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 engine => DBM::Deep::Engine->new( $args ),
119 base_offset => undef,
123 # Grab the parameters we want to use
124 foreach my $param ( keys %$self ) {
125 next unless exists $args->{$param};
126 $self->{$param} = $args->{$param};
129 $self->{engine}->setup_fh( $self );
136 require DBM::Deep::Hash;
137 return DBM::Deep::Hash->TIEHASH( @_ );
142 require DBM::Deep::Array;
143 return DBM::Deep::Array->TIEARRAY( @_ );
147 my $self = shift->_get_self;
148 return $self->_fileobj->lock( $self, @_ );
152 my $self = shift->_get_self;
153 return $self->_fileobj->unlock( $self, @_ );
157 my $self = shift->_get_self;
158 my ($spot, $value) = @_;
163 elsif ( eval { local $SIG{__DIE__}; $value->isa( 'DBM::Deep' ) } ) {
164 ${$spot} = $value->_repr;
165 $value->_copy_node( ${$spot} );
168 my $r = Scalar::Util::reftype( $value );
169 my $c = Scalar::Util::blessed( $value );
170 if ( $r eq 'ARRAY' ) {
171 ${$spot} = [ @{$value} ];
174 ${$spot} = { %{$value} };
176 ${$spot} = bless ${$spot}, $c
184 die "Must be implemented in a child class\n";
188 die "Must be implemented in a child class\n";
193 # Recursively export into standard Perl hashes and arrays.
195 my $self = shift->_get_self;
197 my $temp = $self->_repr;
200 $self->_copy_node( $temp );
208 # Recursively import Perl hash/array structure
210 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
212 my $self = shift->_get_self;
215 # struct is not a reference, so just import based on our type
217 $struct = $self->_repr( @_ );
220 return $self->_import( $struct );
225 # Rebuild entire database into new file, then move
226 # it back on top of original.
228 my $self = shift->_get_self;
230 #XXX Need to create a new test for this
231 # if ($self->_fileobj->{links} > 1) {
232 # $self->_throw_error("Cannot optimize: reference count is greater than 1");
235 my $db_temp = DBM::Deep->new(
236 file => $self->_fileobj->{file} . '.tmp',
241 $self->_copy_node( $db_temp );
245 # Attempt to copy user, group and permissions over to new file
247 my @stats = stat($self->_fh);
248 my $perms = $stats[2] & 07777;
251 chown( $uid, $gid, $self->_fileobj->{file} . '.tmp' );
252 chmod( $perms, $self->_fileobj->{file} . '.tmp' );
254 # q.v. perlport for more information on this variable
255 if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) {
257 # Potential race condition when optmizing on Win32 with locking.
258 # The Windows filesystem requires that the filehandle be closed
259 # before it is overwritten with rename(). This could be redone
263 $self->_fileobj->close;
266 if (!rename $self->_fileobj->{file} . '.tmp', $self->_fileobj->{file}) {
267 unlink $self->_fileobj->{file} . '.tmp';
269 $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!");
273 $self->_fileobj->close;
274 $self->_fileobj->open;
275 $self->{engine}->setup_fh( $self );
282 # Make copy of object and return
284 my $self = shift->_get_self;
286 return DBM::Deep->new(
287 type => $self->_type,
288 base_offset => $self->_base_offset,
289 fileobj => $self->_fileobj,
294 my %is_legal_filter = map {
297 store_key store_value
298 fetch_key fetch_value
303 # Setup filter function for storing or fetching the key or value
305 my $self = shift->_get_self;
309 if ( $is_legal_filter{$type} ) {
310 $self->_fileobj->{"filter_$type"} = $func;
319 my $self = shift->_get_self;
320 $self->_fileobj->begin_transaction;
325 my $self = shift->_get_self;
326 $self->_fileobj->end_transaction;
331 # my $self = shift->_get_self;
340 # Get access to the root structure
342 my $self = $_[0]->_get_self;
343 return $self->{fileobj};
348 # Get type of current node (TYPE_HASH or TYPE_ARRAY)
350 my $self = $_[0]->_get_self;
351 return $self->{type};
356 # Get base_offset of current node (TYPE_HASH or TYPE_ARRAY)
358 my $self = $_[0]->_get_self;
359 return $self->{base_offset};
364 # Get access to the raw fh
366 my $self = $_[0]->_get_self;
367 return $self->_fileobj->{fh};
375 die "DBM::Deep: $_[1]\n";
380 (O_WRONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
385 # (O_RDONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
390 # Store single hash key/value or array element in database.
392 my $self = shift->_get_self;
393 my ($key, $value) = @_;
395 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
396 $self->_throw_error( 'Cannot write to a readonly filehandle' );
400 # Request exclusive lock for writing
402 $self->lock( LOCK_EX );
404 my $md5 = $self->{engine}{digest}->($key);
406 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5, { create => 1 } );
408 # User may be storing a hash, in which case we do not want it run
409 # through the filtering system
410 if ( !ref($value) && $self->_fileobj->{filter_store_value} ) {
411 $value = $self->_fileobj->{filter_store_value}->( $value );
415 # Add key/value to bucket list
417 my $result = $self->{engine}->add_bucket( $tag, $md5, $key, $value );
426 # Fetch single value or element given plain key or array index
428 my $self = shift->_get_self;
431 my $md5 = $self->{engine}{digest}->($key);
434 # Request shared lock for reading
436 $self->lock( LOCK_SH );
438 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
445 # Get value from bucket list
447 my $result = $self->{engine}->get_bucket_value( $tag, $md5 );
451 # Filters only apply to scalar values, so the ref check is making
452 # sure the fetched bucket is a scalar, not a child hash or array.
