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 use Fcntl qw( :DEFAULT :flock :seek );
41 use DBM::Deep::Engine;
44 use vars qw( $VERSION );
45 $VERSION = q(0.99_01);
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( $temp );
214 # Recursively import Perl hash/array structure
216 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
218 my $self = shift->_get_self;
221 # struct is not a reference, so just import based on our type
223 $struct = $self->_repr( @_ );
226 return $self->_import( $struct );
231 # Rebuild entire database into new file, then move
232 # it back on top of original.
234 my $self = shift->_get_self;
236 #XXX Need to create a new test for this
237 # if ($self->_fileobj->{links} > 1) {
238 # $self->_throw_error("Cannot optimize: reference count is greater than 1");
241 my $db_temp = DBM::Deep->new(
242 file => $self->_fileobj->{file} . '.tmp',
247 $self->_copy_node( $db_temp );
251 # Attempt to copy user, group and permissions over to new file
253 my @stats = stat($self->_fh);
254 my $perms = $stats[2] & 07777;
257 chown( $uid, $gid, $self->_fileobj->{file} . '.tmp' );
258 chmod( $perms, $self->_fileobj->{file} . '.tmp' );
260 # q.v. perlport for more information on this variable
261 if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) {
263 # Potential race condition when optmizing on Win32 with locking.
264 # The Windows filesystem requires that the filehandle be closed
265 # before it is overwritten with rename(). This could be redone
269 $self->_fileobj->close;
272 if (!rename $self->_fileobj->{file} . '.tmp', $self->_fileobj->{file}) {
273 unlink $self->_fileobj->{file} . '.tmp';
275 $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!");
279 $self->_fileobj->close;
280 $self->_fileobj->open;
281 $self->{engine}->setup_fh( $self );
288 # Make copy of object and return
290 my $self = shift->_get_self;
292 return DBM::Deep->new(
293 type => $self->_type,
294 base_offset => $self->_base_offset,
295 fileobj => $self->_fileobj,
300 my %is_legal_filter = map {
303 store_key store_value
304 fetch_key fetch_value
309 # Setup filter function for storing or fetching the key or value
311 my $self = shift->_get_self;
315 if ( $is_legal_filter{$type} ) {
316 $self->_fileobj->{"filter_$type"} = $func;
325 my $self = shift->_get_self;
326 $self->_fileobj->begin_transaction;
331 my $self = shift->_get_self;
332 $self->_fileobj->end_transaction;
337 my $self = shift->_get_self;
338 # At this point, we need to replay the actions taken
339 $self->_fileobj->end_transaction;
348 my $self = $_[0]->_get_self;
349 return $self->{fileobj};
353 my $self = $_[0]->_get_self;
354 return $self->{type};
358 my $self = $_[0]->_get_self;
359 return $self->{base_offset};
363 my $self = $_[0]->_get_self;
364 return $self->_fileobj->{fh};
372 die "DBM::Deep: $_[1]\n";
377 (O_WRONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
382 # (O_RDONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
387 if ( $self->{parent} ) {
388 my $base = $self->{parent}->_find_parent();
389 if ( $self->{parent}->_type eq TYPE_HASH ) {
390 return $base . "\{$self->{parent_key}\}";
392 return $base . "\[$self->{parent_key}\]";
399 # Store single hash key/value or array element in database.
401 my $self = shift->_get_self;
402 my ($key, $value, $orig_key) = @_;
405 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
406 $self->_throw_error( 'Cannot write to a readonly filehandle' );
409 if ( my $afh = $self->_fileobj->{audit_fh} ) {
410 if ( defined $orig_key ) {
411 my $lhs = $self->_find_parent;
412 if ( $self->_type eq TYPE_HASH ) {
413 $lhs .= "\{$orig_key\}";
416 $lhs .= "\[$orig_key\]";
421 my $r = Scalar::Util::reftype( $value ) || '';
422 if ( $r eq 'HASH' ) {
425 elsif ( $r eq 'ARRAY' ) {
432 if ( my $c = Scalar::Util::blessed( $value ) ) {
433 $rhs = "bless $rhs, '$c'";
436 $self->_fileobj->audit( "$lhs = $rhs;" );
437 # flock( $afh, LOCK_EX );
438 # print( $afh "$lhs = $rhs; # " . localtime(time) . "\n" );
439 # flock( $afh, LOCK_UN );
444 # Request exclusive lock for writing
446 $self->lock( LOCK_EX );
448 my $md5 = $self->{engine}{digest}->($key);
450 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5, { create => 1 } );
452 # User may be storing a hash, in which case we do not want it run
453 # through the filtering system
454 if ( !ref($value) && $self->_fileobj->{filter_store_value} ) {
455 $value = $self->_fileobj->{filter_store_value}->( $value );
459 # Add key/value to bucket list
461 my $result = $self->{engine}->add_bucket( $tag, $md5, $key, $value, undef, $orig_key );
470 # Fetch single value or element given plain key or array index
472 my $self = shift->_get_self;
473 my ($key, $orig_key) = @_;
475 my $md5 = $self->{engine}{digest}->($key);
478 # Request shared lock for reading
480 $self->lock( LOCK_SH );
482 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
489 # Get value from bucket list
491 my $result = $self->{engine}->get_bucket_value( $tag, $md5, $orig_key );
495 # Filters only apply to scalar values, so the ref check is making
496 # sure the fetched bucket is a scalar, not a child hash or array.
