, where the current value needs to be fetched,
1179 then incremented, then stored again.
1182 my $counter = $db->get("counter");
1184 $db->put("counter", $counter);
1193 You can pass C<lock()> an optional argument, which specifies which mode to use
1194 (exclusive or shared). Use one of these two constants:
1195 C<DBM::Deep-E<gt>LOCK_EX> or C<DBM::Deep-E<gt>LOCK_SH>. These are passed
1196 directly to C<flock()>, and are the same as the constants defined in Perl's
1199 $db->lock( $db->LOCK_SH );
1203 =head1 IMPORTING/EXPORTING
1205 You can import existing complex structures by calling the C<import()> method,
1206 and export an entire database into an in-memory structure using the C<export()>
1207 method. Both are examined here.
1211 Say you have an existing hash with nested hashes/arrays inside it. Instead of
1212 walking the structure and adding keys/elements to the database as you go,
1213 simply pass a reference to the C<import()> method. This recursively adds
1214 everything to an existing DBM::Deep object for you. Here is an example:
1219 array1 => [ "elem0", "elem1", "elem2" ],
1221 subkey1 => "subvalue1",
1222 subkey2 => "subvalue2"
1226 my $db = DBM::Deep->new( "foo.db" );
1227 $db->import( $struct );
1229 print $db->{key1} . "\n"; # prints "value1"
1231 This recursively imports the entire C<$struct> object into C<$db>, including
1232 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
1233 keys are merged with the existing ones, replacing if they already exist.
1234 The C<import()> method can be called on any database level (not just the base
1235 level), and works with both hash and array DB types.
1237 B<Note:> Make sure your existing structure has no circular references in it.
1238 These will cause an infinite loop when importing. There are plans to fix this
1243 Calling the C<export()> method on an existing DBM::Deep object will return
1244 a reference to a new in-memory copy of the database. The export is done
1245 recursively, so all nested hashes/arrays are all exported to standard Perl
1246 objects. Here is an example:
1248 my $db = DBM::Deep->new( "foo.db" );
1250 $db->{key1} = "value1";
1251 $db->{key2} = "value2";
1253 $db->{hash1}->{subkey1} = "subvalue1";
1254 $db->{hash1}->{subkey2} = "subvalue2";
1256 my $struct = $db->export();
1258 print $struct->{key1} . "\n"; # prints "value1"
1260 This makes a complete copy of the database in memory, and returns a reference
1261 to it. The C<export()> method can be called on any database level (not just
1262 the base level), and works with both hash and array DB types. Be careful of
1263 large databases -- you can store a lot more data in a DBM::Deep object than an
1264 in-memory Perl structure.
1266 B<Note:> Make sure your database has no circular references in it.
1267 These will cause an infinite loop when exporting. There are plans to fix this
1272 DBM::Deep has a number of hooks where you can specify your own Perl function
1273 to perform filtering on incoming or outgoing data. This is a perfect
1274 way to extend the engine, and implement things like real-time compression or
1275 encryption. Filtering applies to the base DB level, and all child hashes /
1276 arrays. Filter hooks can be specified when your DBM::Deep object is first
1277 constructed, or by calling the C<set_filter()> method at any time. There are
1278 four available filter hooks, described below:
1282 =item * filter_store_key
1284 This filter is called whenever a hash key is stored. It
1285 is passed the incoming key, and expected to return a transformed key.
1287 =item * filter_store_value
1289 This filter is called whenever a hash key or array element is stored. It
1290 is passed the incoming value, and expected to return a transformed value.
1292 =item * filter_fetch_key
1294 This filter is called whenever a hash key is fetched (i.e. via
1295 C<first_key()> or C<next_key()>). It is passed the transformed key,
1296 and expected to return the plain key.