453 return ($result && !ref($result) && $self->_fileobj->{filter_fetch_value})
454 ? $self->_fileobj->{filter_fetch_value}->($result)
460 # Delete single key/value pair or element given plain key or array index
462 my $self = $_[0]->_get_self;
465 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
466 $self->_throw_error( 'Cannot write to a readonly filehandle' );
470 # Request exclusive lock for writing
472 $self->lock( LOCK_EX );
474 my $md5 = $self->{engine}{digest}->($key);
476 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
485 my $value = $self->{engine}->get_bucket_value( $tag, $md5 );
487 if (defined $value && !ref($value) && $self->_fileobj->{filter_fetch_value}) {
488 $value = $self->_fileobj->{filter_fetch_value}->($value);
491 my $result = $self->{engine}->delete_bucket( $tag, $md5 );
494 # If this object is an array and the key deleted was on the end of the stack,
495 # decrement the length variable.
505 # Check if a single key or element exists given plain key or array index
507 my $self = $_[0]->_get_self;
510 my $md5 = $self->{engine}{digest}->($key);
513 # Request shared lock for reading
515 $self->lock( LOCK_SH );
517 my $tag = $self->{engine}->find_bucket_list( $self->_base_offset, $md5 );
522 # For some reason, the built-in exists() function returns '' for false
528 # Check if bucket exists and return 1 or ''
530 my $result = $self->{engine}->bucket_exists( $tag, $md5 ) || '';
539 # Clear all keys from hash, or all elements from array.
541 my $self = $_[0]->_get_self;
543 if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) {
544 $self->_throw_error( 'Cannot write to a readonly filehandle' );
548 # Request exclusive lock for writing
550 $self->lock( LOCK_EX );
554 seek($fh, $self->_base_offset + $self->_fileobj->{file_offset}, SEEK_SET);
560 #XXX This needs updating to use _release_space
561 $self->{engine}->write_tag(
562 $self->_base_offset, $self->_type,
563 chr(0)x$self->{engine}{index_size},
572 # Public method aliases
574 sub put { (shift)->STORE( @_ ) }
575 sub store { (shift)->STORE( @_ ) }
576 sub get { (shift)->FETCH( @_ ) }
577 sub fetch { (shift)->FETCH( @_ ) }
578 sub delete { (shift)->DELETE( @_ ) }
579 sub exists { (shift)->EXISTS( @_ ) }
580 sub clear { (shift)->CLEAR( @_ ) }
587 DBM::Deep - A pure perl multi-level hash/array DBM
592 my $db = DBM::Deep->new( "foo.db" );
594 $db->{key} = 'value'; # tie() style
597 $db->put('key' => 'value'); # OO style
598 print $db->get('key');
600 # true multi-level support
601 $db->{my_complex} = [
602 'hello', { perl => 'rules' },
608 A unique flat-file database module, written in pure perl. True
609 multi-level hash/array support (unlike MLDBM, which is faked), hybrid
610 OO / tie() interface, cross-platform FTPable files, and quite fast. Can
611 handle millions of keys and unlimited hash levels without significant
612 slow-down. Written from the ground-up in pure perl -- this is NOT a
613 wrapper around a C-based DBM. Out-of-the-box compatibility with Unix,
614 Mac OS X and Windows.
616 =head1 VERSION DIFFERENCES
618 B<NOTE>: 0.99_01 and above have significant file format differences from 0.98 and
619 before. While attempts have been made to be backwards compatible, no guarantees.
623 Hopefully you are using Perl's excellent CPAN module, which will download
624 and install the module for you. If not, get the tarball, and run these
636 Construction can be done OO-style (which is the recommended way), or using
637 Perl's tie() function. Both are examined here.
639 =head2 OO CONSTRUCTION
641 The recommended way to construct a DBM::Deep object is to use the new()
642 method, which gets you a blessed, tied hash or array reference.
644 my $db = DBM::Deep->new( "foo.db" );
646 This opens a new database handle, mapped to the file "foo.db". If this
647 file does not exist, it will automatically be created. DB files are
648 opened in "r+" (read/write) mode, and the type of object returned is a
649 hash, unless otherwise specified (see L<OPTIONS> below).
651 You can pass a number of options to the constructor to specify things like
652 locking, autoflush, etc. This is done by passing an inline hash:
654 my $db = DBM::Deep->new(
660 Notice that the filename is now specified I<inside> the hash with
661 the "file" parameter, as opposed to being the sole argument to the
662 constructor. This is required if any options are specified.
663 See L<OPTIONS> below for the complete list.
667 You can also start with an array instead of a hash. For this, you must
668 specify the C<type> parameter:
670 my $db = DBM::Deep->new(
672 type => DBM::Deep->TYPE_ARRAY
675 B<Note:> Specifing the C<type> parameter only takes effect when beginning
676 a new DB file. If you create a DBM::Deep object with an existing file, the
677 C<type> will be loaded from the file header, and an error will be thrown if
678 the wrong type is passed in.
680 =head2 TIE CONSTRUCTION
682 Alternately, you can create a DBM::Deep handle by using Perl's built-in
683 tie() function. The object returned from tie() can be used to call methods,
684 such as lock() and unlock(), but cannot be used to assign to the DBM::Deep
685 file (as expected with most tie'd objects).
688 my $db = tie %hash, "DBM::Deep", "foo.db";
691 my $db = tie @array, "DBM::Deep", "bar.db";
693 As with the OO constructor, you can replace the DB filename parameter with
694 a hash containing one or more options (see L<OPTIONS> just below for the
697 tie %hash, "DBM::Deep", {
705 There are a number of options that can be passed in when constructing your
706 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
712 Filename of the DB file to link the handle to. You can pass a full absolute
713 filesystem path, partial path, or a plain filename if the file is in the
714 current working directory. This is a required parameter (though q.v. fh).