497 return ($result && !ref($result) && $self->_fileobj->{filter_fetch_value})
498 ? $self->_fileobj->{filter_fetch_value}->($result)
504 # Delete single key/value pair or element given plain key or array index
506 my $self = shift->_get_self;
507 my ($key, $orig_key) = @_;
509 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
510 $self->_throw_error( 'Cannot write to a readonly filehandle' );
513 if ( my $afh = $self->_fileobj->{audit_fh} ) {
514 if ( defined $orig_key ) {
515 my $lhs = $self->_find_parent;
516 if ( $self->_type eq TYPE_HASH ) {
517 $lhs .= "\{$orig_key\}";
520 $lhs .= "\[$orig_key]\]";
523 flock( $afh, LOCK_EX );
524 print( $afh "delete $lhs; # " . localtime(time) . "\n" );
525 flock( $afh, LOCK_UN );
530 # Request exclusive lock for writing
532 $self->lock( LOCK_EX );
534 my $md5 = $self->{engine}{digest}->($key);
536 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
545 my $value = $self->{engine}->get_bucket_value( $tag, $md5 );
547 if (defined $value && !ref($value) && $self->_fileobj->{filter_fetch_value}) {
548 $value = $self->_fileobj->{filter_fetch_value}->($value);
551 my $result = $self->{engine}->delete_bucket( $tag, $md5, $orig_key );
554 # If this object is an array and the key deleted was on the end of the stack,
555 # decrement the length variable.
565 # Check if a single key or element exists given plain key or array index
567 my $self = shift->_get_self;
570 my $md5 = $self->{engine}{digest}->($key);
573 # Request shared lock for reading
575 $self->lock( LOCK_SH );
577 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
582 # For some reason, the built-in exists() function returns '' for false
588 # Check if bucket exists and return 1 or ''
590 my $result = $self->{engine}->bucket_exists( $tag, $md5 ) || '';
599 # Clear all keys from hash, or all elements from array.
601 my $self = shift->_get_self;
603 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
604 $self->_throw_error( 'Cannot write to a readonly filehandle' );
607 if ( my $afh = $self->_fileobj->{audit_fh} ) {
608 my $lhs = $self->_find_parent;
611 if ( $self->_type eq TYPE_HASH ) {
612 $lhs = '%{' . $lhs . '}';
615 $lhs = '@{' . $lhs . '}';
618 flock( $afh, LOCK_EX );
619 print( $afh "$lhs = $rhs; # " . localtime(time) . "\n" );
620 flock( $afh, LOCK_UN );
624 # Request exclusive lock for writing
626 $self->lock( LOCK_EX );
630 seek($fh, $self->_base_offset + $self->_fileobj->{file_offset}, SEEK_SET);
636 #XXX This needs updating to use _release_space
637 $self->{engine}->write_tag(
638 $self->_base_offset, $self->_type,
639 chr(0)x$self->{engine}{index_size},
648 # Public method aliases
650 sub put { (shift)->STORE( @_ ) }
651 sub store { (shift)->STORE( @_ ) }
652 sub get { (shift)->FETCH( @_ ) }
653 sub fetch { (shift)->FETCH( @_ ) }
654 sub delete { (shift)->DELETE( @_ ) }
655 sub exists { (shift)->EXISTS( @_ ) }
656 sub clear { (shift)->CLEAR( @_ ) }
663 DBM::Deep - A pure perl multi-level hash/array DBM
668 my $db = DBM::Deep->new( "foo.db" );
670 $db->{key} = 'value'; # tie() style
673 $db->put('key' => 'value'); # OO style
674 print $db->get('key');
676 # true multi-level support
677 $db->{my_complex} = [
678 'hello', { perl => 'rules' },
684 A unique flat-file database module, written in pure perl. True
685 multi-level hash/array support (unlike MLDBM, which is faked), hybrid
686 OO / tie() interface, cross-platform FTPable files, and quite fast. Can
687 handle millions of keys and unlimited hash levels without significant
688 slow-down. Written from the ground-up in pure perl -- this is NOT a
689 wrapper around a C-based DBM. Out-of-the-box compatibility with Unix,
690 Mac OS X and Windows.
692 =head1 VERSION DIFFERENCES
694 B<NOTE>: 0.99_01 and above have significant file format differences from 0.98 and
695 before. While attempts have been made to be backwards compatible, no guarantees.
699 Hopefully you are using Perl's excellent CPAN module, which will download
700 and install the module for you. If not, get the tarball, and run these
712 Construction can be done OO-style (which is the recommended way), or using
713 Perl's tie() function. Both are examined here.
715 =head2 OO CONSTRUCTION
717 The recommended way to construct a DBM::Deep object is to use the new()
718 method, which gets you a blessed, tied hash or array reference.
720 my $db = DBM::Deep->new( "foo.db" );
722 This opens a new database handle, mapped to the file "foo.db". If this
723 file does not exist, it will automatically be created. DB files are
724 opened in "r+" (read/write) mode, and the type of object returned is a
725 hash, unless otherwise specified (see L<OPTIONS> below).
727 You can pass a number of options to the constructor to specify things like
728 locking, autoflush, etc. This is done by passing an inline hash:
730 my $db = DBM::Deep->new(
736 Notice that the filename is now specified I<inside> the hash with
737 the "file" parameter, as opposed to being the sole argument to the
738 constructor. This is required if any options are specified.
739 See L<OPTIONS> below for the complete list.
743 You can also start with an array instead of a hash. For this, you must
744 specify the C<type> parameter:
746 my $db = DBM::Deep->new(
748 type => DBM::Deep->TYPE_ARRAY
751 B<Note:> Specifing the C<type> parameter only takes effect when beginning
752 a new DB file. If you create a DBM::Deep object with an existing file, the
753 C<type> will be loaded from the file header, and an error will be thrown if
754 the wrong type is passed in.
756 =head2 TIE CONSTRUCTION
758 Alternately, you can create a DBM::Deep handle by using Perl's built-in
759 tie() function. The object returned from tie() can be used to call methods,
760 such as lock() and unlock(), but cannot be used to assign to the DBM::Deep
761 file (as expected with most tie'd objects).
