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.
34 use Fcntl qw( :DEFAULT :flock :seek );
38 use DBM::Deep::Engine;
40 use vars qw( $VERSION );
41 $VERSION = q(0.99_01);
44 # Set to 4 and 'N' for 32-bit offset tags (default). Theoretical limit of 4 GB per file.
45 # (Perl must be compiled with largefile support for files > 2 GB)
47 # Set to 8 and 'Q' for 64-bit offsets. Theoretical limit of 16 XB per file.
48 # (Perl must be compiled with largefile and 64-bit long support)
54 # Set to 4 and 'N' for 32-bit data length prefixes. Limit of 4 GB for each key/value.
55 # Upgrading this is possible (see above) but probably not necessary. If you need
56 # more than 4 GB for a single key or value, this module is really not for you :-)
58 #my $DATA_LENGTH_SIZE = 4;
59 #my $DATA_LENGTH_PACK = 'N';
60 our ($LONG_SIZE, $LONG_PACK, $DATA_LENGTH_SIZE, $DATA_LENGTH_PACK);
63 # Maximum number of buckets per list before another level of indexing is done.
64 # Increase this value for slightly greater speed, but larger database files.
65 # DO NOT decrease this value below 16, due to risk of recursive reindex overrun.
67 our $MAX_BUCKETS = 16;
70 # Better not adjust anything below here, unless you're me :-)
74 # Setup digest function for keys
76 our ($DIGEST_FUNC, $HASH_SIZE);
77 #my $DIGEST_FUNC = \&Digest::MD5::md5;
80 # Precalculate index and bucket sizes based on values above.
83 our ($INDEX_SIZE, $BUCKET_SIZE, $BUCKET_LIST_SIZE);
90 # Setup file and tag signatures. These should never change.
92 sub SIG_FILE () { 'DPDB' }
93 sub SIG_HASH () { 'H' }
94 sub SIG_ARRAY () { 'A' }
95 sub SIG_SCALAR () { 'S' }
96 sub SIG_NULL () { 'N' }
97 sub SIG_DATA () { 'D' }
98 sub SIG_INDEX () { 'I' }
99 sub SIG_BLIST () { 'B' }
100 sub SIG_SIZE () { 1 }
103 # Setup constants for users to pass to new()
105 sub TYPE_HASH () { SIG_HASH }
106 sub TYPE_ARRAY () { SIG_ARRAY }
107 sub TYPE_SCALAR () { SIG_SCALAR }
113 if (scalar(@_) > 1) {
115 $proto->_throw_error( "Odd number of parameters to " . (caller(1))[2] );
119 elsif ( ref $_[0] ) {
120 unless ( eval { local $SIG{'__DIE__'}; %{$_[0]} || 1 } ) {
121 $proto->_throw_error( "Not a hashref in args to " . (caller(1))[2] );
126 $args = { file => shift };
134 # Class constructor method for Perl OO interface.
135 # Calls tie() and returns blessed reference to tied hash or array,
136 # providing a hybrid OO/tie interface.
139 my $args = $class->_get_args( @_ );
142 # Check if we want a tied hash or array.
145 if (defined($args->{type}) && $args->{type} eq TYPE_ARRAY) {
146 $class = 'DBM::Deep::Array';
147 require DBM::Deep::Array;
148 tie @$self, $class, %$args;
151 $class = 'DBM::Deep::Hash';
152 require DBM::Deep::Hash;
153 tie %$self, $class, %$args;
156 return bless $self, $class;
161 # Setup $self and bless into this class.
166 # These are the defaults to be optionally overridden below
169 base_offset => length(SIG_FILE),
170 engine => 'DBM::Deep::Engine',
173 foreach my $param ( keys %$self ) {
174 next unless exists $args->{$param};
175 $self->{$param} = delete $args->{$param}
178 # locking implicitly enables autoflush
179 if ($args->{locking}) { $args->{autoflush} = 1; }
181 $self->{root} = exists $args->{root}
183 : DBM::Deep::_::Root->new( $args );
185 if (!defined($self->_fh)) { $self->{engine}->open( $self ); }
192 require DBM::Deep::Hash;
193 return DBM::Deep::Hash->TIEHASH( @_ );
198 require DBM::Deep::Array;
199 return DBM::Deep::Array->TIEARRAY( @_ );
202 #XXX Unneeded now ...
206 sub _find_bucket_list {
208 # Locate offset for bucket list, given digested key
214 # Locate offset for bucket list using digest index system
217 my $tag = $self->{engine}->load_tag($self, $self->_base_offset);
218 if (!$tag) { return; }
220 while ($tag->{signature} ne SIG_BLIST) {
221 $tag = $self->{engine}->index_lookup($self, $tag, ord(substr($md5, $ch, 1)));
222 if (!$tag) { return; }
229 sub _traverse_index {
231 # Scan index and recursively step into deeper levels, looking for next key.
233 my ($self, $offset, $ch, $force_return_next) = @_;
234 $force_return_next = undef unless $force_return_next;
236 my $tag = $self->{engine}->load_tag($self, $offset );
240 if ($tag->{signature} ne SIG_BLIST) {
241 my $content = $tag->{content};
243 if ($self->{return_next}) { $start = 0; }
244 else { $start = ord(substr($self->{prev_md5}, $ch, 1)); }
246 for (my $index = $start; $index < 256; $index++) {
247 my $subloc = unpack($LONG_PACK, substr($content, $index * $LONG_SIZE, $LONG_SIZE) );
249 my $result = $self->_traverse_index( $subloc, $ch + 1, $force_return_next );
250 if (defined($result)) { return $result; }
254 $self->{return_next} = 1;
257 elsif ($tag->{signature} eq SIG_BLIST) {
258 my $keys = $tag->{content};
259 if ($force_return_next) { $self->{return_next} = 1; }
262 # Iterate through buckets, looking for a key match
264 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
265 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
266 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
270 # End of bucket list -- return to outer loop
272 $self->{return_next} = 1;
275 elsif ($key eq $self->{prev_md5}) {
277 # Located previous key -- return next one found
279 $self->{return_next} = 1;
282 elsif ($self->{return_next}) {
284 # Seek to bucket location and skip over signature
286 seek($fh, $subloc + SIG_SIZE + $self->_root->{file_offset}, SEEK_SET);
289 # Skip over value to get to plain key
292 read( $fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
293 if ($size) { seek($fh, $size, SEEK_CUR); }
296 # Read in plain key and return as scalar
299 read( $fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
300 if ($size) { read( $fh, $plain_key, $size); }
306 $self->{return_next} = 1;
307 } # tag is a bucket list
314 # Locate next key, given digested previous one
316 my $self = $_[0]->_get_self;
318 $self->{prev_md5} = $_[1] ? $_[1] : undef;
319 $self->{return_next} = 0;
322 # If the previous key was not specifed, start at the top and
323 # return the first one found.
325 if (!$self->{prev_md5}) {
326 $self->{prev_md5} = chr(0) x $HASH_SIZE;
327 $self->{return_next} = 1;
330 return $self->_traverse_index( $self->_base_offset, 0 );
335 # If db locking is set, flock() the db file. If called multiple
336 # times before unlock(), then the same number of unlocks() must
337 # be called before the lock is released.
