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 " . $db->{my_complex}->[1]->{perl} . "!\n";
27 # (c) 2002-2005 Joseph Huckaby. All Rights Reserved.
28 # This program is free software; you can redistribute it and/or
29 # modify it under the same terms as Perl itself.
38 use vars qw/$VERSION/;
43 # Set to 4 and 'N' for 32-bit offset tags (default). Theoretical limit of 4 GB per file.
44 # (Perl must be compiled with largefile support for files > 2 GB)
46 # Set to 8 and 'Q' for 64-bit offsets. Theoretical limit of 16 XB per file.
47 # (Perl must be compiled with largefile and 64-bit long support)
53 # Set to 4 and 'N' for 32-bit data length prefixes. Limit of 4 GB for each key/value.
54 # Upgrading this is possible (see above) but probably not necessary. If you need
55 # more than 4 GB for a single key or value, this module is really not for you :-)
57 #my $DATA_LENGTH_SIZE = 4;
58 #my $DATA_LENGTH_PACK = 'N';
59 my ($LONG_SIZE, $LONG_PACK, $DATA_LENGTH_SIZE, $DATA_LENGTH_PACK);
62 # Maximum number of buckets per list before another level of indexing is done.
63 # Increase this value for slightly greater speed, but larger database files.
64 # DO NOT decrease this value below 16, due to risk of recursive reindex overrun.
69 # Better not adjust anything below here, unless you're me :-)
73 # Setup digest function for keys
75 my ($DIGEST_FUNC, $HASH_SIZE);
76 #my $DIGEST_FUNC = \&Digest::MD5::md5;
79 # Precalculate index and bucket sizes based on values above.
82 my ($INDEX_SIZE, $BUCKET_SIZE, $BUCKET_LIST_SIZE);
89 # Setup file and tag signatures. These should never change.
91 sub SIG_FILE () { 'DPDB' }
92 sub SIG_HASH () { 'H' }
93 sub SIG_ARRAY () { 'A' }
94 sub SIG_NULL () { 'N' }
95 sub SIG_DATA () { 'D' }
96 sub SIG_INDEX () { 'I' }
97 sub SIG_BLIST () { 'B' }
101 # Setup constants for users to pass to new()
103 sub TYPE_HASH () { return SIG_HASH; }
104 sub TYPE_ARRAY () { return SIG_ARRAY; }
108 # Class constructor method for Perl OO interface.
109 # Calls tie() and returns blessed reference to tied hash or array,
110 # providing a hybrid OO/tie interface.
114 if (scalar(@_) > 1) { $args = {@_}; }
115 else { $args = { file => shift }; }
118 # Check if we want a tied hash or array.
121 if (defined($args->{type}) && $args->{type} eq TYPE_ARRAY) {
122 tie @$self, $class, %$args;
125 tie %$self, $class, %$args;
128 return bless $self, $class;
132 my @outer_params = qw( type base_offset );
135 # Setup $self and bless into this class.
142 base_offset => length(SIG_FILE),
153 filter_store_key => undef,
154 filter_store_value => undef,
155 filter_fetch_key => undef,
156 filter_fetch_value => undef,
165 foreach my $outer_parm ( @outer_params ) {
166 next unless exists $args->{$outer_parm};
167 $self->{$outer_parm} = $args->{$outer_parm}
170 if ( exists $args->{root} ) {
171 $self->{root} = $args->{root};
174 # This is cleanup based on the fact that the $args
175 # coming in is for both the root and non-root items
176 delete $self->root->{$_} for @outer_params;
178 $self->root->{links}++;
180 if (!defined($self->fh)) { $self->_open(); }
186 sub _get_self { tied( %{$_[0]} ) || $_[0] }
190 # Tied hash constructor method, called by Perl's tie() function.
194 if (scalar(@_) > 1) { $args = {@_}; }
195 #XXX This use of ref() is bad and is a bug
196 elsif (ref($_[0])) { $args = $_[0]; }
197 else { $args = { file => shift }; }
199 $args->{type} = TYPE_HASH;
201 return $class->_init($args);
206 # Tied array constructor method, called by Perl's tie() function.