764 my $db = tie %hash, "DBM::Deep", "foo.db";
767 my $db = tie @array, "DBM::Deep", "bar.db";
769 As with the OO constructor, you can replace the DB filename parameter with
770 a hash containing one or more options (see L<OPTIONS> just below for the
773 tie %hash, "DBM::Deep", {
781 There are a number of options that can be passed in when constructing your
782 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
788 Filename of the DB file to link the handle to. You can pass a full absolute
789 filesystem path, partial path, or a plain filename if the file is in the
790 current working directory. This is a required parameter (though q.v. fh).
794 If you want, you can pass in the fh instead of the file. This is most useful for doing
797 my $db = DBM::Deep->new( { fh => \*DATA } );
799 You are responsible for making sure that the fh has been opened appropriately for your
800 needs. If you open it read-only and attempt to write, an exception will be thrown. If you
801 open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
802 needs to read from the fh.
806 This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
807 not need to set this. However, it's there if you want it.
809 If you pass in fh and do not set this, it will be set appropriately.
813 This parameter specifies what type of object to create, a hash or array. Use
814 one of these two constants:
818 =item * C<DBM::Deep-E<gt>TYPE_HASH>
820 =item * C<DBM::Deep-E<gt>TYPE_ARRAY>.
824 This only takes effect when beginning a new file. This is an optional
825 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
829 Specifies whether locking is to be enabled. DBM::Deep uses Perl's Fnctl flock()
830 function to lock the database in exclusive mode for writes, and shared mode for
831 reads. Pass any true value to enable. This affects the base DB handle I<and
832 any child hashes or arrays> that use the same DB file. This is an optional
833 parameter, and defaults to 0 (disabled). See L<LOCKING> below for more.
837 Specifies whether autoflush is to be enabled on the underlying filehandle.
838 This obviously slows down write operations, but is required if you may have
839 multiple processes accessing the same DB file (also consider enable I<locking>).
840 Pass any true value to enable. This is an optional parameter, and defaults to 0
845 If I<autobless> mode is enabled, DBM::Deep will preserve the class something
846 is blessed into, and restores it when fetched. This is an optional parameter, and defaults to 1 (enabled).
848 B<Note:> If you use the OO-interface, you will not be able to call any methods
849 of DBM::Deep on the blessed item. This is considered to be a feature.
853 See L</FILTERS> below.
859 With DBM::Deep you can access your databases using Perl's standard hash/array
860 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
861 treat them as such. DBM::Deep will intercept all reads/writes and direct them
862 to the right place -- the DB file. This has nothing to do with the
863 L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
864 using regular hashes and arrays, rather than calling functions like C<get()>
865 and C<put()> (although those work too). It is entirely up to you how to want
866 to access your databases.
870 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
871 or even nested hashes (or arrays) using standard Perl syntax:
873 my $db = DBM::Deep->new( "foo.db" );
875 $db->{mykey} = "myvalue";
877 $db->{myhash}->{subkey} = "subvalue";
879 print $db->{myhash}->{subkey} . "\n";
881 You can even step through hash keys using the normal Perl C<keys()> function:
883 foreach my $key (keys %$db) {
884 print "$key: " . $db->{$key} . "\n";
887 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
888 pushes them onto an array, all before the loop even begins. If you have an
889 extra large hash, this may exhaust Perl's memory. Instead, consider using
890 Perl's C<each()> function, which pulls keys/values one at a time, using very
893 while (my ($key, $value) = each %$db) {
894 print "$key: $value\n";
897 Please note that when using C<each()>, you should always pass a direct
898 hash reference, not a lookup. Meaning, you should B<never> do this:
901 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
903 This causes an infinite loop, because for each iteration, Perl is calling
904 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
905 it effectively keeps returning the first key over and over again. Instead,
906 assign a temporary variable to C<$db->{foo}>, then pass that to each().
910 As with hashes, you can treat any DBM::Deep object like a normal Perl array
911 reference. This includes inserting, removing and manipulating elements,
912 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
913 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
914 or simply be a nested array reference inside a hash. Example:
916 my $db = DBM::Deep->new(
917 file => "foo-array.db",
918 type => DBM::Deep->TYPE_ARRAY
922 push @$db, "bar", "baz";
925 my $last_elem = pop @$db; # baz
926 my $first_elem = shift @$db; # bah
927 my $second_elem = $db->[1]; # bar
929 my $num_elements = scalar @$db;
933 In addition to the I<tie()> interface, you can also use a standard OO interface
934 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
935 array) has its own methods, but both types share the following common methods:
936 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>.
940 =item * new() / clone()
942 These are the constructor and copy-functions.
944 =item * put() / store()
946 Stores a new hash key/value pair, or sets an array element value. Takes two
947 arguments, the hash key or array index, and the new value. The value can be
948 a scalar, hash ref or array ref. Returns true on success, false on failure.
950 $db->put("foo", "bar"); # for hashes
951 $db->put(1, "bar"); # for arrays
953 =item * get() / fetch()
955 Fetches the value of a hash key or array element. Takes one argument: the hash
956 key or array index. Returns a scalar, hash ref or array ref, depending on the
959 my $value = $db->get("foo"); # for hashes
960 my $value = $db->get(1); # for arrays
964 Checks if a hash key or array index exists. Takes one argument: the hash key
965 or array index. Returns true if it exists, false if not.
967 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
968 if ($db->exists(1)) { print "yay!\n"; } # for arrays
972 Deletes one hash key/value pair or array element. Takes one argument: the hash
973 key or array index. Returns true on success, false if not found. For arrays,
974 the remaining elements located after the deleted element are NOT moved over.
975 The deleted element is essentially just undefined, which is exactly how Perl's
976 internal arrays work. Please note that the space occupied by the deleted
977 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
978 below for details and workarounds.