1298 =item * filter_fetch_value
1300 This filter is called whenever a hash key or array element is fetched.
1301 It is passed the transformed value, and expected to return the plain value.
1305 Here are the two ways to setup a filter hook:
1307 my $db = DBM::Deep->new(
1309 filter_store_value => \&my_filter_store,
1310 filter_fetch_value => \&my_filter_fetch
1315 $db->set_filter( "filter_store_value", \&my_filter_store );
1316 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
1318 Your filter function will be called only when dealing with SCALAR keys or
1319 values. When nested hashes and arrays are being stored/fetched, filtering
1320 is bypassed. Filters are called as static functions, passed a single SCALAR
1321 argument, and expected to return a single SCALAR value. If you want to
1322 remove a filter, set the function reference to C<undef>:
1324 $db->set_filter( "filter_store_value", undef );
1326 =head2 REAL-TIME ENCRYPTION EXAMPLE
1328 Here is a working example that uses the I<Crypt::Blowfish> module to
1329 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
1330 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
1331 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
1334 use Crypt::Blowfish;
1337 my $cipher = Crypt::CBC->new({
1338 'key' => 'my secret key',
1339 'cipher' => 'Blowfish',
1341 'regenerate_key' => 0,
1342 'padding' => 'space',
1346 my $db = DBM::Deep->new(
1347 file => "foo-encrypt.db",
1348 filter_store_key => \&my_encrypt,
1349 filter_store_value => \&my_encrypt,
1350 filter_fetch_key => \&my_decrypt,
1351 filter_fetch_value => \&my_decrypt,
1354 $db->{key1} = "value1";
1355 $db->{key2} = "value2";
1356 print "key1: " . $db->{key1} . "\n";
1357 print "key2: " . $db->{key2} . "\n";
1363 return $cipher->encrypt( $_[0] );
1366 return $cipher->decrypt( $_[0] );
1369 =head2 REAL-TIME COMPRESSION EXAMPLE
1371 Here is a working example that uses the I<Compress::Zlib> module to do real-time
1372 compression / decompression of keys & values with DBM::Deep Filters.
1373 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
1374 more on I<Compress::Zlib>.
1379 my $db = DBM::Deep->new(
1380 file => "foo-compress.db",
1381 filter_store_key => \&my_compress,
1382 filter_store_value => \&my_compress,
1383 filter_fetch_key => \&my_decompress,
1384 filter_fetch_value => \&my_decompress,
1387 $db->{key1} = "value1";
1388 $db->{key2} = "value2";
1389 print "key1: " . $db->{key1} . "\n";
1390 print "key2: " . $db->{key2} . "\n";
1396 return Compress::Zlib::memGzip( $_[0] ) ;
1399 return Compress::Zlib::memGunzip( $_[0] ) ;
1402 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
1403 actually numerical index numbers, and are not filtered.
1405 =head1 ERROR HANDLING
1407 Most DBM::Deep methods return a true value for success, and call die() on
1408 failure. You can wrap calls in an eval block to catch the die.
1410 my $db = DBM::Deep->new( "foo.db" ); # create hash
1411 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
1413 print $@; # prints error message
1415 =head1 LARGEFILE SUPPORT
1417 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
1418 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
1419 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
1420 by specifying the 'pack_size' parameter when constructing the file.
1423 filename => $filename,
1424 pack_size => 'large',
1427 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
1428 instead of 32-bit longs. After setting these values your DB files have a
1429 theoretical maximum size of 16 XB (exabytes).
1431 You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file
1434 B<Note:> Changing these values will B<NOT> work for existing database files.
1435 Only change this for new files. Once the value has been set, it is stored in
1436 the file's header and cannot be changed for the life of the file. These
1437 parameters are per-file, meaning you can access 32-bit and 64-bit files, as
1440 B<Note:> We have not personally tested files larger than 2 GB -- all my
1441 systems have only a 32-bit Perl. However, I have received user reports that
1442 this does indeed work!
1444 =head1 LOW-LEVEL ACCESS
1446 If you require low-level access to the underlying filehandle that DBM::Deep uses,
1447 you can call the C<_fh()> method, which returns the handle:
1449 my $fh = $db->_fh();
1451 This method can be called on the root level of the datbase, or any child
1452 hashes or arrays. All levels share a I<root> structure, which contains things
1453 like the filehandle, a reference counter, and all the options specified
1454 when you created the object. You can get access to this file object by
1455 calling the C<_fileobj()> method.
1457 my $file_obj = $db->_fileobj();
1459 This is useful for changing options after the object has already been created,
1460 such as enabling/disabling locking. You can also store your own temporary user
1461 data in this structure (be wary of name collision), which is then accessible from
1462 any child hash or array.
1464 =head1 CUSTOM DIGEST ALGORITHM
1466 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
1467 keys. However you can override this, and use another algorithm (such as SHA-256)
1468 or even write your own. But please note that DBM::Deep currently expects zero
1469 collisions, so your algorithm has to be I<perfect>, so to speak. Collision
1470 detection may be introduced in a later version.