210 if (scalar(@_) > 1) { $args = {@_}; }
211 #XXX This use of ref() is bad and is a bug
212 elsif (ref($_[0])) { $args = $_[0]; }
213 else { $args = { file => shift }; }
215 $args->{type} = TYPE_ARRAY;
217 return $class->_init($args);
222 # Class deconstructor. Close file handle if there are no more refs.
224 my $self = _get_self($_[0]);
227 $self->root->{links}--;
229 if (!$self->root->{links}) {
236 # Open a FileHandle to the database, create if nonexistent.
237 # Make sure file signature matches DeepDB spec.
239 my $self = _get_self($_[0]);
241 if (defined($self->fh)) { $self->_close(); }
244 if (!(-e $self->root->{file}) && $self->root->{mode} eq 'r+') {
245 my $temp = FileHandle->new( $self->root->{file}, 'w' );
248 #XXX Convert to set_fh()
249 $self->root->{fh} = FileHandle->new( $self->root->{file}, $self->root->{mode} );
250 # }; if ($@ ) { $self->_throw_error( "Received error: $@\n" ); }
251 if (! defined($self->fh)) {
252 return $self->_throw_error("Cannot open file: " . $self->root->{file} . ": $!");
255 binmode $self->fh; # for win32
256 if ($self->root->{autoflush}) {
257 $self->fh->autoflush();
261 seek($self->fh, 0, 0);
262 my $bytes_read = read( $self->fh, $signature, length(SIG_FILE));
265 # File is empty -- write signature and master index
268 seek($self->fh, 0, 0);
269 $self->fh->print(SIG_FILE);
270 $self->root->{end} = length(SIG_FILE);
271 $self->_create_tag($self->base_offset, $self->type, chr(0) x $INDEX_SIZE);
273 my $plain_key = "[base]";
274 $self->fh->print( pack($DATA_LENGTH_PACK, length($plain_key)) . $plain_key );
275 $self->root->{end} += $DATA_LENGTH_SIZE + length($plain_key);
282 # Check signature was valid
284 unless ($signature eq SIG_FILE) {
286 return $self->_throw_error("Signature not found -- file is not a Deep DB");
289 $self->root->{end} = (stat($self->fh))[7];
292 # Get our type from master index signature
294 my $tag = $self->_load_tag($self->base_offset);
295 #XXX We probably also want to store the hash algorithm name and not assume anything
297 return $self->_throw_error("Corrupted file, no master index record");
299 if ($self->{type} ne $tag->{signature}) {
300 return $self->_throw_error("File type mismatch");
308 # Close database FileHandle
310 my $self = _get_self($_[0]);
311 undef $self->root->{fh};
316 # Given offset, signature and content, create tag and write to disk
318 my ($self, $offset, $sig, $content) = @_;
319 my $size = length($content);
321 seek($self->fh, $offset, 0);
322 $self->fh->print( $sig . pack($DATA_LENGTH_PACK, $size) . $content );
324 if ($offset == $self->root->{end}) {
325 $self->root->{end} += SIG_SIZE + $DATA_LENGTH_SIZE + $size;
331 offset => $offset + SIG_SIZE + $DATA_LENGTH_SIZE,
338 # Given offset, load single tag and return signature, size and data
343 seek($self->fh, $offset, 0);
344 if ($self->fh->eof()) { return undef; }
347 read( $self->fh, $sig, SIG_SIZE);
350 read( $self->fh, $size, $DATA_LENGTH_SIZE);
351 $size = unpack($DATA_LENGTH_PACK, $size);
354 read( $self->fh, $buffer, $size);
359 offset => $offset + SIG_SIZE + $DATA_LENGTH_SIZE,
366 # Given index tag, lookup single entry in index and return .
369 my ($tag, $index) = @_;
371 my $location = unpack($LONG_PACK, substr($tag->{content}, $index * $LONG_SIZE, $LONG_SIZE) );
372 if (!$location) { return; }
374 return $self->_load_tag( $location );
379 # Adds one key/value pair to bucket list, given offset, MD5 digest of key,
380 # plain (undigested) key and value.
383 my ($tag, $md5, $plain_key, $value) = @_;
384 my $keys = $tag->{content};
388 my $is_dbm_deep = eval { $value->isa( 'DBM::Deep' ) };
389 my $internal_ref = $is_dbm_deep && ($value->root eq $self->root);
392 # Iterate through buckets, seeing if this is a new entry or a replace.