980 $db->delete("foo"); # for hashes
981 $db->delete(1); # for arrays
985 Deletes B<all> hash keys or array elements. Takes no arguments. No return
986 value. Please note that the space occupied by the deleted keys/values or
987 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
988 details and workarounds.
990 $db->clear(); # hashes or arrays
992 =item * lock() / unlock()
998 Recover lost disk space.
1000 =item * import() / export()
1002 Data going in and out.
1008 For hashes, DBM::Deep supports all the common methods described above, and the
1009 following additional methods: C<first_key()> and C<next_key()>.
1015 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
1016 fetched in an undefined order (which appears random). Takes no arguments,
1017 returns the key as a scalar value.
1019 my $key = $db->first_key();
1023 Returns the "next" key in the hash, given the previous one as the sole argument.
1024 Returns undef if there are no more keys to be fetched.
1026 $key = $db->next_key($key);
1030 Here are some examples of using hashes:
1032 my $db = DBM::Deep->new( "foo.db" );
1034 $db->put("foo", "bar");
1035 print "foo: " . $db->get("foo") . "\n";
1037 $db->put("baz", {}); # new child hash ref
1038 $db->get("baz")->put("buz", "biz");
1039 print "buz: " . $db->get("baz")->get("buz") . "\n";
1041 my $key = $db->first_key();
1043 print "$key: " . $db->get($key) . "\n";
1044 $key = $db->next_key($key);
1047 if ($db->exists("foo")) { $db->delete("foo"); }
1051 For arrays, DBM::Deep supports all the common methods described above, and the
1052 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
1053 C<unshift()> and C<splice()>.
1059 Returns the number of elements in the array. Takes no arguments.
1061 my $len = $db->length();
1065 Adds one or more elements onto the end of the array. Accepts scalars, hash
1066 refs or array refs. No return value.
1068 $db->push("foo", "bar", {});
1072 Fetches the last element in the array, and deletes it. Takes no arguments.
1073 Returns undef if array is empty. Returns the element value.
1075 my $elem = $db->pop();
1079 Fetches the first element in the array, deletes it, then shifts all the
1080 remaining elements over to take up the space. Returns the element value. This
1081 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
1084 my $elem = $db->shift();
1088 Inserts one or more elements onto the beginning of the array, shifting all
1089 existing elements over to make room. Accepts scalars, hash refs or array refs.
1090 No return value. This method is not recommended with large arrays -- see
1091 <LARGE ARRAYS> below for details.
1093 $db->unshift("foo", "bar", {});
1097 Performs exactly like Perl's built-in function of the same name. See L<perldoc
1098 -f splice> for usage -- it is too complicated to document here. This method is
1099 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
1103 Here are some examples of using arrays:
1105 my $db = DBM::Deep->new(
1107 type => DBM::Deep->TYPE_ARRAY
1110 $db->push("bar", "baz");
1111 $db->unshift("foo");
1114 my $len = $db->length();
1115 print "length: $len\n"; # 4
1117 for (my $k=0; $k<$len; $k++) {
1118 print "$k: " . $db->get($k) . "\n";
1121 $db->splice(1, 2, "biz", "baf");
1123 while (my $elem = shift @$db) {
1124 print "shifted: $elem\n";
1129 Enable automatic file locking by passing a true value to the C<locking>
1130 parameter when constructing your DBM::Deep object (see L<SETUP> above).
1132 my $db = DBM::Deep->new(
1137 This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
1138 mode for writes, and shared mode for reads. This is required if you have
1139 multiple processes accessing the same database file, to avoid file corruption.
1140 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
1141 NFS> below for more.
1143 =head2 EXPLICIT LOCKING
1145 You can explicitly lock a database, so it remains locked for multiple
1146 transactions. This is done by calling the C<lock()> method, and passing an
1147 optional lock mode argument (defaults to exclusive mode). This is particularly
1148 useful for things like counters, where the current value needs to be fetched,
1149 then incremented, then stored again.
1152 my $counter = $db->get("counter");
1154 $db->put("counter", $counter);
1163 You can pass C<lock()> an optional argument, which specifies which mode to use
1164 (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX>
1165 or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the
1166 same as the constants defined in Perl's C<Fcntl> module.
1168 $db->lock( DBM::Deep->LOCK_SH );
1172 =head1 IMPORTING/EXPORTING
1174 You can import existing complex structures by calling the C<import()> method,
1175 and export an entire database into an in-memory structure using the C<export()>
1176 method. Both are examined here.
1180 Say you have an existing hash with nested hashes/arrays inside it. Instead of
1181 walking the structure and adding keys/elements to the database as you go,
1182 simply pass a reference to the C<import()> method. This recursively adds
1183 everything to an existing DBM::Deep object for you. Here is an example:
1188 array1 => [ "elem0", "elem1", "elem2" ],
1190 subkey1 => "subvalue1",
1191 subkey2 => "subvalue2"
1195 my $db = DBM::Deep->new( "foo.db" );
1196 $db->import( $struct );
1198 print $db->{key1} . "\n"; # prints "value1"
1200 This recursively imports the entire C<$struct> object into C<$db>, including
1201 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
1202 keys are merged with the existing ones, replacing if they already exist.
1203 The C<import()> method can be called on any database level (not just the base
1204 level), and works with both hash and array DB types.