339 my $self = $_[0]->_get_self;
341 $type = LOCK_EX unless defined $type;
343 if (!defined($self->_fh)) { return; }
345 if ($self->_root->{locking}) {
346 if (!$self->_root->{locked}) {
347 flock($self->_fh, $type);
349 # refresh end counter in case file has changed size
350 my @stats = stat($self->_root->{file});
351 $self->_root->{end} = $stats[7];
353 # double-check file inode, in case another process
354 # has optimize()d our file while we were waiting.
355 if ($stats[1] != $self->_root->{inode}) {
356 $self->{engine}->open( $self ); # re-open
357 flock($self->_fh, $type); # re-lock
358 $self->_root->{end} = (stat($self->_fh))[7]; # re-end
361 $self->_root->{locked}++;
371 # If db locking is set, unlock the db file. See note in lock()
372 # regarding calling lock() multiple times.
374 my $self = $_[0]->_get_self;
376 if (!defined($self->_fh)) { return; }
378 if ($self->_root->{locking} && $self->_root->{locked} > 0) {
379 $self->_root->{locked}--;
380 if (!$self->_root->{locked}) { flock($self->_fh, LOCK_UN); }
389 my $self = shift->_get_self;
390 my ($spot, $value) = @_;
395 elsif ( eval { local $SIG{__DIE__}; $value->isa( 'DBM::Deep' ) } ) {
396 my $type = $value->_type;
397 ${$spot} = $type eq TYPE_HASH ? {} : [];
398 $value->_copy_node( ${$spot} );
401 my $r = Scalar::Util::reftype( $value );
402 my $c = Scalar::Util::blessed( $value );
403 if ( $r eq 'ARRAY' ) {
404 ${$spot} = [ @{$value} ];
407 ${$spot} = { %{$value} };
409 ${$spot} = bless ${$spot}, $c
418 # Copy single level of keys or elements to new DB handle.
419 # Recurse for nested structures
421 my $self = shift->_get_self;
424 if ($self->_type eq TYPE_HASH) {
425 my $key = $self->first_key();
427 my $value = $self->get($key);
428 $self->_copy_value( \$db_temp->{$key}, $value );
429 $key = $self->next_key($key);
433 my $length = $self->length();
434 for (my $index = 0; $index < $length; $index++) {
435 my $value = $self->get($index);
436 $self->_copy_value( \$db_temp->[$index], $value );
445 # Recursively export into standard Perl hashes and arrays.
447 my $self = $_[0]->_get_self;
450 if ($self->_type eq TYPE_HASH) { $temp = {}; }
451 elsif ($self->_type eq TYPE_ARRAY) { $temp = []; }
454 $self->_copy_node( $temp );
462 # Recursively import Perl hash/array structure
464 #XXX This use of ref() seems to be ok
465 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
467 my $self = $_[0]->_get_self;
470 #XXX This use of ref() seems to be ok
473 # struct is not a reference, so just import based on our type
477 if ($self->_type eq TYPE_HASH) { $struct = {@_}; }
478 elsif ($self->_type eq TYPE_ARRAY) { $struct = [@_]; }
481 my $r = Scalar::Util::reftype($struct) || '';
482 if ($r eq "HASH" && $self->_type eq TYPE_HASH) {
483 foreach my $key (keys %$struct) { $self->put($key, $struct->{$key}); }
485 elsif ($r eq "ARRAY" && $self->_type eq TYPE_ARRAY) {
486 $self->push( @$struct );
489 return $self->_throw_error("Cannot import: type mismatch");
497 # Rebuild entire database into new file, then move
498 # it back on top of original.
500 my $self = $_[0]->_get_self;
502 #XXX Need to create a new test for this
503 # if ($self->_root->{links} > 1) {
504 # return $self->_throw_error("Cannot optimize: reference count is greater than 1");
507 my $db_temp = DBM::Deep->new(
508 file => $self->_root->{file} . '.tmp',
512 return $self->_throw_error("Cannot optimize: failed to open temp file: $!");
516 $self->_copy_node( $db_temp );
520 # Attempt to copy user, group and permissions over to new file
522 my @stats = stat($self->_fh);
523 my $perms = $stats[2] & 07777;
526 chown( $uid, $gid, $self->_root->{file} . '.tmp' );
527 chmod( $perms, $self->_root->{file} . '.tmp' );
529 # q.v. perlport for more information on this variable
530 if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) {
532 # Potential race condition when optmizing on Win32 with locking.
533 # The Windows filesystem requires that the filehandle be closed
534 # before it is overwritten with rename(). This could be redone
538 $self->{engine}->close( $self );
541 if (!rename $self->_root->{file} . '.tmp', $self->_root->{file}) {
542 unlink $self->_root->{file} . '.tmp';
544 return $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!");
548 $self->{engine}->close( $self );
549 $self->{engine}->open( $self );
556 # Make copy of object and return
558 my $self = $_[0]->_get_self;
560 return DBM::Deep->new(
561 type => $self->_type,
562 base_offset => $self->_base_offset,
568 my %is_legal_filter = map {
571 store_key store_value
572 fetch_key fetch_value
577 # Setup filter function for storing or fetching the key or value
579 my $self = $_[0]->_get_self;
581 my $func = $_[2] ? $_[2] : undef;
583 if ( $is_legal_filter{$type} ) {
584 $self->_root->{"filter_$type"} = $func;
598 # Get access to the root structure
600 my $self = $_[0]->_get_self;
601 return $self->{root};
606 # Get access to the raw fh
608 #XXX It will be useful, though, when we split out HASH and ARRAY
609 my $self = $_[0]->_get_self;
610 return $self->_root->{fh};
615 # Get type of current node (TYPE_HASH or TYPE_ARRAY)
617 my $self = $_[0]->_get_self;
618 return $self->{type};
623 # Get base_offset of current node (TYPE_HASH or TYPE_ARRAY)
625 my $self = $_[0]->_get_self;
626 return $self->{base_offset};
634 die "DBM::Deep: $_[1]\n";
639 # Precalculate index, bucket and bucket list sizes
642 #XXX I don't like this ...
643 set_pack() unless defined $LONG_SIZE;
645 $INDEX_SIZE = 256 * $LONG_SIZE;
646 $BUCKET_SIZE = $HASH_SIZE + $LONG_SIZE;
647 $BUCKET_LIST_SIZE = $MAX_BUCKETS * $BUCKET_SIZE;
652 # Set pack/unpack modes (see file header for more)
654 my ($long_s, $long_p, $data_s, $data_p) = @_;
656 $LONG_SIZE = $long_s ? $long_s : 4;
657 $LONG_PACK = $long_p ? $long_p : 'N';
659 $DATA_LENGTH_SIZE = $data_s ? $data_s : 4;
660 $DATA_LENGTH_PACK = $data_p ? $data_p : 'N';
667 # Set key digest function (default is MD5)
669 my ($digest_func, $hash_size) = @_;
671 $DIGEST_FUNC = $digest_func ? $digest_func : \&Digest::MD5::md5;
672 $HASH_SIZE = $hash_size ? $hash_size : 16;
679 (O_WRONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
684 # (O_RDONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0);
688 # tie() methods (hashes and arrays)
693 # Store single hash key/value or array element in database.