394 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
395 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
396 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
399 # Found empty bucket (end of list). Populate and exit loop.
403 $location = $internal_ref
404 ? $value->base_offset
405 : $self->root->{end};
407 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
408 $self->fh->print( $md5 . pack($LONG_PACK, $location) );
411 elsif ($md5 eq $key) {
413 # Found existing bucket with same key. Replace with new value.
418 $location = $value->base_offset;
419 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
420 $self->fh->print( $md5 . pack($LONG_PACK, $location) );
423 seek($self->fh, $subloc + SIG_SIZE, 0);
425 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
428 # If value is a hash, array, or raw value with equal or less size, we can
429 # reuse the same content area of the database. Otherwise, we have to create
430 # a new content area at the EOF.
433 my $r = Scalar::Util::reftype( $value ) || '';
434 if ( $r eq 'HASH' || $r eq 'ARRAY' ) { $actual_length = $INDEX_SIZE; }
435 else { $actual_length = length($value); }
437 if ($actual_length <= $size) {
441 $location = $self->root->{end};
442 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE) + $HASH_SIZE, 0);
443 $self->fh->print( pack($LONG_PACK, $location) );
451 # If this is an internal reference, return now.
452 # No need to write value or plain key
459 # If bucket didn't fit into list, split into a new index level
462 seek($self->fh, $tag->{ref_loc}, 0);
463 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
465 my $index_tag = $self->_create_tag($self->root->{end}, SIG_INDEX, chr(0) x $INDEX_SIZE);
468 $keys .= $md5 . pack($LONG_PACK, 0);
470 for (my $i=0; $i<=$MAX_BUCKETS; $i++) {
471 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
473 my $old_subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
474 my $num = ord(substr($key, $tag->{ch} + 1, 1));
476 if ($offsets[$num]) {
477 my $offset = $offsets[$num] + SIG_SIZE + $DATA_LENGTH_SIZE;
478 seek($self->fh, $offset, 0);
480 read( $self->fh, $subkeys, $BUCKET_LIST_SIZE);
482 for (my $k=0; $k<$MAX_BUCKETS; $k++) {
483 my $subloc = unpack($LONG_PACK, substr($subkeys, ($k * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
485 seek($self->fh, $offset + ($k * $BUCKET_SIZE), 0);
486 $self->fh->print( $key . pack($LONG_PACK, $old_subloc || $self->root->{end}) );
492 $offsets[$num] = $self->root->{end};
493 seek($self->fh, $index_tag->{offset} + ($num * $LONG_SIZE), 0);
494 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
496 my $blist_tag = $self->_create_tag($self->root->{end}, SIG_BLIST, chr(0) x $BUCKET_LIST_SIZE);
498 seek($self->fh, $blist_tag->{offset}, 0);
499 $self->fh->print( $key . pack($LONG_PACK, $old_subloc || $self->root->{end}) );
504 $location ||= $self->root->{end};
505 } # re-index bucket list
508 # Seek to content area and store signature, value and plaintext key
512 seek($self->fh, $location, 0);
515 # Write signature based on content type, set content length and write actual value.
517 my $r = Scalar::Util::reftype($value) || '';
519 $self->fh->print( TYPE_HASH );
520 $self->fh->print( pack($DATA_LENGTH_PACK, $INDEX_SIZE) . chr(0) x $INDEX_SIZE );
521 $content_length = $INDEX_SIZE;
523 elsif ($r eq 'ARRAY') {
524 $self->fh->print( TYPE_ARRAY );
525 $self->fh->print( pack($DATA_LENGTH_PACK, $INDEX_SIZE) . chr(0) x $INDEX_SIZE );
526 $content_length = $INDEX_SIZE;
528 elsif (!defined($value)) {
529 $self->fh->print( SIG_NULL );
530 $self->fh->print( pack($DATA_LENGTH_PACK, 0) );
534 $self->fh->print( SIG_DATA );
535 $self->fh->print( pack($DATA_LENGTH_PACK, length($value)) . $value );
536 $content_length = length($value);
540 # Plain key is stored AFTER value, as keys are typically fetched less often.