718 If you want, you can pass in the fh instead of the file. This is most useful for doing
721 my $db = DBM::Deep->new( { fh => \*DATA } );
723 You are responsible for making sure that the fh has been opened appropriately for your
724 needs. If you open it read-only and attempt to write, an exception will be thrown. If you
725 open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
726 needs to read from the fh.
730 This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
731 not need to set this. However, it's there if you want it.
733 If you pass in fh and do not set this, it will be set appropriately.
737 This parameter specifies what type of object to create, a hash or array. Use
738 one of these two constants: C<DBM::Deep-E<gt>TYPE_HASH> or C<DBM::Deep-E<gt>TYPE_ARRAY>.
739 This only takes effect when beginning a new file. This is an optional
740 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
744 Specifies whether locking is to be enabled. DBM::Deep uses Perl's Fnctl flock()
745 function to lock the database in exclusive mode for writes, and shared mode for
746 reads. Pass any true value to enable. This affects the base DB handle I<and
747 any child hashes or arrays> that use the same DB file. This is an optional
748 parameter, and defaults to 0 (disabled). See L<LOCKING> below for more.
752 Specifies whether autoflush is to be enabled on the underlying filehandle.
753 This obviously slows down write operations, but is required if you may have
754 multiple processes accessing the same DB file (also consider enable I<locking>).
755 Pass any true value to enable. This is an optional parameter, and defaults to 0
760 If I<autobless> mode is enabled, DBM::Deep will preserve blessed hashes, and
761 restore them when fetched. This is an B<experimental> feature, and does have
762 side-effects. Basically, when hashes are re-blessed into their original
763 classes, they are no longer blessed into the DBM::Deep class! So you won't be
764 able to call any DBM::Deep methods on them. You have been warned.
765 This is an optional parameter, and defaults to 0 (disabled).
769 See L<FILTERS> below.
775 With DBM::Deep you can access your databases using Perl's standard hash/array
776 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
777 treat them as such. DBM::Deep will intercept all reads/writes and direct them
778 to the right place -- the DB file. This has nothing to do with the
779 L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
780 using regular hashes and arrays, rather than calling functions like C<get()>
781 and C<put()> (although those work too). It is entirely up to you how to want
782 to access your databases.
786 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
787 or even nested hashes (or arrays) using standard Perl syntax:
789 my $db = DBM::Deep->new( "foo.db" );
791 $db->{mykey} = "myvalue";
793 $db->{myhash}->{subkey} = "subvalue";
795 print $db->{myhash}->{subkey} . "\n";
797 You can even step through hash keys using the normal Perl C<keys()> function:
799 foreach my $key (keys %$db) {
800 print "$key: " . $db->{$key} . "\n";
803 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
804 pushes them onto an array, all before the loop even begins. If you have an
805 extra large hash, this may exhaust Perl's memory. Instead, consider using
806 Perl's C<each()> function, which pulls keys/values one at a time, using very
809 while (my ($key, $value) = each %$db) {
810 print "$key: $value\n";
813 Please note that when using C<each()>, you should always pass a direct
814 hash reference, not a lookup. Meaning, you should B<never> do this:
817 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
819 This causes an infinite loop, because for each iteration, Perl is calling
820 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
821 it effectively keeps returning the first key over and over again. Instead,
822 assign a temporary variable to C<$db->{foo}>, then pass that to each().
826 As with hashes, you can treat any DBM::Deep object like a normal Perl array
827 reference. This includes inserting, removing and manipulating elements,
828 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
829 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
830 or simply be a nested array reference inside a hash. Example:
832 my $db = DBM::Deep->new(
833 file => "foo-array.db",
834 type => DBM::Deep->TYPE_ARRAY
838 push @$db, "bar", "baz";
841 my $last_elem = pop @$db; # baz
842 my $first_elem = shift @$db; # bah
843 my $second_elem = $db->[1]; # bar
845 my $num_elements = scalar @$db;
849 In addition to the I<tie()> interface, you can also use a standard OO interface
850 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
851 array) has its own methods, but both types share the following common methods:
852 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>.
856 =item * new() / clone()
858 These are the constructor and copy-functions.
860 =item * put() / store()
862 Stores a new hash key/value pair, or sets an array element value. Takes two
863 arguments, the hash key or array index, and the new value. The value can be
864 a scalar, hash ref or array ref. Returns true on success, false on failure.
866 $db->put("foo", "bar"); # for hashes
867 $db->put(1, "bar"); # for arrays
869 =item * get() / fetch()
871 Fetches the value of a hash key or array element. Takes one argument: the hash
872 key or array index. Returns a scalar, hash ref or array ref, depending on the
875 my $value = $db->get("foo"); # for hashes
876 my $value = $db->get(1); # for arrays
880 Checks if a hash key or array index exists. Takes one argument: the hash key
881 or array index. Returns true if it exists, false if not.
883 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
884 if ($db->exists(1)) { print "yay!\n"; } # for arrays
888 Deletes one hash key/value pair or array element. Takes one argument: the hash
889 key or array index. Returns true on success, false if not found. For arrays,
890 the remaining elements located after the deleted element are NOT moved over.
891 The deleted element is essentially just undefined, which is exactly how Perl's
892 internal arrays work. Please note that the space occupied by the deleted
893 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
894 below for details and workarounds.
896 $db->delete("foo"); # for hashes
897 $db->delete(1); # for arrays
901 Deletes B<all> hash keys or array elements. Takes no arguments. No return
902 value. Please note that the space occupied by the deleted keys/values or
903 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
904 details and workarounds.