695 my $self = $_[0]->_get_self;
698 # User may be storing a hash, in which case we do not want it run
699 # through the filtering system
700 my $value = ($self->_root->{filter_store_value} && !ref($_[2]))
701 ? $self->_root->{filter_store_value}->($_[2])
704 my $md5 = $DIGEST_FUNC->($key);
706 unless ( _is_writable( $self->_fh ) ) {
707 $self->_throw_error( 'Cannot write to a readonly filehandle' );
711 # Request exclusive lock for writing
713 $self->lock( LOCK_EX );
718 # Locate offset for bucket list using digest index system
720 my $tag = $self->{engine}->load_tag($self, $self->_base_offset);
722 $tag = $self->{engine}->create_tag($self, $self->_base_offset, SIG_INDEX, chr(0) x $INDEX_SIZE);
726 while ($tag->{signature} ne SIG_BLIST) {
727 my $num = ord(substr($md5, $ch, 1));
729 my $ref_loc = $tag->{offset} + ($num * $LONG_SIZE);
730 my $new_tag = $self->{engine}->index_lookup($self, $tag, $num);
733 seek($fh, $ref_loc + $self->_root->{file_offset}, SEEK_SET);
734 print( $fh pack($LONG_PACK, $self->_root->{end}) );
736 $tag = $self->{engine}->create_tag($self, $self->_root->{end}, SIG_BLIST, chr(0) x $BUCKET_LIST_SIZE);
738 $tag->{ref_loc} = $ref_loc;
746 $tag->{ref_loc} = $ref_loc;
753 # Add key/value to bucket list
755 my $result = $self->{engine}->add_bucket( $self, $tag, $md5, $key, $value );
764 # Fetch single value or element given plain key or array index
766 my $self = shift->_get_self;
769 my $md5 = $DIGEST_FUNC->($key);
772 # Request shared lock for reading
774 $self->lock( LOCK_SH );
776 my $tag = $self->_find_bucket_list( $md5 );
783 # Get value from bucket list
785 my $result = $self->{engine}->get_bucket_value( $self, $tag, $md5 );
789 #XXX What is ref() checking here?
790 #YYY Filters only apply on scalar values, so the ref check is making
791 #YYY sure the fetched bucket is a scalar, not a child hash or array.
792 return ($result && !ref($result) && $self->_root->{filter_fetch_value})
793 ? $self->_root->{filter_fetch_value}->($result)
799 # Delete single key/value pair or element given plain key or array index
801 my $self = $_[0]->_get_self;
804 my $md5 = $DIGEST_FUNC->($key);
807 # Request exclusive lock for writing
809 $self->lock( LOCK_EX );
811 my $tag = $self->_find_bucket_list( $md5 );
820 my $value = $self->{engine}->get_bucket_value($self, $tag, $md5 );
821 if ($value && !ref($value) && $self->_root->{filter_fetch_value}) {
822 $value = $self->_root->{filter_fetch_value}->($value);
825 my $result = $self->{engine}->delete_bucket( $self, $tag, $md5 );
828 # If this object is an array and the key deleted was on the end of the stack,
829 # decrement the length variable.
839 # Check if a single key or element exists given plain key or array index
841 my $self = $_[0]->_get_self;
844 my $md5 = $DIGEST_FUNC->($key);
847 # Request shared lock for reading
849 $self->lock( LOCK_SH );
851 my $tag = $self->_find_bucket_list( $md5 );
854 # For some reason, the built-in exists() function returns '' for false
862 # Check if bucket exists and return 1 or ''
864 my $result = $self->{engine}->bucket_exists( $self, $tag, $md5 ) || '';
873 # Clear all keys from hash, or all elements from array.
875 my $self = $_[0]->_get_self;
878 # Request exclusive lock for writing
880 $self->lock( LOCK_EX );
884 seek($fh, $self->_base_offset + $self->_root->{file_offset}, SEEK_SET);
890 $self->{engine}->create_tag($self, $self->_base_offset, $self->_type, chr(0) x $INDEX_SIZE);
898 # Public method aliases
900 sub put { (shift)->STORE( @_ ) }
901 sub store { (shift)->STORE( @_ ) }
902 sub get { (shift)->FETCH( @_ ) }
903 sub fetch { (shift)->FETCH( @_ ) }
904 sub delete { (shift)->DELETE( @_ ) }
905 sub exists { (shift)->EXISTS( @_ ) }
906 sub clear { (shift)->CLEAR( @_ ) }
908 package DBM::Deep::_::Root;
922 filter_store_key => undef,
923 filter_store_value => undef,
924 filter_fetch_key => undef,
925 filter_fetch_value => undef,
931 if ( $self->{fh} && !$self->{file_offset} ) {
932 $self->{file_offset} = tell( $self->{fh} );
942 close $self->{fh} if $self->{fh};
953 DBM::Deep - A pure perl multi-level hash/array DBM
958 my $db = DBM::Deep->new( "foo.db" );
960 $db->{key} = 'value'; # tie() style
963 $db->put('key' => 'value'); # OO style
964 print $db->get('key');
966 # true multi-level support
967 $db->{my_complex} = [
968 'hello', { perl => 'rules' },
974 A unique flat-file database module, written in pure perl. True
975 multi-level hash/array support (unlike MLDBM, which is faked), hybrid
976 OO / tie() interface, cross-platform FTPable files, and quite fast. Can
977 handle millions of keys and unlimited hash levels without significant
978 slow-down. Written from the ground-up in pure perl -- this is NOT a
979 wrapper around a C-based DBM. Out-of-the-box compatibility with Unix,
980 Mac OS X and Windows.
984 Hopefully you are using Perl's excellent CPAN module, which will download
985 and install the module for you. If not, get the tarball, and run these
997 Construction can be done OO-style (which is the recommended way), or using
998 Perl's tie() function. Both are examined here.
1000 =head2 OO CONSTRUCTION
1002 The recommended way to construct a DBM::Deep object is to use the new()
1003 method, which gets you a blessed, tied hash or array reference.
1005 my $db = DBM::Deep->new( "foo.db" );
1007 This opens a new database handle, mapped to the file "foo.db". If this
1008 file does not exist, it will automatically be created. DB files are
1009 opened in "r+" (read/write) mode, and the type of object returned is a
1010 hash, unless otherwise specified (see L<OPTIONS> below).
1012 You can pass a number of options to the constructor to specify things like
1013 locking, autoflush, etc. This is done by passing an inline hash:
1015 my $db = DBM::Deep->new(
1021 Notice that the filename is now specified I<inside> the hash with
1022 the "file" parameter, as opposed to being the sole argument to the
1023 constructor. This is required if any options are specified.