542 $self->fh->print( pack($DATA_LENGTH_PACK, length($plain_key)) . $plain_key );
545 # If value is blessed, preserve class name
547 if ( $self->root->{autobless} ) {
548 my $value_class = Scalar::Util::blessed($value);
549 if ( defined $value_class && $value_class ne 'DBM::Deep' ) {
551 # Blessed ref -- will restore later
553 $self->fh->print( chr(1) );
554 $self->fh->print( pack($DATA_LENGTH_PACK, length($value_class)) . $value_class );
555 $content_length += 1;
556 $content_length += $DATA_LENGTH_SIZE + length($value_class);
559 $self->fh->print( chr(0) );
560 $content_length += 1;
565 # If this is a new content area, advance EOF counter
567 if ($location == $self->root->{end}) {
568 $self->root->{end} += SIG_SIZE;
569 $self->root->{end} += $DATA_LENGTH_SIZE + $content_length;
570 $self->root->{end} += $DATA_LENGTH_SIZE + length($plain_key);
574 # If content is a hash or array, create new child DeepDB object and
575 # pass each key or element to it.
578 my $branch = DBM::Deep->new(
580 base_offset => $location,
583 foreach my $key (keys %{$value}) {
584 $branch->{$key} = $value->{$key};
587 elsif ($r eq 'ARRAY') {
588 my $branch = DBM::Deep->new(
590 base_offset => $location,
594 foreach my $element (@{$value}) {
595 $branch->[$index] = $element;
603 return $self->_throw_error("Fatal error: indexing failed -- possibly due to corruption in file");
606 sub _get_bucket_value {
608 # Fetch single value given tag and MD5 digested key.
611 my ($tag, $md5) = @_;
612 my $keys = $tag->{content};
615 # Iterate through buckets, looking for a key match
618 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
619 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
620 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
624 # Hit end of list, no match
629 if ( $md5 ne $key ) {
634 # Found match -- seek to offset and read signature
637 seek($self->fh, $subloc, 0);
638 read( $self->fh, $signature, SIG_SIZE);
641 # If value is a hash or array, return new DeepDB object with correct offset
643 if (($signature eq TYPE_HASH) || ($signature eq TYPE_ARRAY)) {
644 my $obj = DBM::Deep->new(
646 base_offset => $subloc,
650 if ($self->root->{autobless}) {
652 # Skip over value and plain key to see if object needs
655 seek($self->fh, $DATA_LENGTH_SIZE + $INDEX_SIZE, 1);
658 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
659 if ($size) { seek($self->fh, $size, 1); }
662 read( $self->fh, $bless_bit, 1);
663 if (ord($bless_bit)) {
665 # Yes, object needs to be re-blessed
668 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
669 if ($size) { read( $self->fh, $class_name, $size); }
670 if ($class_name) { $obj = bless( $obj, $class_name ); }
678 # Otherwise return actual value
680 elsif ($signature eq SIG_DATA) {
683 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
684 if ($size) { read( $self->fh, $value, $size); }
689 # Key exists, but content is null
699 # Delete single key/value pair given tag and MD5 digested key.
702 my ($tag, $md5) = @_;
703 my $keys = $tag->{content};
706 # Iterate through buckets, looking for a key match
709 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
710 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
711 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
715 # Hit end of list, no match
720 if ( $md5 ne $key ) {
725 # Matched key -- delete bucket and return
727 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
728 $self->fh->print( substr($keys, ($i+1) * $BUCKET_SIZE ) );
729 $self->fh->print( chr(0) x $BUCKET_SIZE );
739 # Check existence of single key given tag and MD5 digested key.
742 my ($tag, $md5) = @_;
743 my $keys = $tag->{content};
746 # Iterate through buckets, looking for a key match
749 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
750 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
751 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
755 # Hit end of list, no match
760 if ( $md5 ne $key ) {
765 # Matched key -- return true
773 sub _find_bucket_list {
775 # Locate offset for bucket list, given digested key
781 # Locate offset for bucket list using digest index system
784 my $tag = $self->_load_tag($self->base_offset);
785 if (!$tag) { return; }
787 while ($tag->{signature} ne SIG_BLIST) {
788 $tag = $self->_index_lookup($tag, ord(substr($md5, $ch, 1)));
789 if (!$tag) { return; }
796 sub _traverse_index {
798 # Scan index and recursively step into deeper levels, looking for next key.