1472 You can specify a custom digest algorithm by passing it into the parameter
1473 list for new(), passing a reference to a subroutine as the 'digest' parameter,
1474 and the length of the algorithm's hashes (in bytes) as the 'hash_size'
1475 parameter. Here is a working example that uses a 256-bit hash from the
1476 I<Digest::SHA256> module. Please see
1477 L<http://search.cpan.org/search?module=Digest::SHA256> for more information.
1482 my $context = Digest::SHA256::new(256);
1484 my $db = DBM::Deep->new(
1485 filename => "foo-sha.db",
1486 digest => \&my_digest,
1490 $db->{key1} = "value1";
1491 $db->{key2} = "value2";
1492 print "key1: " . $db->{key1} . "\n";
1493 print "key2: " . $db->{key2} . "\n";
1499 return substr( $context->hash($_[0]), 0, 32 );
1502 B<Note:> Your returned digest strings must be B<EXACTLY> the number
1503 of bytes you specify in the hash_size parameter (in this case 32).
1505 B<Note:> If you do choose to use a custom digest algorithm, you must set it
1506 every time you access this file. Otherwise, the default (MD5) will be used.
1508 =head1 CIRCULAR REFERENCES
1510 DBM::Deep has B<experimental> support for circular references. Meaning you
1511 can have a nested hash key or array element that points to a parent object.
1512 This relationship is stored in the DB file, and is preserved between sessions.
1515 my $db = DBM::Deep->new( "foo.db" );
1518 $db->{circle} = $db; # ref to self
1520 print $db->{foo} . "\n"; # prints "bar"
1521 print $db->{circle}->{foo} . "\n"; # prints "bar" again
1523 B<Note>: Passing the object to a function that recursively walks the
1524 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
1525 C<export()> methods) will result in an infinite loop. This will be fixed in
1530 New in 0.99_01 is the ability to audit your databases actions. By passing in
1531 audit_file (or audit_fh) to the constructor, all actions will be logged to
1532 that file. The format is one that is suitable for eval'ing against the
1533 database to replay the actions. Please see t/33_audit_trail.t for an example
1538 New in 0.99_01 is ACID transactions. Every DBM::Deep object is completely
1539 transaction-ready - it is not an option you have to turn on. Three new methods
1540 have been added to support them. They are:
1544 =item * begin_work()
1546 This starts a transaction.
1550 This applies the changes done within the transaction to the mainline and ends
1555 This discards the changes done within the transaction to the mainline and ends
1560 Transactions in DBM::Deep are done using the MVCC method, the same method used
1561 by the InnoDB MySQL table type.
1563 =head1 CAVEATS / ISSUES / BUGS
1565 This section describes all the known issues with DBM::Deep. It you have found
1566 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
1568 =head2 UNUSED SPACE RECOVERY
1570 One major caveat with DBM::Deep is that space occupied by existing keys and
1571 values is not recovered when they are deleted. Meaning if you keep deleting
1572 and adding new keys, your file will continuously grow. I am working on this,
1573 but in the meantime you can call the built-in C<optimize()> method from time to
1574 time (perhaps in a crontab or something) to recover all your unused space.
1576 $db->optimize(); # returns true on success
1578 This rebuilds the ENTIRE database into a new file, then moves it on top of
1579 the original. The new file will have no unused space, thus it will take up as
1580 little disk space as possible. Please note that this operation can take
1581 a long time for large files, and you need enough disk space to temporarily hold
1582 2 copies of your DB file. The temporary file is created in the same directory
1583 as the original, named with a ".tmp" extension, and is deleted when the
1584 operation completes. Oh, and if locking is enabled, the DB is automatically
1585 locked for the entire duration of the copy.
1587 B<WARNING:> Only call optimize() on the top-level node of the database, and
1588 make sure there are no child references lying around. DBM::Deep keeps a reference
1589 counter, and if it is greater than 1, optimize() will abort and return undef.
1593 (The reasons given assume a high level of Perl understanding, specifically of
1594 references. You can safely skip this section.)
1596 Currently, the only references supported are HASH and ARRAY. The other reference
1597 types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons.
1603 These are things like filehandles and other sockets. They can't be supported
1604 because it's completely unclear how DBM::Deep should serialize them.