1024 See L<OPTIONS> below for the complete list.
1028 You can also start with an array instead of a hash. For this, you must
1029 specify the C<type> parameter:
1031 my $db = DBM::Deep->new(
1033 type => DBM::Deep->TYPE_ARRAY
1036 B<Note:> Specifing the C<type> parameter only takes effect when beginning
1037 a new DB file. If you create a DBM::Deep object with an existing file, the
1038 C<type> will be loaded from the file header, and an error will be thrown if
1039 the wrong type is passed in.
1041 =head2 TIE CONSTRUCTION
1043 Alternately, you can create a DBM::Deep handle by using Perl's built-in
1044 tie() function. The object returned from tie() can be used to call methods,
1045 such as lock() and unlock(), but cannot be used to assign to the DBM::Deep
1046 file (as expected with most tie'd objects).
1049 my $db = tie %hash, "DBM::Deep", "foo.db";
1052 my $db = tie @array, "DBM::Deep", "bar.db";
1054 As with the OO constructor, you can replace the DB filename parameter with
1055 a hash containing one or more options (see L<OPTIONS> just below for the
1058 tie %hash, "DBM::Deep", {
1066 There are a number of options that can be passed in when constructing your
1067 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
1073 Filename of the DB file to link the handle to. You can pass a full absolute
1074 filesystem path, partial path, or a plain filename if the file is in the
1075 current working directory. This is a required parameter (though q.v. fh).
1079 If you want, you can pass in the fh instead of the file. This is most useful for doing
1082 my $db = DBM::Deep->new( { fh => \*DATA } );
1084 You are responsible for making sure that the fh has been opened appropriately for your
1085 needs. If you open it read-only and attempt to write, an exception will be thrown. If you
1086 open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
1087 needs to read from the fh.
1091 This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
1092 not need to set this. However, it's there if you want it.
1094 If you pass in fh and do not set this, it will be set appropriately.
1098 This parameter specifies what type of object to create, a hash or array. Use
1099 one of these two constants: C<DBM::Deep-E<gt>TYPE_HASH> or C<DBM::Deep-E<gt>TYPE_ARRAY>.
1100 This only takes effect when beginning a new file. This is an optional
1101 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
1105 Specifies whether locking is to be enabled. DBM::Deep uses Perl's Fnctl flock()
1106 function to lock the database in exclusive mode for writes, and shared mode for
1107 reads. Pass any true value to enable. This affects the base DB handle I<and
1108 any child hashes or arrays> that use the same DB file. This is an optional
1109 parameter, and defaults to 0 (disabled). See L<LOCKING> below for more.
1113 Specifies whether autoflush is to be enabled on the underlying filehandle.
1114 This obviously slows down write operations, but is required if you may have
1115 multiple processes accessing the same DB file (also consider enable I<locking>).
1116 Pass any true value to enable. This is an optional parameter, and defaults to 0
1121 If I<autobless> mode is enabled, DBM::Deep will preserve blessed hashes, and
1122 restore them when fetched. This is an B<experimental> feature, and does have
1123 side-effects. Basically, when hashes are re-blessed into their original
1124 classes, they are no longer blessed into the DBM::Deep class! So you won't be
1125 able to call any DBM::Deep methods on them. You have been warned.
1126 This is an optional parameter, and defaults to 0 (disabled).
1130 See L<FILTERS> below.
1134 Setting I<debug> mode will make all errors non-fatal, dump them out to
1135 STDERR, and continue on. This is for debugging purposes only, and probably
1136 not what you want. This is an optional parameter, and defaults to 0 (disabled).
1138 B<NOTE>: This parameter is considered deprecated and should not be used anymore.
1142 =head1 TIE INTERFACE
1144 With DBM::Deep you can access your databases using Perl's standard hash/array
1145 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
1146 treat them as such. DBM::Deep will intercept all reads/writes and direct them
1147 to the right place -- the DB file. This has nothing to do with the
1148 L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
1149 using regular hashes and arrays, rather than calling functions like C<get()>
1150 and C<put()> (although those work too). It is entirely up to you how to want
1151 to access your databases.
1155 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
1156 or even nested hashes (or arrays) using standard Perl syntax:
1158 my $db = DBM::Deep->new( "foo.db" );
1160 $db->{mykey} = "myvalue";
1162 $db->{myhash}->{subkey} = "subvalue";
1164 print $db->{myhash}->{subkey} . "\n";
1166 You can even step through hash keys using the normal Perl C<keys()> function:
1168 foreach my $key (keys %$db) {
1169 print "$key: " . $db->{$key} . "\n";
1172 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
1173 pushes them onto an array, all before the loop even begins. If you have an
1174 extra large hash, this may exhaust Perl's memory. Instead, consider using
1175 Perl's C<each()> function, which pulls keys/values one at a time, using very
1178 while (my ($key, $value) = each %$db) {
1179 print "$key: $value\n";
1182 Please note that when using C<each()>, you should always pass a direct
1183 hash reference, not a lookup. Meaning, you should B<never> do this:
1186 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
1188 This causes an infinite loop, because for each iteration, Perl is calling
1189 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
1190 it effectively keeps returning the first key over and over again. Instead,
1191 assign a temporary variable to C<$db->{foo}>, then pass that to each().
1195 As with hashes, you can treat any DBM::Deep object like a normal Perl array
1196 reference. This includes inserting, removing and manipulating elements,
1197 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
1198 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
1199 or simply be a nested array reference inside a hash. Example:
1201 my $db = DBM::Deep->new(
1202 file => "foo-array.db",
1203 type => DBM::Deep->TYPE_ARRAY
1207 push @$db, "bar", "baz";
1208 unshift @$db, "bah";
1210 my $last_elem = pop @$db; # baz
1211 my $first_elem = shift @$db; # bah
1212 my $second_elem = $db->[1]; # bar
1214 my $num_elements = scalar @$db;
1218 In addition to the I<tie()> interface, you can also use a standard OO interface
1219 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
1220 array) has its own methods, but both types share the following common methods:
1221 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>.
1225 =item * new() / clone()
1227 These are the constructor and copy-functions.
1229 =item * put() / store()
1231 Stores a new hash key/value pair, or sets an array element value. Takes two
1232 arguments, the hash key or array index, and the new value. The value can be
1233 a scalar, hash ref or array ref. Returns true on success, false on failure.
1235 $db->put("foo", "bar"); # for hashes
1236 $db->put(1, "bar"); # for arrays
1238 =item * get() / fetch()
1240 Fetches the value of a hash key or array element. Takes one argument: the hash
1241 key or array index. Returns a scalar, hash ref or array ref, depending on the
1244 my $value = $db->get("foo"); # for hashes
1245 my $value = $db->get(1); # for arrays
1249 Checks if a hash key or array index exists. Takes one argument: the hash key
1250 or array index. Returns true if it exists, false if not.
1252 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
1253 if ($db->exists(1)) { print "yay!\n"; } # for arrays
1257 Deletes one hash key/value pair or array element. Takes one argument: the hash
1258 key or array index. Returns true on success, false if not found. For arrays,
1259 the remaining elements located after the deleted element are NOT moved over.
1260 The deleted element is essentially just undefined, which is exactly how Perl's
1261 internal arrays work. Please note that the space occupied by the deleted
1262 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
1263 below for details and workarounds.