1206 B<Note:> Make sure your existing structure has no circular references in it.
1207 These will cause an infinite loop when importing.
1211 Calling the C<export()> method on an existing DBM::Deep object will return
1212 a reference to a new in-memory copy of the database. The export is done
1213 recursively, so all nested hashes/arrays are all exported to standard Perl
1214 objects. Here is an example:
1216 my $db = DBM::Deep->new( "foo.db" );
1218 $db->{key1} = "value1";
1219 $db->{key2} = "value2";
1221 $db->{hash1}->{subkey1} = "subvalue1";
1222 $db->{hash1}->{subkey2} = "subvalue2";
1224 my $struct = $db->export();
1226 print $struct->{key1} . "\n"; # prints "value1"
1228 This makes a complete copy of the database in memory, and returns a reference
1229 to it. The C<export()> method can be called on any database level (not just
1230 the base level), and works with both hash and array DB types. Be careful of
1231 large databases -- you can store a lot more data in a DBM::Deep object than an
1232 in-memory Perl structure.
1234 B<Note:> Make sure your database has no circular references in it.
1235 These will cause an infinite loop when exporting.
1239 DBM::Deep has a number of hooks where you can specify your own Perl function
1240 to perform filtering on incoming or outgoing data. This is a perfect
1241 way to extend the engine, and implement things like real-time compression or
1242 encryption. Filtering applies to the base DB level, and all child hashes /
1243 arrays. Filter hooks can be specified when your DBM::Deep object is first
1244 constructed, or by calling the C<set_filter()> method at any time. There are
1245 four available filter hooks, described below:
1249 =item * filter_store_key
1251 This filter is called whenever a hash key is stored. It
1252 is passed the incoming key, and expected to return a transformed key.
1254 =item * filter_store_value
1256 This filter is called whenever a hash key or array element is stored. It
1257 is passed the incoming value, and expected to return a transformed value.
1259 =item * filter_fetch_key
1261 This filter is called whenever a hash key is fetched (i.e. via
1262 C<first_key()> or C<next_key()>). It is passed the transformed key,
1263 and expected to return the plain key.
1265 =item * filter_fetch_value
1267 This filter is called whenever a hash key or array element is fetched.
1268 It is passed the transformed value, and expected to return the plain value.
1272 Here are the two ways to setup a filter hook:
1274 my $db = DBM::Deep->new(
1276 filter_store_value => \&my_filter_store,
1277 filter_fetch_value => \&my_filter_fetch
1282 $db->set_filter( "filter_store_value", \&my_filter_store );
1283 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
1285 Your filter function will be called only when dealing with SCALAR keys or
1286 values. When nested hashes and arrays are being stored/fetched, filtering
1287 is bypassed. Filters are called as static functions, passed a single SCALAR
1288 argument, and expected to return a single SCALAR value. If you want to
1289 remove a filter, set the function reference to C<undef>:
1291 $db->set_filter( "filter_store_value", undef );
1293 =head2 REAL-TIME ENCRYPTION EXAMPLE
1295 Here is a working example that uses the I<Crypt::Blowfish> module to
1296 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
1297 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
1298 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
1301 use Crypt::Blowfish;
1304 my $cipher = Crypt::CBC->new({
1305 'key' => 'my secret key',
1306 'cipher' => 'Blowfish',
1308 'regenerate_key' => 0,
1309 'padding' => 'space',
1313 my $db = DBM::Deep->new(
1314 file => "foo-encrypt.db",
1315 filter_store_key => \&my_encrypt,
1316 filter_store_value => \&my_encrypt,
1317 filter_fetch_key => \&my_decrypt,
1318 filter_fetch_value => \&my_decrypt,
1321 $db->{key1} = "value1";
1322 $db->{key2} = "value2";
1323 print "key1: " . $db->{key1} . "\n";
1324 print "key2: " . $db->{key2} . "\n";
1330 return $cipher->encrypt( $_[0] );
1333 return $cipher->decrypt( $_[0] );
1336 =head2 REAL-TIME COMPRESSION EXAMPLE
1338 Here is a working example that uses the I<Compress::Zlib> module to do real-time
1339 compression / decompression of keys & values with DBM::Deep Filters.
1340 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
1341 more on I<Compress::Zlib>.
1346 my $db = DBM::Deep->new(
1347 file => "foo-compress.db",
1348 filter_store_key => \&my_compress,
1349 filter_store_value => \&my_compress,
1350 filter_fetch_key => \&my_decompress,
1351 filter_fetch_value => \&my_decompress,
1354 $db->{key1} = "value1";
1355 $db->{key2} = "value2";
1356 print "key1: " . $db->{key1} . "\n";
1357 print "key2: " . $db->{key2} . "\n";
1363 return Compress::Zlib::memGzip( $_[0] ) ;
1366 return Compress::Zlib::memGunzip( $_[0] ) ;
1369 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
1370 actually numerical index numbers, and are not filtered.
1372 =head1 ERROR HANDLING
1374 Most DBM::Deep methods return a true value for success, and call die() on
1375 failure. You can wrap calls in an eval block to catch the die.
1377 my $db = DBM::Deep->new( "foo.db" ); # create hash
1378 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
1380 print $@; # prints error message
1382 =head1 LARGEFILE SUPPORT
1384 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
1385 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
1386 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
1387 by specifying the 'pack_size' parameter when constructing the file.
1390 filename => $filename,
1391 pack_size => 'large',
1394 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
1395 instead of 32-bit longs. After setting these values your DB files have a
1396 theoretical maximum size of 16 XB (exabytes).
1398 You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file
1401 B<Note:> Changing these values will B<NOT> work for existing database files.
1402 Only change this for new files. Once the value has been set, it is stored in
1403 the file's header and cannot be changed for the life of the file. These
1404 parameters are per-file, meaning you can access 32-bit and 64-bit files, as
1407 B<Note:> We have not personally tested files larger than 2 GB -- all my
1408 systems have only a 32-bit Perl. However, I have received user reports that
1409 this does indeed work!