1606 =item * SCALAR / REF
1608 The discussion here refers to the following type of example:
1615 # In some other process ...
1617 my $val = ${ $db->{key1} };
1619 is( $val, 50, "What actually gets stored in the DB file?" );
1621 The problem is one of synchronization. When the variable being referred to
1622 changes value, the reference isn't notified. This means that the new value won't
1623 be stored in the datafile for other processes to read. There is no TIEREF.
1625 It is theoretically possible to store references to values already within a
1626 DBM::Deep object because everything already is synchronized, but the change to
1627 the internals would be quite large. Specifically, DBM::Deep would have to tie
1628 every single value that is stored. This would bloat the RAM footprint of
1629 DBM::Deep at least twofold (if not more) and be a significant performance drain,
1630 all to support a feature that has never been requested.
1634 L<Data::Dump::Streamer/> provides a mechanism for serializing coderefs,
1635 including saving off all closure state. However, just as for SCALAR and REF,
1636 that closure state may change without notifying the DBM::Deep object storing
1641 =head2 FILE CORRUPTION
1643 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
1644 for a 32-bit signature when opened, but other corruption in files can cause
1645 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
1646 stuck in an infinite loop depending on the level of corruption. File write
1647 operations are not checked for failure (for speed), so if you happen to run
1648 out of disk space, DBM::Deep will probably fail in a bad way. These things will
1649 be addressed in a later version of DBM::Deep.
1653 Beware of using DBM::Deep files over NFS. DBM::Deep uses flock(), which works
1654 well on local filesystems, but will NOT protect you from file corruption over
1655 NFS. I've heard about setting up your NFS server with a locking daemon, then
1656 using lockf() to lock your files, but your mileage may vary there as well.
1657 From what I understand, there is no real way to do it. However, if you need
1658 access to the underlying filehandle in DBM::Deep for using some other kind of
1659 locking scheme like lockf(), see the L<LOW-LEVEL ACCESS> section above.
1661 =head2 COPYING OBJECTS
1663 Beware of copying tied objects in Perl. Very strange things can happen.
1664 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
1665 returns a new, blessed, tied hash or array to the same level in the DB.
1667 my $copy = $db->clone();
1669 B<Note>: Since clone() here is cloning the object, not the database location, any
1670 modifications to either $db or $copy will be visible to both.
1674 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
1675 These functions cause every element in the array to move, which can be murder
1676 on DBM::Deep, as every element has to be fetched from disk, then stored again in
1677 a different location. This will be addressed in the forthcoming version 1.00.
1679 =head2 WRITEONLY FILES
1681 If you pass in a filehandle to new(), you may have opened it in either a readonly or
1682 writeonly mode. STORE will verify that the filehandle is writable. However, there
1683 doesn't seem to be a good way to determine if a filehandle is readable. And, if the
1684 filehandle isn't readable, it's not clear what will happen. So, don't do that.
1686 =head1 CODE COVERAGE
1688 B<Devel::Cover> is used to test the code coverage of the tests. Below is the
1689 B<Devel::Cover> report on this distribution's test suite.
1691 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1692 File stmt bran cond sub pod time total
1693 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1694 blib/lib/DBM/Deep.pm 96.2 89.0 75.0 95.8 89.5 36.0 92.9
1695 blib/lib/DBM/Deep/Array.pm 96.1 88.3 100.0 96.4 100.0 15.9 94.7
1696 blib/lib/DBM/Deep/Engine.pm 96.6 86.6 89.5 100.0 0.0 20.0 91.0
1697 blib/lib/DBM/Deep/File.pm 99.4 88.3 55.6 100.0 0.0 19.6 89.5
1698 blib/lib/DBM/Deep/Hash.pm 98.5 83.3 100.0 100.0 100.0 8.5 96.3
1699 Total 96.9 87.4 81.2 98.0 38.5 100.0 92.1
1700 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1702 =head1 MORE INFORMATION
1704 Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
1705 or send email to L<DBM-Deep@googlegroups.com>. You can also visit #dbm-deep on
1708 The source code repository is at L<http://svn.perl.org/modules/DBM-Deep>
1712 Rob Kinyon, L<rkinyon@cpan.org>
1714 Originally written by Joseph Huckaby, L<jhuckaby@cpan.org>
1716 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
1720 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
1721 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
1725 Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved.
1726 This is free software, you may use it and distribute it under the
1727 same terms as Perl itself.