1265 $db->delete("foo"); # for hashes
1266 $db->delete(1); # for arrays
1270 Deletes B<all> hash keys or array elements. Takes no arguments. No return
1271 value. Please note that the space occupied by the deleted keys/values or
1272 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
1273 details and workarounds.
1275 $db->clear(); # hashes or arrays
1277 =item * lock() / unlock()
1283 Recover lost disk space.
1285 =item * import() / export()
1287 Data going in and out.
1289 =item * set_digest() / set_pack() / set_filter()
1291 q.v. adjusting the interal parameters.
1297 For hashes, DBM::Deep supports all the common methods described above, and the
1298 following additional methods: C<first_key()> and C<next_key()>.
1304 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
1305 fetched in an undefined order (which appears random). Takes no arguments,
1306 returns the key as a scalar value.
1308 my $key = $db->first_key();
1312 Returns the "next" key in the hash, given the previous one as the sole argument.
1313 Returns undef if there are no more keys to be fetched.
1315 $key = $db->next_key($key);
1319 Here are some examples of using hashes:
1321 my $db = DBM::Deep->new( "foo.db" );
1323 $db->put("foo", "bar");
1324 print "foo: " . $db->get("foo") . "\n";
1326 $db->put("baz", {}); # new child hash ref
1327 $db->get("baz")->put("buz", "biz");
1328 print "buz: " . $db->get("baz")->get("buz") . "\n";
1330 my $key = $db->first_key();
1332 print "$key: " . $db->get($key) . "\n";
1333 $key = $db->next_key($key);
1336 if ($db->exists("foo")) { $db->delete("foo"); }
1340 For arrays, DBM::Deep supports all the common methods described above, and the
1341 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
1342 C<unshift()> and C<splice()>.
1348 Returns the number of elements in the array. Takes no arguments.
1350 my $len = $db->length();
1354 Adds one or more elements onto the end of the array. Accepts scalars, hash
1355 refs or array refs. No return value.
1357 $db->push("foo", "bar", {});
1361 Fetches the last element in the array, and deletes it. Takes no arguments.
1362 Returns undef if array is empty. Returns the element value.
1364 my $elem = $db->pop();
1368 Fetches the first element in the array, deletes it, then shifts all the
1369 remaining elements over to take up the space. Returns the element value. This
1370 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
1373 my $elem = $db->shift();
1377 Inserts one or more elements onto the beginning of the array, shifting all
1378 existing elements over to make room. Accepts scalars, hash refs or array refs.
1379 No return value. This method is not recommended with large arrays -- see
1380 <LARGE ARRAYS> below for details.
1382 $db->unshift("foo", "bar", {});
1386 Performs exactly like Perl's built-in function of the same name. See L<perldoc
1387 -f splice> for usage -- it is too complicated to document here. This method is
1388 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
1392 Here are some examples of using arrays:
1394 my $db = DBM::Deep->new(
1396 type => DBM::Deep->TYPE_ARRAY
1399 $db->push("bar", "baz");
1400 $db->unshift("foo");
1403 my $len = $db->length();
1404 print "length: $len\n"; # 4
1406 for (my $k=0; $k<$len; $k++) {
1407 print "$k: " . $db->get($k) . "\n";
1410 $db->splice(1, 2, "biz", "baf");
1412 while (my $elem = shift @$db) {
1413 print "shifted: $elem\n";
1418 Enable automatic file locking by passing a true value to the C<locking>
1419 parameter when constructing your DBM::Deep object (see L<SETUP> above).
1421 my $db = DBM::Deep->new(
1426 This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
1427 mode for writes, and shared mode for reads. This is required if you have
1428 multiple processes accessing the same database file, to avoid file corruption.
1429 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
1430 NFS> below for more.
1432 =head2 EXPLICIT LOCKING
1434 You can explicitly lock a database, so it remains locked for multiple
1435 transactions. This is done by calling the C<lock()> method, and passing an
1436 optional lock mode argument (defaults to exclusive mode). This is particularly
1437 useful for things like counters, where the current value needs to be fetched,
1438 then incremented, then stored again.
1441 my $counter = $db->get("counter");
1443 $db->put("counter", $counter);
1452 You can pass C<lock()> an optional argument, which specifies which mode to use
1453 (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX>
1454 or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the
1455 same as the constants defined in Perl's C<Fcntl> module.
1457 $db->lock( DBM::Deep->LOCK_SH );
1461 =head1 IMPORTING/EXPORTING
1463 You can import existing complex structures by calling the C<import()> method,
1464 and export an entire database into an in-memory structure using the C<export()>
1465 method. Both are examined here.
1469 Say you have an existing hash with nested hashes/arrays inside it. Instead of
1470 walking the structure and adding keys/elements to the database as you go,
1471 simply pass a reference to the C<import()> method. This recursively adds
1472 everything to an existing DBM::Deep object for you. Here is an example:
1477 array1 => [ "elem0", "elem1", "elem2" ],
1479 subkey1 => "subvalue1",
1480 subkey2 => "subvalue2"
1484 my $db = DBM::Deep->new( "foo.db" );
1485 $db->import( $struct );
1487 print $db->{key1} . "\n"; # prints "value1"
1489 This recursively imports the entire C<$struct> object into C<$db>, including
1490 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
1491 keys are merged with the existing ones, replacing if they already exist.
1492 The C<import()> method can be called on any database level (not just the base
1493 level), and works with both hash and array DB types.
1495 B<Note:> Make sure your existing structure has no circular references in it.
1496 These will cause an infinite loop when importing.
1500 Calling the C<export()> method on an existing DBM::Deep object will return
1501 a reference to a new in-memory copy of the database. The export is done
1502 recursively, so all nested hashes/arrays are all exported to standard Perl
1503 objects. Here is an example:
1505 my $db = DBM::Deep->new( "foo.db" );
1507 $db->{key1} = "value1";
1508 $db->{key2} = "value2";
1510 $db->{hash1}->{subkey1} = "subvalue1";
1511 $db->{hash1}->{subkey2} = "subvalue2";
1513 my $struct = $db->export();
1515 print $struct->{key1} . "\n"; # prints "value1"
1517 This makes a complete copy of the database in memory, and returns a reference
1518 to it. The C<export()> method can be called on any database level (not just
1519 the base level), and works with both hash and array DB types. Be careful of
1520 large databases -- you can store a lot more data in a DBM::Deep object than an
1521 in-memory Perl structure.
1523 B<Note:> Make sure your database has no circular references in it.
1524 These will cause an infinite loop when exporting.
1528 DBM::Deep has a number of hooks where you can specify your own Perl function
1529 to perform filtering on incoming or outgoing data. This is a perfect
1530 way to extend the engine, and implement things like real-time compression or
1531 encryption. Filtering applies to the base DB level, and all child hashes /
1532 arrays. Filter hooks can be specified when your DBM::Deep object is first
1533 constructed, or by calling the C<set_filter()> method at any time. There are
1534 four available filter hooks, described below:
1538 =item * filter_store_key
1540 This filter is called whenever a hash key is stored. It
1541 is passed the incoming key, and expected to return a transformed key.