800 my ($self, $offset, $ch, $force_return_next) = @_;
801 $force_return_next = undef unless $force_return_next;
803 my $tag = $self->_load_tag( $offset );
805 if ($tag->{signature} ne SIG_BLIST) {
806 my $content = $tag->{content};
808 if ($self->{return_next}) { $start = 0; }
809 else { $start = ord(substr($self->{prev_md5}, $ch, 1)); }
811 for (my $index = $start; $index < 256; $index++) {
812 my $subloc = unpack($LONG_PACK, substr($content, $index * $LONG_SIZE, $LONG_SIZE) );
814 my $result = $self->_traverse_index( $subloc, $ch + 1, $force_return_next );
815 if (defined($result)) { return $result; }
819 $self->{return_next} = 1;
822 elsif ($tag->{signature} eq SIG_BLIST) {
823 my $keys = $tag->{content};
824 if ($force_return_next) { $self->{return_next} = 1; }
827 # Iterate through buckets, looking for a key match
829 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
830 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
831 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
835 # End of bucket list -- return to outer loop
837 $self->{return_next} = 1;
840 elsif ($key eq $self->{prev_md5}) {
842 # Located previous key -- return next one found
844 $self->{return_next} = 1;
847 elsif ($self->{return_next}) {
849 # Seek to bucket location and skip over signature
851 seek($self->fh, $subloc + SIG_SIZE, 0);
854 # Skip over value to get to plain key
857 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
858 if ($size) { seek($self->fh, $size, 1); }
861 # Read in plain key and return as scalar
864 read( $self->fh, $size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
865 if ($size) { read( $self->fh, $plain_key, $size); }
871 $self->{return_next} = 1;
872 } # tag is a bucket list
879 # Locate next key, given digested previous one
881 my $self = _get_self($_[0]);
883 $self->{prev_md5} = $_[1] ? $_[1] : undef;
884 $self->{return_next} = 0;
887 # If the previous key was not specifed, start at the top and
888 # return the first one found.
890 if (!$self->{prev_md5}) {
891 $self->{prev_md5} = chr(0) x $HASH_SIZE;
892 $self->{return_next} = 1;
895 return $self->_traverse_index( $self->base_offset, 0 );
900 # If db locking is set, flock() the db file. If called multiple
901 # times before unlock(), then the same number of unlocks() must
902 # be called before the lock is released.
904 my $self = _get_self($_[0]);
906 $type = LOCK_EX unless defined $type;
908 if ($self->root->{locking}) {
909 if (!$self->root->{locked}) { flock($self->fh, $type); }
910 $self->root->{locked}++;
916 # If db locking is set, unlock the db file. See note in lock()
917 # regarding calling lock() multiple times.
919 my $self = _get_self($_[0]);
921 if ($self->root->{locking} && $self->root->{locked} > 0) {
922 $self->root->{locked}--;
923 if (!$self->root->{locked}) { flock($self->fh, LOCK_UN); }
927 #XXX These uses of ref() need verified
930 # Copy single level of keys or elements to new DB handle.
931 # Recurse for nested structures
933 my $self = _get_self($_[0]);
936 if ($self->type eq TYPE_HASH) {
937 my $key = $self->first_key();
939 my $value = $self->get($key);
940 #XXX This doesn't work with autobless
941 if (!ref($value)) { $db_temp->{$key} = $value; }
943 my $type = $value->type;
944 if ($type eq TYPE_HASH) { $db_temp->{$key} = {}; }
945 else { $db_temp->{$key} = []; }
946 $value->_copy_node( $db_temp->{$key} );
948 $key = $self->next_key($key);
952 my $length = $self->length();
953 for (my $index = 0; $index < $length; $index++) {
954 my $value = $self->get($index);
955 if (!ref($value)) { $db_temp->[$index] = $value; }
956 #XXX NO tests for this code
958 my $type = $value->type;
959 if ($type eq TYPE_HASH) { $db_temp->[$index] = {}; }
960 else { $db_temp->[$index] = []; }
961 $value->_copy_node( $db_temp->[$index] );
969 # Recursively export into standard Perl hashes and arrays.