1411 =head1 LOW-LEVEL ACCESS
1413 If you require low-level access to the underlying filehandle that DBM::Deep uses,
1414 you can call the C<_fh()> method, which returns the handle:
1416 my $fh = $db->_fh();
1418 This method can be called on the root level of the datbase, or any child
1419 hashes or arrays. All levels share a I<root> structure, which contains things
1420 like the filehandle, a reference counter, and all the options specified
1421 when you created the object. You can get access to this file object by
1422 calling the C<_fileobj()> method.
1424 my $file_obj = $db->_fileobj();
1426 This is useful for changing options after the object has already been created,
1427 such as enabling/disabling locking. You can also store your own temporary user
1428 data in this structure (be wary of name collision), which is then accessible from
1429 any child hash or array.
1431 =head1 CUSTOM DIGEST ALGORITHM
1433 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
1434 keys. However you can override this, and use another algorithm (such as SHA-256)
1435 or even write your own. But please note that DBM::Deep currently expects zero
1436 collisions, so your algorithm has to be I<perfect>, so to speak. Collision
1437 detection may be introduced in a later version.
1439 You can specify a custom digest algorithm by passing it into the parameter
1440 list for new(), passing a reference to a subroutine as the 'digest' parameter,
1441 and the length of the algorithm's hashes (in bytes) as the 'hash_size'
1442 parameter. Here is a working example that uses a 256-bit hash from the
1443 I<Digest::SHA256> module. Please see
1444 L<http://search.cpan.org/search?module=Digest::SHA256> for more information.
1449 my $context = Digest::SHA256::new(256);
1451 my $db = DBM::Deep->new(
1452 filename => "foo-sha.db",
1453 digest => \&my_digest,
1457 $db->{key1} = "value1";
1458 $db->{key2} = "value2";
1459 print "key1: " . $db->{key1} . "\n";
1460 print "key2: " . $db->{key2} . "\n";
1466 return substr( $context->hash($_[0]), 0, 32 );
1469 B<Note:> Your returned digest strings must be B<EXACTLY> the number
1470 of bytes you specify in the hash_size parameter (in this case 32).
1472 B<Note:> If you do choose to use a custom digest algorithm, you must set it
1473 every time you access this file. Otherwise, the default (MD5) will be used.
1475 =head1 CIRCULAR REFERENCES
1477 DBM::Deep has B<experimental> support for circular references. Meaning you
1478 can have a nested hash key or array element that points to a parent object.
1479 This relationship is stored in the DB file, and is preserved between sessions.
1482 my $db = DBM::Deep->new( "foo.db" );
1485 $db->{circle} = $db; # ref to self
1487 print $db->{foo} . "\n"; # prints "bar"
1488 print $db->{circle}->{foo} . "\n"; # prints "bar" again
1490 B<Note>: Passing the object to a function that recursively walks the
1491 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
1492 C<export()> methods) will result in an infinite loop. This will be fixed in
1495 =head1 CAVEATS / ISSUES / BUGS
1497 This section describes all the known issues with DBM::Deep. It you have found
1498 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
1500 =head2 UNUSED SPACE RECOVERY
1502 One major caveat with DBM::Deep is that space occupied by existing keys and
1503 values is not recovered when they are deleted. Meaning if you keep deleting
1504 and adding new keys, your file will continuously grow. I am working on this,
1505 but in the meantime you can call the built-in C<optimize()> method from time to
1506 time (perhaps in a crontab or something) to recover all your unused space.
1508 $db->optimize(); # returns true on success
1510 This rebuilds the ENTIRE database into a new file, then moves it on top of
1511 the original. The new file will have no unused space, thus it will take up as
1512 little disk space as possible. Please note that this operation can take
1513 a long time for large files, and you need enough disk space to temporarily hold
1514 2 copies of your DB file. The temporary file is created in the same directory
1515 as the original, named with a ".tmp" extension, and is deleted when the
1516 operation completes. Oh, and if locking is enabled, the DB is automatically
1517 locked for the entire duration of the copy.
1519 B<WARNING:> Only call optimize() on the top-level node of the database, and
1520 make sure there are no child references lying around. DBM::Deep keeps a reference
1521 counter, and if it is greater than 1, optimize() will abort and return undef.
1525 (The reasons given assume a high level of Perl understanding, specifically of
1526 references. You can safely skip this section.)
1528 Currently, the only references supported are HASH and ARRAY. The other reference
1529 types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons.
1535 These are things like filehandles and other sockets. They can't be supported
1536 because it's completely unclear how DBM::Deep should serialize them.
1538 =item * SCALAR / REF
1540 The discussion here refers to the following type of example:
1547 # In some other process ...
1549 my $val = ${ $db->{key1} };
1551 is( $val, 50, "What actually gets stored in the DB file?" );
1553 The problem is one of synchronization. When the variable being referred to
1554 changes value, the reference isn't notified. This means that the new value won't
1555 be stored in the datafile for other processes to read. There is no TIEREF.
1557 It is theoretically possible to store references to values already within a
1558 DBM::Deep object because everything already is synchronized, but the change to
1559 the internals would be quite large. Specifically, DBM::Deep would have to tie
1560 every single value that is stored. This would bloat the RAM footprint of
1561 DBM::Deep at least twofold (if not more) and be a significant performance drain,
1562 all to support a feature that has never been requested.
1566 L<http://search.cpan.org/search?module=Data::Dump::Streamer> provides a
1567 mechanism for serializing coderefs, including saving off all closure state.
1568 However, just as for SCALAR and REF, that closure state may change without
1569 notifying the DBM::Deep object storing the reference.
1573 =head2 FILE CORRUPTION
1575 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
1576 for a 32-bit signature when opened, but other corruption in files can cause
1577 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
1578 stuck in an infinite loop depending on the level of corruption. File write
1579 operations are not checked for failure (for speed), so if you happen to run
1580 out of disk space, DBM::Deep will probably fail in a bad way. These things will
1581 be addressed in a later version of DBM::Deep.