1543 =item * filter_store_value
1545 This filter is called whenever a hash key or array element is stored. It
1546 is passed the incoming value, and expected to return a transformed value.
1548 =item * filter_fetch_key
1550 This filter is called whenever a hash key is fetched (i.e. via
1551 C<first_key()> or C<next_key()>). It is passed the transformed key,
1552 and expected to return the plain key.
1554 =item * filter_fetch_value
1556 This filter is called whenever a hash key or array element is fetched.
1557 It is passed the transformed value, and expected to return the plain value.
1561 Here are the two ways to setup a filter hook:
1563 my $db = DBM::Deep->new(
1565 filter_store_value => \&my_filter_store,
1566 filter_fetch_value => \&my_filter_fetch
1571 $db->set_filter( "filter_store_value", \&my_filter_store );
1572 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
1574 Your filter function will be called only when dealing with SCALAR keys or
1575 values. When nested hashes and arrays are being stored/fetched, filtering
1576 is bypassed. Filters are called as static functions, passed a single SCALAR
1577 argument, and expected to return a single SCALAR value. If you want to
1578 remove a filter, set the function reference to C<undef>:
1580 $db->set_filter( "filter_store_value", undef );
1582 =head2 REAL-TIME ENCRYPTION EXAMPLE
1584 Here is a working example that uses the I<Crypt::Blowfish> module to
1585 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
1586 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
1587 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
1590 use Crypt::Blowfish;
1593 my $cipher = Crypt::CBC->new({
1594 'key' => 'my secret key',
1595 'cipher' => 'Blowfish',
1597 'regenerate_key' => 0,
1598 'padding' => 'space',
1602 my $db = DBM::Deep->new(
1603 file => "foo-encrypt.db",
1604 filter_store_key => \&my_encrypt,
1605 filter_store_value => \&my_encrypt,
1606 filter_fetch_key => \&my_decrypt,
1607 filter_fetch_value => \&my_decrypt,
1610 $db->{key1} = "value1";
1611 $db->{key2} = "value2";
1612 print "key1: " . $db->{key1} . "\n";
1613 print "key2: " . $db->{key2} . "\n";
1619 return $cipher->encrypt( $_[0] );
1622 return $cipher->decrypt( $_[0] );
1625 =head2 REAL-TIME COMPRESSION EXAMPLE
1627 Here is a working example that uses the I<Compress::Zlib> module to do real-time
1628 compression / decompression of keys & values with DBM::Deep Filters.
1629 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
1630 more on I<Compress::Zlib>.
1635 my $db = DBM::Deep->new(
1636 file => "foo-compress.db",
1637 filter_store_key => \&my_compress,
1638 filter_store_value => \&my_compress,
1639 filter_fetch_key => \&my_decompress,
1640 filter_fetch_value => \&my_decompress,
1643 $db->{key1} = "value1";
1644 $db->{key2} = "value2";
1645 print "key1: " . $db->{key1} . "\n";
1646 print "key2: " . $db->{key2} . "\n";
1652 return Compress::Zlib::memGzip( $_[0] ) ;
1655 return Compress::Zlib::memGunzip( $_[0] ) ;
1658 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
1659 actually numerical index numbers, and are not filtered.
1661 =head1 ERROR HANDLING
1663 Most DBM::Deep methods return a true value for success, and call die() on
1664 failure. You can wrap calls in an eval block to catch the die.
1666 my $db = DBM::Deep->new( "foo.db" ); # create hash
1667 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
1669 print $@; # prints error message
1671 =head1 LARGEFILE SUPPORT
1673 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
1674 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
1675 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
1676 by calling the static C<set_pack()> method before you do anything else.
1678 DBM::Deep::set_pack(8, 'Q');
1680 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
1681 instead of 32-bit longs. After setting these values your DB files have a
1682 theoretical maximum size of 16 XB (exabytes).
1684 B<Note:> Changing these values will B<NOT> work for existing database files.
1685 Only change this for new files, and make sure it stays set consistently
1686 throughout the file's life. If you do set these values, you can no longer
1687 access 32-bit DB files. You can, however, call C<set_pack(4, 'N')> to change
1688 back to 32-bit mode.
1690 B<Note:> I have not personally tested files > 2 GB -- all my systems have
1691 only a 32-bit Perl. However, I have received user reports that this does
1694 =head1 LOW-LEVEL ACCESS
1696 If you require low-level access to the underlying filehandle that DBM::Deep uses,
1697 you can call the C<_fh()> method, which returns the handle:
1699 my $fh = $db->_fh();
1701 This method can be called on the root level of the datbase, or any child
1702 hashes or arrays. All levels share a I<root> structure, which contains things
1703 like the filehandle, a reference counter, and all the options specified
1704 when you created the object. You can get access to this root structure by
1705 calling the C<root()> method.
1707 my $root = $db->_root();
1709 This is useful for changing options after the object has already been created,
1710 such as enabling/disabling locking, or debug modes. You can also
1711 store your own temporary user data in this structure (be wary of name
1712 collision), which is then accessible from any child hash or array.
1714 =head1 CUSTOM DIGEST ALGORITHM
1716 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
1717 keys. However you can override this, and use another algorithm (such as SHA-256)
1718 or even write your own. But please note that DBM::Deep currently expects zero
1719 collisions, so your algorithm has to be I<perfect>, so to speak.
1720 Collision detection may be introduced in a later version.
1724 You can specify a custom digest algorithm by calling the static C<set_digest()>
1725 function, passing a reference to a subroutine, and the length of the algorithm's
1726 hashes (in bytes). This is a global static function, which affects ALL DBM::Deep
1727 objects. Here is a working example that uses a 256-bit hash from the
1728 I<Digest::SHA256> module. Please see
1729 L<http://search.cpan.org/search?module=Digest::SHA256> for more.
1734 my $context = Digest::SHA256::new(256);
1736 DBM::Deep::set_digest( \&my_digest, 32 );
1738 my $db = DBM::Deep->new( "foo-sha.db" );
1740 $db->{key1} = "value1";
1741 $db->{key2} = "value2";
1742 print "key1: " . $db->{key1} . "\n";
1743 print "key2: " . $db->{key2} . "\n";
1749 return substr( $context->hash($_[0]), 0, 32 );
1752 B<Note:> Your returned digest strings must be B<EXACTLY> the number
1753 of bytes you specify in the C<set_digest()> function (in this case 32).
1755 =head1 CIRCULAR REFERENCES
1757 DBM::Deep has B<experimental> support for circular references. Meaning you
1758 can have a nested hash key or array element that points to a parent object.