906 $db->clear(); # hashes or arrays
908 =item * lock() / unlock()
914 Recover lost disk space.
916 =item * import() / export()
918 Data going in and out.
924 For hashes, DBM::Deep supports all the common methods described above, and the
925 following additional methods: C<first_key()> and C<next_key()>.
931 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
932 fetched in an undefined order (which appears random). Takes no arguments,
933 returns the key as a scalar value.
935 my $key = $db->first_key();
939 Returns the "next" key in the hash, given the previous one as the sole argument.
940 Returns undef if there are no more keys to be fetched.
942 $key = $db->next_key($key);
946 Here are some examples of using hashes:
948 my $db = DBM::Deep->new( "foo.db" );
950 $db->put("foo", "bar");
951 print "foo: " . $db->get("foo") . "\n";
953 $db->put("baz", {}); # new child hash ref
954 $db->get("baz")->put("buz", "biz");
955 print "buz: " . $db->get("baz")->get("buz") . "\n";
957 my $key = $db->first_key();
959 print "$key: " . $db->get($key) . "\n";
960 $key = $db->next_key($key);
963 if ($db->exists("foo")) { $db->delete("foo"); }
967 For arrays, DBM::Deep supports all the common methods described above, and the
968 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
969 C<unshift()> and C<splice()>.
975 Returns the number of elements in the array. Takes no arguments.
977 my $len = $db->length();
981 Adds one or more elements onto the end of the array. Accepts scalars, hash
982 refs or array refs. No return value.
984 $db->push("foo", "bar", {});
988 Fetches the last element in the array, and deletes it. Takes no arguments.
989 Returns undef if array is empty. Returns the element value.
991 my $elem = $db->pop();
995 Fetches the first element in the array, deletes it, then shifts all the
996 remaining elements over to take up the space. Returns the element value. This
997 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
1000 my $elem = $db->shift();
1004 Inserts one or more elements onto the beginning of the array, shifting all
1005 existing elements over to make room. Accepts scalars, hash refs or array refs.
1006 No return value. This method is not recommended with large arrays -- see
1007 <LARGE ARRAYS> below for details.
1009 $db->unshift("foo", "bar", {});
1013 Performs exactly like Perl's built-in function of the same name. See L<perldoc
1014 -f splice> for usage -- it is too complicated to document here. This method is
1015 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
1019 Here are some examples of using arrays:
1021 my $db = DBM::Deep->new(
1023 type => DBM::Deep->TYPE_ARRAY
1026 $db->push("bar", "baz");
1027 $db->unshift("foo");
1030 my $len = $db->length();
1031 print "length: $len\n"; # 4
1033 for (my $k=0; $k<$len; $k++) {
1034 print "$k: " . $db->get($k) . "\n";
1037 $db->splice(1, 2, "biz", "baf");
1039 while (my $elem = shift @$db) {
1040 print "shifted: $elem\n";
1045 Enable automatic file locking by passing a true value to the C<locking>
1046 parameter when constructing your DBM::Deep object (see L<SETUP> above).
1048 my $db = DBM::Deep->new(
1053 This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
1054 mode for writes, and shared mode for reads. This is required if you have
1055 multiple processes accessing the same database file, to avoid file corruption.
1056 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
1057 NFS> below for more.
1059 =head2 EXPLICIT LOCKING
1061 You can explicitly lock a database, so it remains locked for multiple
1062 transactions. This is done by calling the C<lock()> method, and passing an
1063 optional lock mode argument (defaults to exclusive mode). This is particularly
1064 useful for things like counters, where the current value needs to be fetched,
1065 then incremented, then stored again.
1068 my $counter = $db->get("counter");
1070 $db->put("counter", $counter);
1079 You can pass C<lock()> an optional argument, which specifies which mode to use
1080 (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX>
1081 or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the
1082 same as the constants defined in Perl's C<Fcntl> module.
1084 $db->lock( DBM::Deep->LOCK_SH );
1088 =head1 IMPORTING/EXPORTING
1090 You can import existing complex structures by calling the C<import()> method,
1091 and export an entire database into an in-memory structure using the C<export()>
1092 method. Both are examined here.
1096 Say you have an existing hash with nested hashes/arrays inside it. Instead of
1097 walking the structure and adding keys/elements to the database as you go,
1098 simply pass a reference to the C<import()> method. This recursively adds
1099 everything to an existing DBM::Deep object for you. Here is an example:
1104 array1 => [ "elem0", "elem1", "elem2" ],
1106 subkey1 => "subvalue1",
1107 subkey2 => "subvalue2"
1111 my $db = DBM::Deep->new( "foo.db" );
1112 $db->import( $struct );
1114 print $db->{key1} . "\n"; # prints "value1"
1116 This recursively imports the entire C<$struct> object into C<$db>, including
1117 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
1118 keys are merged with the existing ones, replacing if they already exist.
1119 The C<import()> method can be called on any database level (not just the base
1120 level), and works with both hash and array DB types.
1122 B<Note:> Make sure your existing structure has no circular references in it.
1123 These will cause an infinite loop when importing.
1127 Calling the C<export()> method on an existing DBM::Deep object will return
1128 a reference to a new in-memory copy of the database. The export is done
1129 recursively, so all nested hashes/arrays are all exported to standard Perl
1130 objects. Here is an example:
1132 my $db = DBM::Deep->new( "foo.db" );
1134 $db->{key1} = "value1";
1135 $db->{key2} = "value2";
1137 $db->{hash1}->{subkey1} = "subvalue1";
1138 $db->{hash1}->{subkey2} = "subvalue2";
1140 my $struct = $db->export();
1142 print $struct->{key1} . "\n"; # prints "value1"
1144 This makes a complete copy of the database in memory, and returns a reference
1145 to it. The C<export()> method can be called on any database level (not just
1146 the base level), and works with both hash and array DB types. Be careful of
1147 large databases -- you can store a lot more data in a DBM::Deep object than an
1148 in-memory Perl structure.