1585 Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local
1586 filesystems, but will NOT protect you from file corruption over NFS. I've heard
1587 about setting up your NFS server with a locking daemon, then using lockf() to
1588 lock your files, but your mileage may vary there as well. From what I
1589 understand, there is no real way to do it. However, if you need access to the
1590 underlying filehandle in DBM::Deep for using some other kind of locking scheme like
1591 lockf(), see the L<LOW-LEVEL ACCESS> section above.
1593 =head2 COPYING OBJECTS
1595 Beware of copying tied objects in Perl. Very strange things can happen.
1596 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
1597 returns a new, blessed, tied hash or array to the same level in the DB.
1599 my $copy = $db->clone();
1601 B<Note>: Since clone() here is cloning the object, not the database location, any
1602 modifications to either $db or $copy will be visible in both.
1606 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
1607 These functions cause every element in the array to move, which can be murder
1608 on DBM::Deep, as every element has to be fetched from disk, then stored again in
1609 a different location. This will be addressed in the forthcoming version 1.00.
1611 =head2 WRITEONLY FILES
1613 If you pass in a filehandle to new(), you may have opened it in either a readonly or
1614 writeonly mode. STORE will verify that the filehandle is writable. However, there
1615 doesn't seem to be a good way to determine if a filehandle is readable. And, if the
1616 filehandle isn't readable, it's not clear what will happen. So, don't do that.
1620 This section discusses DBM::Deep's speed and memory usage.
1624 Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as
1625 the almighty I<BerkeleyDB>. But it makes up for it in features like true
1626 multi-level hash/array support, and cross-platform FTPable files. Even so,
1627 DBM::Deep is still pretty fast, and the speed stays fairly consistent, even
1628 with huge databases. Here is some test data:
1630 Adding 1,000,000 keys to new DB file...
1632 At 100 keys, avg. speed is 2,703 keys/sec
1633 At 200 keys, avg. speed is 2,642 keys/sec
1634 At 300 keys, avg. speed is 2,598 keys/sec
1635 At 400 keys, avg. speed is 2,578 keys/sec
1636 At 500 keys, avg. speed is 2,722 keys/sec
1637 At 600 keys, avg. speed is 2,628 keys/sec
1638 At 700 keys, avg. speed is 2,700 keys/sec
1639 At 800 keys, avg. speed is 2,607 keys/sec
1640 At 900 keys, avg. speed is 2,190 keys/sec
1641 At 1,000 keys, avg. speed is 2,570 keys/sec
1642 At 2,000 keys, avg. speed is 2,417 keys/sec
1643 At 3,000 keys, avg. speed is 1,982 keys/sec
1644 At 4,000 keys, avg. speed is 1,568 keys/sec
1645 At 5,000 keys, avg. speed is 1,533 keys/sec
1646 At 6,000 keys, avg. speed is 1,787 keys/sec
1647 At 7,000 keys, avg. speed is 1,977 keys/sec
1648 At 8,000 keys, avg. speed is 2,028 keys/sec
1649 At 9,000 keys, avg. speed is 2,077 keys/sec
1650 At 10,000 keys, avg. speed is 2,031 keys/sec
1651 At 20,000 keys, avg. speed is 1,970 keys/sec
1652 At 30,000 keys, avg. speed is 2,050 keys/sec
1653 At 40,000 keys, avg. speed is 2,073 keys/sec
1654 At 50,000 keys, avg. speed is 1,973 keys/sec
1655 At 60,000 keys, avg. speed is 1,914 keys/sec
1656 At 70,000 keys, avg. speed is 2,091 keys/sec
1657 At 80,000 keys, avg. speed is 2,103 keys/sec
1658 At 90,000 keys, avg. speed is 1,886 keys/sec
1659 At 100,000 keys, avg. speed is 1,970 keys/sec
1660 At 200,000 keys, avg. speed is 2,053 keys/sec
1661 At 300,000 keys, avg. speed is 1,697 keys/sec
1662 At 400,000 keys, avg. speed is 1,838 keys/sec
1663 At 500,000 keys, avg. speed is 1,941 keys/sec
1664 At 600,000 keys, avg. speed is 1,930 keys/sec
1665 At 700,000 keys, avg. speed is 1,735 keys/sec
1666 At 800,000 keys, avg. speed is 1,795 keys/sec
1667 At 900,000 keys, avg. speed is 1,221 keys/sec
1668 At 1,000,000 keys, avg. speed is 1,077 keys/sec
1670 This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl
1671 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and
1672 values were between 6 - 12 chars in length. The DB file ended up at 210MB.
1673 Run time was 12 min 3 sec.
1677 One of the great things about DBM::Deep is that it uses very little memory.
1678 Even with huge databases (1,000,000+ keys) you will not see much increased
1679 memory on your process. DBM::Deep relies solely on the filesystem for storing
1680 and fetching data. Here is output from I</usr/bin/top> before even opening a
1683 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1684 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl
1686 Basically the process is taking 2,716K of memory. And here is the same
1687 process after storing and fetching 1,000,000 keys:
1689 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1690 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl
1692 Notice the memory usage increased by only 56K. Test was performed on a 700mHz
1693 x86 box running Linux RedHat 7.2 & Perl 5.6.1.
1695 =head1 DB FILE FORMAT
1697 In case you were interested in the underlying DB file format, it is documented
1698 here in this section. You don't need to know this to use the module, it's just
1699 included for reference.
1703 DBM::Deep files always start with a 32-bit signature to identify the file type.
1704 This is at offset 0. The signature is "DPDB" in network byte order. This is
1705 checked for when the file is opened and an error will be thrown if it's not found.