1759 This relationship is stored in the DB file, and is preserved between sessions.
1762 my $db = DBM::Deep->new( "foo.db" );
1765 $db->{circle} = $db; # ref to self
1767 print $db->{foo} . "\n"; # prints "foo"
1768 print $db->{circle}->{foo} . "\n"; # prints "foo" again
1770 One catch is, passing the object to a function that recursively walks the
1771 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
1772 C<export()> methods) will result in an infinite loop. The other catch is,
1773 if you fetch the I<key> of a circular reference (i.e. using the C<first_key()>
1774 or C<next_key()> methods), you will get the I<target object's key>, not the
1775 ref's key. This gets even more interesting with the above example, where
1776 the I<circle> key points to the base DB object, which technically doesn't
1777 have a key. So I made DBM::Deep return "[base]" as the key name in that
1780 =head1 CAVEATS / ISSUES / BUGS
1782 This section describes all the known issues with DBM::Deep. It you have found
1783 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
1785 =head2 UNUSED SPACE RECOVERY
1787 One major caveat with DBM::Deep is that space occupied by existing keys and
1788 values is not recovered when they are deleted. Meaning if you keep deleting
1789 and adding new keys, your file will continuously grow. I am working on this,
1790 but in the meantime you can call the built-in C<optimize()> method from time to
1791 time (perhaps in a crontab or something) to recover all your unused space.
1793 $db->optimize(); # returns true on success
1795 This rebuilds the ENTIRE database into a new file, then moves it on top of
1796 the original. The new file will have no unused space, thus it will take up as
1797 little disk space as possible. Please note that this operation can take
1798 a long time for large files, and you need enough disk space to temporarily hold
1799 2 copies of your DB file. The temporary file is created in the same directory
1800 as the original, named with a ".tmp" extension, and is deleted when the
1801 operation completes. Oh, and if locking is enabled, the DB is automatically
1802 locked for the entire duration of the copy.
1804 B<WARNING:> Only call optimize() on the top-level node of the database, and
1805 make sure there are no child references lying around. DBM::Deep keeps a reference
1806 counter, and if it is greater than 1, optimize() will abort and return undef.
1808 =head2 AUTOVIVIFICATION
1810 Unfortunately, autovivification doesn't work with tied hashes. This appears to
1811 be a bug in Perl's tie() system, as I<Jakob Schmidt> encountered the very same
1812 issue with his I<DWH_FIle> module (see L<http://search.cpan.org/search?module=DWH_File>),
1813 and it is also mentioned in the BUGS section for the I<MLDBM> module <see
1814 L<http://search.cpan.org/search?module=MLDBM>). Basically, on a new db file,
1817 $db->{foo}->{bar} = "hello";
1819 Since "foo" doesn't exist, you cannot add "bar" to it. You end up with "foo"
1820 being an empty hash. Try this instead, which works fine:
1822 $db->{foo} = { bar => "hello" };
1824 As of Perl 5.8.7, this bug still exists. I have walked very carefully through
1825 the execution path, and Perl indeed passes an empty hash to the STORE() method.
1826 Probably a bug in Perl.
1828 =head2 FILE CORRUPTION
1830 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
1831 for a 32-bit signature when opened, but other corruption in files can cause
1832 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
1833 stuck in an infinite loop depending on the level of corruption. File write
1834 operations are not checked for failure (for speed), so if you happen to run
1835 out of disk space, DBM::Deep will probably fail in a bad way. These things will
1836 be addressed in a later version of DBM::Deep.
1840 Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local
1841 filesystems, but will NOT protect you from file corruption over NFS. I've heard
1842 about setting up your NFS server with a locking daemon, then using lockf() to
1843 lock your files, but your mileage may vary there as well. From what I
1844 understand, there is no real way to do it. However, if you need access to the
1845 underlying filehandle in DBM::Deep for using some other kind of locking scheme like
1846 lockf(), see the L<LOW-LEVEL ACCESS> section above.
1848 =head2 COPYING OBJECTS
1850 Beware of copying tied objects in Perl. Very strange things can happen.
1851 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
1852 returns a new, blessed, tied hash or array to the same level in the DB.
1854 my $copy = $db->clone();
1856 B<Note>: Since clone() here is cloning the object, not the database location, any
1857 modifications to either $db or $copy will be visible in both.
1861 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
1862 These functions cause every element in the array to move, which can be murder
1863 on DBM::Deep, as every element has to be fetched from disk, then stored again in
1864 a different location. This will be addressed in the forthcoming version 1.00.
1866 =head2 WRITEONLY FILES
1868 If you pass in a filehandle to new(), you may have opened it in either a readonly or
1869 writeonly mode. STORE will verify that the filehandle is writable. However, there
1870 doesn't seem to be a good way to determine if a filehandle is readable. And, if the
1871 filehandle isn't readable, it's not clear what will happen. So, don't do that.
1875 This section discusses DBM::Deep's speed and memory usage.
1879 Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as
1880 the almighty I<BerkeleyDB>. But it makes up for it in features like true
1881 multi-level hash/array support, and cross-platform FTPable files. Even so,
1882 DBM::Deep is still pretty fast, and the speed stays fairly consistent, even
1883 with huge databases. Here is some test data:
1885 Adding 1,000,000 keys to new DB file...
1887 At 100 keys, avg. speed is 2,703 keys/sec
1888 At 200 keys, avg. speed is 2,642 keys/sec
1889 At 300 keys, avg. speed is 2,598 keys/sec
1890 At 400 keys, avg. speed is 2,578 keys/sec
1891 At 500 keys, avg. speed is 2,722 keys/sec
1892 At 600 keys, avg. speed is 2,628 keys/sec
1893 At 700 keys, avg. speed is 2,700 keys/sec
1894 At 800 keys, avg. speed is 2,607 keys/sec
1895 At 900 keys, avg. speed is 2,190 keys/sec
1896 At 1,000 keys, avg. speed is 2,570 keys/sec
1897 At 2,000 keys, avg. speed is 2,417 keys/sec
1898 At 3,000 keys, avg. speed is 1,982 keys/sec
1899 At 4,000 keys, avg. speed is 1,568 keys/sec
1900 At 5,000 keys, avg. speed is 1,533 keys/sec
1901 At 6,000 keys, avg. speed is 1,787 keys/sec
1902 At 7,000 keys, avg. speed is 1,977 keys/sec
1903 At 8,000 keys, avg. speed is 2,028 keys/sec
1904 At 9,000 keys, avg. speed is 2,077 keys/sec
1905 At 10,000 keys, avg. speed is 2,031 keys/sec
1906 At 20,000 keys, avg. speed is 1,970 keys/sec
1907 At 30,000 keys, avg. speed is 2,050 keys/sec
1908 At 40,000 keys, avg. speed is 2,073 keys/sec
1909 At 50,000 keys, avg. speed is 1,973 keys/sec
1910 At 60,000 keys, avg. speed is 1,914 keys/sec
1911 At 70,000 keys, avg. speed is 2,091 keys/sec
1912 At 80,000 keys, avg. speed is 2,103 keys/sec
1913 At 90,000 keys, avg. speed is 1,886 keys/sec
1914 At 100,000 keys, avg. speed is 1,970 keys/sec
1915 At 200,000 keys, avg. speed is 2,053 keys/sec
1916 At 300,000 keys, avg. speed is 1,697 keys/sec
1917 At 400,000 keys, avg. speed is 1,838 keys/sec
1918 At 500,000 keys, avg. speed is 1,941 keys/sec
1919 At 600,000 keys, avg. speed is 1,930 keys/sec
1920 At 700,000 keys, avg. speed is 1,735 keys/sec
1921 At 800,000 keys, avg. speed is 1,795 keys/sec
1922 At 900,000 keys, avg. speed is 1,221 keys/sec
1923 At 1,000,000 keys, avg. speed is 1,077 keys/sec
1925 This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl
1926 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and
1927 values were between 6 - 12 chars in length. The DB file ended up at 210MB.