1150 B<Note:> Make sure your database has no circular references in it.
1151 These will cause an infinite loop when exporting.
1155 DBM::Deep has a number of hooks where you can specify your own Perl function
1156 to perform filtering on incoming or outgoing data. This is a perfect
1157 way to extend the engine, and implement things like real-time compression or
1158 encryption. Filtering applies to the base DB level, and all child hashes /
1159 arrays. Filter hooks can be specified when your DBM::Deep object is first
1160 constructed, or by calling the C<set_filter()> method at any time. There are
1161 four available filter hooks, described below:
1165 =item * filter_store_key
1167 This filter is called whenever a hash key is stored. It
1168 is passed the incoming key, and expected to return a transformed key.
1170 =item * filter_store_value
1172 This filter is called whenever a hash key or array element is stored. It
1173 is passed the incoming value, and expected to return a transformed value.
1175 =item * filter_fetch_key
1177 This filter is called whenever a hash key is fetched (i.e. via
1178 C<first_key()> or C<next_key()>). It is passed the transformed key,
1179 and expected to return the plain key.
1181 =item * filter_fetch_value
1183 This filter is called whenever a hash key or array element is fetched.
1184 It is passed the transformed value, and expected to return the plain value.
1188 Here are the two ways to setup a filter hook:
1190 my $db = DBM::Deep->new(
1192 filter_store_value => \&my_filter_store,
1193 filter_fetch_value => \&my_filter_fetch
1198 $db->set_filter( "filter_store_value", \&my_filter_store );
1199 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
1201 Your filter function will be called only when dealing with SCALAR keys or
1202 values. When nested hashes and arrays are being stored/fetched, filtering
1203 is bypassed. Filters are called as static functions, passed a single SCALAR
1204 argument, and expected to return a single SCALAR value. If you want to
1205 remove a filter, set the function reference to C<undef>:
1207 $db->set_filter( "filter_store_value", undef );
1209 =head2 REAL-TIME ENCRYPTION EXAMPLE
1211 Here is a working example that uses the I<Crypt::Blowfish> module to
1212 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
1213 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
1214 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
1217 use Crypt::Blowfish;
1220 my $cipher = Crypt::CBC->new({
1221 'key' => 'my secret key',
1222 'cipher' => 'Blowfish',
1224 'regenerate_key' => 0,
1225 'padding' => 'space',
1229 my $db = DBM::Deep->new(
1230 file => "foo-encrypt.db",
1231 filter_store_key => \&my_encrypt,
1232 filter_store_value => \&my_encrypt,
1233 filter_fetch_key => \&my_decrypt,
1234 filter_fetch_value => \&my_decrypt,
1237 $db->{key1} = "value1";
1238 $db->{key2} = "value2";
1239 print "key1: " . $db->{key1} . "\n";
1240 print "key2: " . $db->{key2} . "\n";
1246 return $cipher->encrypt( $_[0] );
1249 return $cipher->decrypt( $_[0] );
1252 =head2 REAL-TIME COMPRESSION EXAMPLE
1254 Here is a working example that uses the I<Compress::Zlib> module to do real-time
1255 compression / decompression of keys & values with DBM::Deep Filters.
1256 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
1257 more on I<Compress::Zlib>.
1262 my $db = DBM::Deep->new(
1263 file => "foo-compress.db",
1264 filter_store_key => \&my_compress,
1265 filter_store_value => \&my_compress,
1266 filter_fetch_key => \&my_decompress,
1267 filter_fetch_value => \&my_decompress,
1270 $db->{key1} = "value1";
1271 $db->{key2} = "value2";
1272 print "key1: " . $db->{key1} . "\n";
1273 print "key2: " . $db->{key2} . "\n";
1279 return Compress::Zlib::memGzip( $_[0] ) ;
1282 return Compress::Zlib::memGunzip( $_[0] ) ;
1285 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
1286 actually numerical index numbers, and are not filtered.
1288 =head1 ERROR HANDLING
1290 Most DBM::Deep methods return a true value for success, and call die() on
1291 failure. You can wrap calls in an eval block to catch the die.
1293 my $db = DBM::Deep->new( "foo.db" ); # create hash
1294 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
1296 print $@; # prints error message
1298 =head1 LARGEFILE SUPPORT
1300 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
1301 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
1302 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
1303 by specifying the 'pack_size' parameter when constructing the file.
1306 filename => $filename,
1307 pack_size => 'large',
1310 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
1311 instead of 32-bit longs. After setting these values your DB files have a
1312 theoretical maximum size of 16 XB (exabytes).
1314 You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file
1317 B<Note:> Changing these values will B<NOT> work for existing database files.
1318 Only change this for new files. Once the value has been set, it is stored in
1319 the file's header and cannot be changed for the life of the file. These
1320 parameters are per-file, meaning you can access 32-bit and 64-bit files, as
1323 B<Note:> We have not personally tested files larger than 2 GB -- all my
1324 systems have only a 32-bit Perl. However, I have received user reports that
1325 this does indeed work!
1327 =head1 LOW-LEVEL ACCESS
1329 If you require low-level access to the underlying filehandle that DBM::Deep uses,
1330 you can call the C<_fh()> method, which returns the handle:
1332 my $fh = $db->_fh();
1334 This method can be called on the root level of the datbase, or any child
1335 hashes or arrays. All levels share a I<root> structure, which contains things
1336 like the filehandle, a reference counter, and all the options specified
1337 when you created the object. You can get access to this file object by
1338 calling the C<_fileobj()> method.