971 my $self = _get_self($_[0]);
974 if ($self->type eq TYPE_HASH) { $temp = {}; }
975 elsif ($self->type eq TYPE_ARRAY) { $temp = []; }
978 $self->_copy_node( $temp );
986 # Recursively import Perl hash/array structure
988 #XXX This use of ref() seems to be ok
989 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
991 my $self = _get_self($_[0]);
994 #XXX This use of ref() seems to be ok
997 # struct is not a reference, so just import based on our type
1001 if ($self->type eq TYPE_HASH) { $struct = {@_}; }
1002 elsif ($self->type eq TYPE_ARRAY) { $struct = [@_]; }
1005 my $r = Scalar::Util::reftype($struct) || '';
1006 if ($r eq "HASH" && $self->type eq TYPE_HASH) {
1007 foreach my $key (keys %$struct) { $self->put($key, $struct->{$key}); }
1009 elsif ($r eq "ARRAY" && $self->type eq TYPE_ARRAY) {
1010 $self->push( @$struct );
1013 return $self->_throw_error("Cannot import: type mismatch");
1021 # Rebuild entire database into new file, then move
1022 # it back on top of original.
1024 my $self = _get_self($_[0]);
1025 if ($self->root->{links} > 1) {
1026 return $self->_throw_error("Cannot optimize: reference count is greater than 1");
1029 my $db_temp = DBM::Deep->new(
1030 file => $self->root->{file} . '.tmp',
1034 return $self->_throw_error("Cannot optimize: failed to open temp file: $!");
1038 $self->_copy_node( $db_temp );
1042 # Attempt to copy user, group and permissions over to new file
1044 my @stats = stat($self->fh);
1045 my $perms = $stats[2] & 07777;
1046 my $uid = $stats[4];
1047 my $gid = $stats[5];
1048 chown( $uid, $gid, $self->root->{file} . '.tmp' );
1049 chmod( $perms, $self->root->{file} . '.tmp' );
1051 # q.v. perlport for more information on this variable
1052 if ( $^O eq 'MSWin32' ) {
1054 # Potential race condition when optmizing on Win32 with locking.
1055 # The Windows filesystem requires that the filehandle be closed
1056 # before it is overwritten with rename(). This could be redone
1063 if (!rename $self->root->{file} . '.tmp', $self->root->{file}) {
1064 unlink $self->root->{file} . '.tmp';
1066 return $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!");
1078 # Make copy of object and return
1080 my $self = _get_self($_[0]);
1082 return DBM::Deep->new(
1083 type => $self->type,
1084 base_offset => $self->base_offset,
1090 my %is_legal_filter = map {
1093 store_key store_value
1094 fetch_key fetch_value
1099 # Setup filter function for storing or fetching the key or value
1101 my $self = _get_self($_[0]);
1102 my $type = lc $_[1];
1103 my $func = $_[2] ? $_[2] : undef;
1105 if ( $is_legal_filter{$type} ) {
1106 $self->root->{"filter_$type"} = $func;
1120 # Get access to the root structure
1122 my $self = _get_self($_[0]);
1123 return $self->{root};
1128 # Get access to the raw FileHandle
1130 #XXX It will be useful, though, when we split out HASH and ARRAY
1131 my $self = _get_self($_[0]);
1132 return $self->root->{fh};
1137 # Get type of current node (TYPE_HASH or TYPE_ARRAY)
1139 my $self = _get_self($_[0]);
1140 return $self->{type};
1145 # Get base_offset of current node (TYPE_HASH or TYPE_ARRAY)
1147 my $self = _get_self($_[0]);
1148 return $self->{base_offset};
1153 # Get last error string, or undef if no error
1156 ? ( _get_self($_[0])->{root}->{error} or undef )
1166 # Store error string in self
1168 my $self = _get_self($_[0]);
1169 my $error_text = $_[1];
1171 $self->root->{error} = $error_text;
1173 unless ($self->root->{debug}) {
1174 die "DBM::Deep: $error_text\n";
1177 warn "DBM::Deep: $error_text\n";
1185 my $self = _get_self($_[0]);
1187 undef $self->root->{error};
1192 # Precalculate index, bucket and bucket list sizes
1195 #XXX I don't like this ...