1928 Run time was 12 min 3 sec.
1932 One of the great things about DBM::Deep is that it uses very little memory.
1933 Even with huge databases (1,000,000+ keys) you will not see much increased
1934 memory on your process. DBM::Deep relies solely on the filesystem for storing
1935 and fetching data. Here is output from I</usr/bin/top> before even opening a
1938 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1939 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl
1941 Basically the process is taking 2,716K of memory. And here is the same
1942 process after storing and fetching 1,000,000 keys:
1944 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
1945 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl
1947 Notice the memory usage increased by only 56K. Test was performed on a 700mHz
1948 x86 box running Linux RedHat 7.2 & Perl 5.6.1.
1950 =head1 DB FILE FORMAT
1952 In case you were interested in the underlying DB file format, it is documented
1953 here in this section. You don't need to know this to use the module, it's just
1954 included for reference.
1958 DBM::Deep files always start with a 32-bit signature to identify the file type.
1959 This is at offset 0. The signature is "DPDB" in network byte order. This is
1960 checked for when the file is opened and an error will be thrown if it's not found.
1964 The DBM::Deep file is in a I<tagged format>, meaning each section of the file
1965 has a standard header containing the type of data, the length of data, and then
1966 the data itself. The type is a single character (1 byte), the length is a
1967 32-bit unsigned long in network byte order, and the data is, well, the data.
1968 Here is how it unfolds:
1972 Immediately after the 32-bit file signature is the I<Master Index> record.
1973 This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048
1974 bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array,
1975 depending on how the DBM::Deep object was constructed.
1977 The index works by looking at a I<MD5 Hash> of the hash key (or array index
1978 number). The first 8-bit char of the MD5 signature is the offset into the
1979 index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the
1980 index element is a file offset of the next tag for the key/element in question,
1981 which is usually a I<Bucket List> tag (see below).
1983 The next tag I<could> be another index, depending on how many keys/elements
1984 exist. See L<RE-INDEXING> below for details.
1988 A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus
1989 file offsets to where the actual data is stored. It starts with a standard
1990 tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or
1991 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus
1992 the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data.
1993 When the list fills up, a I<Re-Index> operation is performed (See
1994 L<RE-INDEXING> below).
1998 A I<Bucket> is a tag containing a key/value pair (in hash mode), or a
1999 index/value pair (in array mode). It starts with a standard tag header with
2000 type I<D> for scalar data (string, binary, etc.), or it could be a nested
2001 hash (type I<H>) or array (type I<A>). The value comes just after the tag
2002 header. The size reported in the tag header is only for the value, but then,
2003 just after the value is another size (32-bit unsigned long) and then the plain
2004 key itself. Since the value is likely to be fetched more often than the plain
2005 key, I figured it would be I<slightly> faster to store the value first.
2007 If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index>
2008 record for the nested structure, where the process begins all over again.
2012 After a I<Bucket List> grows to 16 records, its allocated space in the file is
2013 exhausted. Then, when another key/element comes in, the list is converted to a
2014 new index record. However, this index will look at the next char in the MD5
2015 hash, and arrange new Bucket List pointers accordingly. This process is called
2016 I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all
2017 17 (16 + new one) keys/elements are removed from the old Bucket List and
2018 inserted into the new index. Several new Bucket Lists are created in the
2019 process, as a new MD5 char from the key is being examined (it is unlikely that
2020 the keys will all share the same next char of their MD5s).
2022 Because of the way the I<MD5> algorithm works, it is impossible to tell exactly
2023 when the Bucket Lists will turn into indexes, but the first round tends to
2024 happen right around 4,000 keys. You will see a I<slight> decrease in
2025 performance here, but it picks back up pretty quick (see L<SPEED> above). Then
2026 it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's
2027 right around 900,000 keys. This process can continue nearly indefinitely --
2028 right up until the point the I<MD5> signatures start colliding with each other,
2029 and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND
2030 getting struck by lightning while you are walking to cash in your tickets.
2031 Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to
2032 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe
2033 this is 340 unodecillion, but don't quote me).
2037 When a new key/element is stored, the key (or index number) is first run through
2038 I<Digest::MD5> to get a 128-bit signature (example, in hex:
2039 b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked
2040 for the first char of the signature (in this case I<b0>). If it does not exist,
2041 a new I<Bucket List> is created for our key (and the next 15 future keys that
2042 happen to also have I<b> as their first MD5 char). The entire MD5 is written
2043 to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at
2044 this point, unless we are replacing an existing I<Bucket>), where the actual
2045 data will be stored.
2049 Fetching an existing key/element involves getting a I<Digest::MD5> of the key
2050 (or index number), then walking along the indexes. If there are enough
2051 keys/elements in this DB level, there might be nested indexes, each linked to
2052 a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which
2053 contains up to 16 full MD5 hashes. Each is checked for equality to the key in
2054 question. If we found a match, the I<Bucket> tag is loaded, where the value and
2055 plain key are stored.
2057 Fetching the plain key occurs when calling the I<first_key()> and I<next_key()>
2058 methods. In this process the indexes are walked systematically, and each key
2059 fetched in increasing MD5 order (which is why it appears random). Once the
2060 I<Bucket> is found, the value is skipped and the plain key returned instead.
2061 B<Note:> Do not count on keys being fetched as if the MD5 hashes were
2062 alphabetically sorted. This only happens on an index-level -- as soon as the
2063 I<Bucket Lists> are hit, the keys will come out in the order they went in --
2064 so it's pretty much undefined how the keys will come out -- just like Perl's
2067 =head1 CODE COVERAGE
2069 We use B<Devel::Cover> to test the code coverage of our tests, below is the
2070 B<Devel::Cover> report on this module's test suite.
2072 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2073 File stmt bran cond sub pod time total
2074 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2075 blib/lib/DBM/Deep.pm 95.2 83.8 70.0 98.2 100.0 58.0 91.0
2076 blib/lib/DBM/Deep/Array.pm 100.0 91.1 100.0 100.0 n/a 26.7 98.0
2077 blib/lib/DBM/Deep/Hash.pm 95.3 80.0 100.0 100.0 n/a 15.3 92.4
2078 Total 96.2 84.8 74.4 98.8 100.0 100.0 92.4
2079 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2081 =head1 MORE INFORMATION
2083 Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
2084 or send email to L<DBM-Deep@googlegroups.com>.
2088 Joseph Huckaby, L<jhuckaby@cpan.org>
2090 Rob Kinyon, L<rkinyon@cpan.org>
2092 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
2096 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
2097 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
2101 Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved.
2102 This is free software, you may use it and distribute it under the
2103 same terms as Perl itself.