1340 my $file_obj = $db->_fileobj();
1342 This is useful for changing options after the object has already been created,
1343 such as enabling/disabling locking. You can also store your own temporary user
1344 data in this structure (be wary of name collision), which is then accessible from
1345 any child hash or array.
1347 =head1 CUSTOM DIGEST ALGORITHM
1349 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
1350 keys. However you can override this, and use another algorithm (such as SHA-256)
1351 or even write your own. But please note that DBM::Deep currently expects zero
1352 collisions, so your algorithm has to be I<perfect>, so to speak. Collision
1353 detection may be introduced in a later version.
1355 You can specify a custom digest algorithm by passing it into the parameter
1356 list for new(), passing a reference to a subroutine as the 'digest' parameter,
1357 and the length of the algorithm's hashes (in bytes) as the 'hash_size'
1358 parameter. Here is a working example that uses a 256-bit hash from the
1359 I<Digest::SHA256> module. Please see
1360 L<http://search.cpan.org/search?module=Digest::SHA256> for more information.
1365 my $context = Digest::SHA256::new(256);
1367 my $db = DBM::Deep->new(
1368 filename => "foo-sha.db",
1369 digest => \&my_digest,
1373 $db->{key1} = "value1";
1374 $db->{key2} = "value2";
1375 print "key1: " . $db->{key1} . "\n";
1376 print "key2: " . $db->{key2} . "\n";
1382 return substr( $context->hash($_[0]), 0, 32 );
1385 B<Note:> Your returned digest strings must be B<EXACTLY> the number
1386 of bytes you specify in the hash_size parameter (in this case 32).
1388 B<Note:> If you do choose to use a custom digest algorithm, you must set it
1389 every time you access this file. Otherwise, the default (MD5) will be used.
1391 =head1 CIRCULAR REFERENCES
1393 DBM::Deep has B<experimental> support for circular references. Meaning you
1394 can have a nested hash key or array element that points to a parent object.
1395 This relationship is stored in the DB file, and is preserved between sessions.
1398 my $db = DBM::Deep->new( "foo.db" );
1401 $db->{circle} = $db; # ref to self
1403 print $db->{foo} . "\n"; # prints "bar"
1404 print $db->{circle}->{foo} . "\n"; # prints "bar" again
1406 B<Note>: Passing the object to a function that recursively walks the
1407 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
1408 C<export()> methods) will result in an infinite loop. This will be fixed in
1411 =head1 CAVEATS / ISSUES / BUGS
1413 This section describes all the known issues with DBM::Deep. It you have found
1414 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
1416 =head2 UNUSED SPACE RECOVERY
1418 One major caveat with DBM::Deep is that space occupied by existing keys and
1419 values is not recovered when they are deleted. Meaning if you keep deleting
1420 and adding new keys, your file will continuously grow. I am working on this,
1421 but in the meantime you can call the built-in C<optimize()> method from time to
1422 time (perhaps in a crontab or something) to recover all your unused space.
1424 $db->optimize(); # returns true on success
1426 This rebuilds the ENTIRE database into a new file, then moves it on top of
1427 the original. The new file will have no unused space, thus it will take up as
1428 little disk space as possible. Please note that this operation can take
1429 a long time for large files, and you need enough disk space to temporarily hold
1430 2 copies of your DB file. The temporary file is created in the same directory
1431 as the original, named with a ".tmp" extension, and is deleted when the
1432 operation completes. Oh, and if locking is enabled, the DB is automatically
1433 locked for the entire duration of the copy.
1435 B<WARNING:> Only call optimize() on the top-level node of the database, and
1436 make sure there are no child references lying around. DBM::Deep keeps a reference
1437 counter, and if it is greater than 1, optimize() will abort and return undef.
1441 (The reasons given assume a high level of Perl understanding, specifically of
1442 references. You can safely skip this section.)
1444 Currently, the only references supported are HASH and ARRAY. The other reference
1445 types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons.
1451 These are things like filehandles and other sockets. They can't be supported
1452 because it's completely unclear how DBM::Deep should serialize them.
1454 =item * SCALAR / REF
1456 The discussion here refers to the following type of example:
1463 # In some other process ...
1465 my $val = ${ $db->{key1} };
1467 is( $val, 50, "What actually gets stored in the DB file?" );
1469 The problem is one of synchronization. When the variable being referred to
1470 changes value, the reference isn't notified. This means that the new value won't
1471 be stored in the datafile for other processes to read. There is no TIEREF.
1473 It is theoretically possible to store references to values already within a
1474 DBM::Deep object because everything already is synchronized, but the change to
1475 the internals would be quite large. Specifically, DBM::Deep would have to tie
1476 every single value that is stored. This would bloat the RAM footprint of
1477 DBM::Deep at least twofold (if not more) and be a significant performance drain,
1478 all to support a feature that has never been requested.
1482 L<http://search.cpan.org/search?module=Data::Dump::Streamer> provides a
1483 mechanism for serializing coderefs, including saving off all closure state.
1484 However, just as for SCALAR and REF, that closure state may change without
1485 notifying the DBM::Deep object storing the reference.
1489 =head2 FILE CORRUPTION
1491 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
1492 for a 32-bit signature when opened, but other corruption in files can cause
1493 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
1494 stuck in an infinite loop depending on the level of corruption. File write
1495 operations are not checked for failure (for speed), so if you happen to run
1496 out of disk space, DBM::Deep will probably fail in a bad way. These things will
1497 be addressed in a later version of DBM::Deep.