1709 The DBM::Deep file is in a I<tagged format>, meaning each section of the file
1710 has a standard header containing the type of data, the length of data, and then
1711 the data itself. The type is a single character (1 byte), the length is a
1712 32-bit unsigned long in network byte order, and the data is, well, the data.
1713 Here is how it unfolds:
1717 Immediately after the 32-bit file signature is the I<Master Index> record.
1718 This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048
1719 bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array,
1720 depending on how the DBM::Deep object was constructed.
1722 The index works by looking at a I<MD5 Hash> of the hash key (or array index
1723 number). The first 8-bit char of the MD5 signature is the offset into the
1724 index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the
1725 index element is a file offset of the next tag for the key/element in question,
1726 which is usually a I<Bucket List> tag (see below).
1728 The next tag I<could> be another index, depending on how many keys/elements
1729 exist. See L<RE-INDEXING> below for details.
1733 A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus
1734 file offsets to where the actual data is stored. It starts with a standard
1735 tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or
1736 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus
1737 the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data.
1738 When the list fills up, a I<Re-Index> operation is performed (See
1739 L<RE-INDEXING> below).
1743 A I<Bucket> is a tag containing a key/value pair (in hash mode), or a
1744 index/value pair (in array mode). It starts with a standard tag header with
1745 type I<D> for scalar data (string, binary, etc.), or it could be a nested
1746 hash (type I<H>) or array (type I<A>). The value comes just after the tag
1747 header. The size reported in the tag header is only for the value, but then,
1748 just after the value is another size (32-bit unsigned long) and then the plain
1749 key itself. Since the value is likely to be fetched more often than the plain
1750 key, I figured it would be I<slightly> faster to store the value first.
1752 If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index>
1753 record for the nested structure, where the process begins all over again.
1757 After a I<Bucket List> grows to 16 records, its allocated space in the file is
1758 exhausted. Then, when another key/element comes in, the list is converted to a
1759 new index record. However, this index will look at the next char in the MD5
1760 hash, and arrange new Bucket List pointers accordingly. This process is called
1761 I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all
1762 17 (16 + new one) keys/elements are removed from the old Bucket List and
1763 inserted into the new index. Several new Bucket Lists are created in the
1764 process, as a new MD5 char from the key is being examined (it is unlikely that
1765 the keys will all share the same next char of their MD5s).
1767 Because of the way the I<MD5> algorithm works, it is impossible to tell exactly
1768 when the Bucket Lists will turn into indexes, but the first round tends to
1769 happen right around 4,000 keys. You will see a I<slight> decrease in
1770 performance here, but it picks back up pretty quick (see L<SPEED> above). Then
1771 it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's
1772 right around 900,000 keys. This process can continue nearly indefinitely --
1773 right up until the point the I<MD5> signatures start colliding with each other,
1774 and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND
1775 getting struck by lightning while you are walking to cash in your tickets.
1776 Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to
1777 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe
1778 this is 340 unodecillion, but don't quote me).
1782 When a new key/element is stored, the key (or index number) is first run through
1783 I<Digest::MD5> to get a 128-bit signature (example, in hex:
1784 b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked
1785 for the first char of the signature (in this case I<b0>). If it does not exist,
1786 a new I<Bucket List> is created for our key (and the next 15 future keys that
1787 happen to also have I<b> as their first MD5 char). The entire MD5 is written
1788 to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at
1789 this point, unless we are replacing an existing I<Bucket>), where the actual
1790 data will be stored.
1794 Fetching an existing key/element involves getting a I<Digest::MD5> of the key
1795 (or index number), then walking along the indexes. If there are enough
1796 keys/elements in this DB level, there might be nested indexes, each linked to
1797 a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which
1798 contains up to 16 full MD5 hashes. Each is checked for equality to the key in
1799 question. If we found a match, the I<Bucket> tag is loaded, where the value and
1800 plain key are stored.
1802 Fetching the plain key occurs when calling the I<first_key()> and I<next_key()>
1803 methods. In this process the indexes are walked systematically, and each key
1804 fetched in increasing MD5 order (which is why it appears random). Once the
1805 I<Bucket> is found, the value is skipped and the plain key returned instead.
1806 B<Note:> Do not count on keys being fetched as if the MD5 hashes were
1807 alphabetically sorted. This only happens on an index-level -- as soon as the
1808 I<Bucket Lists> are hit, the keys will come out in the order they went in --
1809 so it's pretty much undefined how the keys will come out -- just like Perl's
1812 =head1 CODE COVERAGE
1814 We use B<Devel::Cover> to test the code coverage of our tests, below is the
1815 B<Devel::Cover> report on this module's test suite.
1817 ----------------------------------- ------ ------ ------ ------ ------ ------
1818 File stmt bran cond sub time total
1819 ----------------------------------- ------ ------ ------ ------ ------ ------
1820 blib/lib/DBM/Deep.pm 94.9 80.6 73.0 100.0 37.9 90.4
1821 blib/lib/DBM/Deep/Array.pm 100.0 91.1 100.0 100.0 18.2 98.1
1822 blib/lib/DBM/Deep/Engine.pm 98.9 87.3 80.0 100.0 34.2 95.2
1823 blib/lib/DBM/Deep/Hash.pm 100.0 87.5 100.0 100.0 9.7 97.3
1824 Total 97.9 85.9 79.7 100.0 100.0 94.3
1825 ----------------------------------- ------ ------ ------ ------ ------ ------
1827 =head1 MORE INFORMATION
1829 Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
1830 or send email to L<DBM-Deep@googlegroups.com>.
1834 Joseph Huckaby, L<jhuckaby@cpan.org>
1836 Rob Kinyon, L<rkinyon@cpan.org>
1838 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
1842 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
1843 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
1847 Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved.
1848 This is free software, you may use it and distribute it under the
1849 same terms as Perl itself.