1501 Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local
1502 filesystems, but will NOT protect you from file corruption over NFS. I've heard
1503 about setting up your NFS server with a locking daemon, then using lockf() to
1504 lock your files, but your mileage may vary there as well. From what I
1505 understand, there is no real way to do it. However, if you need access to the
1506 underlying filehandle in DBM::Deep for using some other kind of locking scheme like
1507 lockf(), see the L<LOW-LEVEL ACCESS> section above.
1509 =head2 COPYING OBJECTS
1511 Beware of copying tied objects in Perl. Very strange things can happen.
1512 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
1513 returns a new, blessed, tied hash or array to the same level in the DB.
1515 my $copy = $db->clone();
1517 B<Note>: Since clone() here is cloning the object, not the database location, any
1518 modifications to either $db or $copy will be visible in both.
1522 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
1523 These functions cause every element in the array to move, which can be murder
1524 on DBM::Deep, as every element has to be fetched from disk, then stored again in
1525 a different location. This will be addressed in the forthcoming version 1.00.
1527 =head2 WRITEONLY FILES
1529 If you pass in a filehandle to new(), you may have opened it in either a readonly or
1530 writeonly mode. STORE will verify that the filehandle is writable. However, there
1531 doesn't seem to be a good way to determine if a filehandle is readable. And, if the
1532 filehandle isn't readable, it's not clear what will happen. So, don't do that.
1536 This section discusses DBM::Deep's speed and memory usage.
1540 Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as
1541 the almighty I<BerkeleyDB>. But it makes up for it in features like true
1542 multi-level hash/array support, and cross-platform FTPable files. Even so,
1543 DBM::Deep is still pretty fast, and the speed stays fairly consistent, even
1544 with huge databases. Here is some test data:
1546 Adding 1,000,000 keys to new DB file...
1548 At 100 keys, avg. speed is 2,703 keys/sec
1549 At 200 keys, avg. speed is 2,642 keys/sec
1550 At 300 keys, avg. speed is 2,598 keys/sec
1551 At 400 keys, avg. speed is 2,578 keys/sec
1552 At 500 keys, avg. speed is 2,722 keys/sec
1553 At 600 keys, avg. speed is 2,628 keys/sec
1554 At 700 keys, avg. speed is 2,700 keys/sec
1555 At 800 keys, avg. speed is 2,607 keys/sec
1556 At 900 keys, avg. speed is 2,190 keys/sec
1557 At 1,000 keys, avg. speed is 2,570 keys/sec
1558 At 2,000 keys, avg. speed is 2,417 keys/sec
1559 At 3,000 keys, avg. speed is 1,982 keys/sec
1560 At 4,000 keys, avg. speed is 1,568 keys/sec
1561 At 5,000 keys, avg. speed is 1,533 keys/sec
1562 At 6,000 keys, avg. speed is 1,787 keys/sec
1563 At 7,000 keys, avg. speed is 1,977 keys/sec
1564 At 8,000 keys, avg. speed is 2,028 keys/sec
1565 At 9,000 keys, avg. speed is 2,077 keys/sec
1566 At 10,000 keys, avg. speed is 2,031 keys/sec
1567 At 20,000 keys, avg. speed is 1,970 keys/sec
1568 At 30,000 keys, avg. speed is 2,050 keys/sec
1569 At 40,000 keys, avg. speed is 2,073 keys/sec
1570 At 50,000 keys, avg. speed is 1,973 keys/sec
1571 At 60,000 keys, avg. speed is 1,914 keys/sec
1572 At 70,000 keys, avg. speed is 2,091 keys/sec
1573 At 80,000 keys, avg. speed is 2,103 keys/sec
1574 At 90,000 keys, avg. speed is 1,886 keys/sec
1575 At 100,000 keys, avg. speed is 1,970 keys/sec
1576 At 200,000 keys, avg. speed is 2,053 keys/sec
1577 At 300,000 keys, avg. speed is 1,697 keys/sec
1578 At 400,000 keys, avg. speed is 1,838 keys/sec
1579 At 500,000 keys, avg. speed is 1,941 keys/sec
1580 At 600,000 keys, avg. speed is 1,930 keys/sec
1581 At 700,000 keys, avg. speed is 1,735 keys/sec
1582 At 800,000 keys, avg. speed is 1,795 keys/sec
1583 At 900,000 keys, avg. speed is 1,221 keys/sec
1584 At 1,000,000 keys, avg. speed is 1,077 keys/sec
1586 This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl
1587 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and
1588 values were between 6 - 12 chars in length. The DB file ended up at 210MB.
1589 Run time was 12 min 3 sec.
1593 One of the great things about DBM::Deep is that it uses very little memory.
1594 Even with huge databases (1,000,000+ keys) you will not see much increased
1595 memory on your process. DBM::Deep relies solely on the filesystem for storing
1596 and fetching data. Here is output from I</usr/bin/top> before even opening a
1599 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1600 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl
1602 Basically the process is taking 2,716K of memory. And here is the same
1603 process after storing and fetching 1,000,000 keys:
1605 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1606 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl
1608 Notice the memory usage increased by only 56K. Test was performed on a 700mHz
1609 x86 box running Linux RedHat 7.2 & Perl 5.6.1.
1611 =head1 DB FILE FORMAT
1613 In case you were interested in the underlying DB file format, it is documented
1614 here in this section. You don't need to know this to use the module, it's just
1615 included for reference.
1619 DBM::Deep files always start with a 32-bit signature to identify the file type.
1620 This is at offset 0. The signature is "DPDB" in network byte order. This is
1621 checked for when the file is opened and an error will be thrown if it's not found.
1625 The DBM::Deep file is in a I<tagged format>, meaning each section of the file
1626 has a standard header containing the type of data, the length of data, and then
1627 the data itself. The type is a single character (1 byte), the length is a
1628 32-bit unsigned long in network byte order, and the data is, well, the data.
1629 Here is how it unfolds:
1633 Immediately after the 32-bit file signature is the I<Master Index> record.
1634 This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048
1635 bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array,
1636 depending on how the DBM::Deep object was constructed.
1638 The index works by looking at a I<MD5 Hash> of the hash key (or array index
1639 number). The first 8-bit char of the MD5 signature is the offset into the
1640 index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the
1641 index element is a file offset of the next tag for the key/element in question,
1642 which is usually a I<Bucket List> tag (see below).
1644 The next tag I<could> be another index, depending on how many keys/elements
1645 exist. See L<RE-INDEXING> below for details.
1649 A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus
1650 file offsets to where the actual data is stored. It starts with a standard
1651 tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or
1652 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus
1653 the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data.
1654 When the list fills up, a I<Re-Index> operation is performed (See
1655 L<RE-INDEXING> below).
1659 A I<Bucket> is a tag containing a key/value pair (in hash mode), or a
1660 index/value pair (in array mode). It starts with a standard tag header with
1661 type I<D> for scalar data (string, binary, etc.), or it could be a nested
1662 hash (type I<H>) or array (type I<A>). The value comes just after the tag
1663 header. The size reported in the tag header is only for the value, but then,
1664 just after the value is another size (32-bit unsigned long) and then the plain
1665 key itself. Since the value is likely to be fetched more often than the plain
1666 key, I figured it would be I<slightly> faster to store the value first.
1668 If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index>
1669 record for the nested structure, where the process begins all over again.
1673 After a I<Bucket List> grows to 16 records, its allocated space in the file is
1674 exhausted. Then, when another key/element comes in, the list is converted to a
1675 new index record. However, this index will look at the next char in the MD5
1676 hash, and arrange new Bucket List pointers accordingly. This process is called
1677 I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all
1678 17 (16 + new one) keys/elements are removed from the old Bucket List and
1679 inserted into the new index. Several new Bucket Lists are created in the
1680 process, as a new MD5 char from the key is being examined (it is unlikely that
1681 the keys will all share the same next char of their MD5s).
1683 Because of the way the I<MD5> algorithm works, it is impossible to tell exactly
1684 when the Bucket Lists will turn into indexes, but the first round tends to
1685 happen right around 4,000 keys. You will see a I<slight> decrease in
1686 performance here, but it picks back up pretty quick (see L<SPEED> above). Then
1687 it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's
1688 right around 900,000 keys. This process can continue nearly indefinitely --
1689 right up until the point the I<MD5> signatures start colliding with each other,
1690 and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND
1691 getting struck by lightning while you are walking to cash in your tickets.
1692 Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to
1693 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe
1694 this is 340 unodecillion, but don't quote me).
1698 When a new key/element is stored, the key (or index number) is first run through
1699 I<Digest::MD5> to get a 128-bit signature (example, in hex:
1700 b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked
1701 for the first char of the signature (in this case I<b0>). If it does not exist,
1702 a new I<Bucket List> is created for our key (and the next 15 future keys that
1703 happen to also have I<b> as their first MD5 char). The entire MD5 is written
1704 to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at
1705 this point, unless we are replacing an existing I<Bucket>), where the actual
1706 data will be stored.
1710 Fetching an existing key/element involves getting a I<Digest::MD5> of the key
1711 (or index number), then walking along the indexes. If there are enough
1712 keys/elements in this DB level, there might be nested indexes, each linked to
1713 a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which
1714 contains up to 16 full MD5 hashes. Each is checked for equality to the key in
1715 question. If we found a match, the I<Bucket> tag is loaded, where the value and
1716 plain key are stored.
1718 Fetching the plain key occurs when calling the I<first_key()> and I<next_key()>
1719 methods. In this process the indexes are walked systematically, and each key
1720 fetched in increasing MD5 order (which is why it appears random). Once the
1721 I<Bucket> is found, the value is skipped and the plain key returned instead.
1722 B<Note:> Do not count on keys being fetched as if the MD5 hashes were
1723 alphabetically sorted. This only happens on an index-level -- as soon as the
1724 I<Bucket Lists> are hit, the keys will come out in the order they went in --
1725 so it's pretty much undefined how the keys will come out -- just like Perl's
1728 =head1 CODE COVERAGE
1730 We use B<Devel::Cover> to test the code coverage of our tests, below is the
1731 B<Devel::Cover> report on this module's test suite.
1733 ----------------------------------- ------ ------ ------ ------ ------ ------
1734 File stmt bran cond sub time total
1735 ----------------------------------- ------ ------ ------ ------ ------ ------
1736 blib/lib/DBM/Deep.pm 94.9 80.6 73.0 100.0 37.9 90.4
1737 blib/lib/DBM/Deep/Array.pm 100.0 91.1 100.0 100.0 18.2 98.1
1738 blib/lib/DBM/Deep/Engine.pm 98.9 87.3 80.0 100.0 34.2 95.2
1739 blib/lib/DBM/Deep/Hash.pm 100.0 87.5 100.0 100.0 9.7 97.3
1740 Total 97.9 85.9 79.7 100.0 100.0 94.3
1741 ----------------------------------- ------ ------ ------ ------ ------ ------
1743 =head1 MORE INFORMATION
1745 Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
1746 or send email to L<DBM-Deep@googlegroups.com>.
1750 Joseph Huckaby, L<jhuckaby@cpan.org>
1752 Rob Kinyon, L<rkinyon@cpan.org>
1754 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
1758 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
1759 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
1763 Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved.
1764 This is free software, you may use it and distribute it under the
1765 same terms as Perl itself.