1196 set_pack() unless defined $LONG_SIZE;
1198 $INDEX_SIZE = 256 * $LONG_SIZE;
1199 $BUCKET_SIZE = $HASH_SIZE + $LONG_SIZE;
1200 $BUCKET_LIST_SIZE = $MAX_BUCKETS * $BUCKET_SIZE;
1205 # Set pack/unpack modes (see file header for more)
1207 my ($long_s, $long_p, $data_s, $data_p) = @_;
1209 $LONG_SIZE = $long_s ? $long_s : 4;
1210 $LONG_PACK = $long_p ? $long_p : 'N';
1212 $DATA_LENGTH_SIZE = $data_s ? $data_s : 4;
1213 $DATA_LENGTH_PACK = $data_p ? $data_p : 'N';
1220 # Set key digest function (default is MD5)
1222 my ($digest_func, $hash_size) = @_;
1224 $DIGEST_FUNC = $digest_func ? $digest_func : \&Digest::MD5::md5;
1225 $HASH_SIZE = $hash_size ? $hash_size : 16;
1231 # tie() methods (hashes and arrays)
1236 # Store single hash key/value or array element in database.
1238 my $self = _get_self($_[0]);
1239 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1240 #XXX What is ref() checking here?
1241 #YYY User may be storing a hash, in which case we do not want it run
1242 #YYY through the filtering system
1243 my $value = ($self->root->{filter_store_value} && !ref($_[2])) ? $self->root->{filter_store_value}->($_[2]) : $_[2];
1245 my $unpacked_key = $key;
1246 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1247 my $md5 = $DIGEST_FUNC->($key);
1250 # Make sure file is open
1252 if (!defined($self->fh) && !$self->_open()) {
1257 # Request exclusive lock for writing
1259 $self->lock( LOCK_EX );
1262 # If locking is enabled, set 'end' parameter again, in case another
1263 # DB instance appended to our file while we were unlocked.
1265 if ($self->root->{locking} || $self->root->{volatile}) {
1266 $self->root->{end} = (stat($self->fh))[7];
1270 # Locate offset for bucket list using digest index system
1272 my $tag = $self->_load_tag($self->base_offset);
1274 $tag = $self->_create_tag($self->base_offset, SIG_INDEX, chr(0) x $INDEX_SIZE);
1278 while ($tag->{signature} ne SIG_BLIST) {
1279 my $num = ord(substr($md5, $ch, 1));
1280 my $new_tag = $self->_index_lookup($tag, $num);
1282 my $ref_loc = $tag->{offset} + ($num * $LONG_SIZE);
1283 seek($self->fh, $ref_loc, 0);
1284 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
1286 $tag = $self->_create_tag($self->root->{end}, SIG_BLIST, chr(0) x $BUCKET_LIST_SIZE);
1287 $tag->{ref_loc} = $ref_loc;
1292 my $ref_loc = $tag->{offset} + ($num * $LONG_SIZE);
1294 $tag->{ref_loc} = $ref_loc;
1301 # Add key/value to bucket list
1303 my $result = $self->_add_bucket( $tag, $md5, $key, $value );
1306 # If this object is an array, and bucket was not a replace, and key is numerical,
1307 # and index is equal or greater than current length, advance length variable.
1309 if (($result == 2) && ($self->type eq TYPE_ARRAY) && ($unpacked_key =~ /^\d+$/) && ($unpacked_key >= $self->FETCHSIZE())) {
1310 $self->STORESIZE( $unpacked_key + 1 );
1320 # Fetch single value or element given plain key or array index
1322 my $self = _get_self($_[0]);
1325 if ( $self->type eq TYPE_HASH ) {
1326 if ( my $filter = $self->root->{filter_store_key} ) {
1327 $key = $filter->( $key );
1330 elsif ( $self->type eq TYPE_ARRAY ) {
1331 if ( $key =~ /^\d+$/ ) {
1332 $key = pack($LONG_PACK, $key);
1336 my $md5 = $DIGEST_FUNC->($key);
1339 # Make sure file is open
1341 if (!defined($self->fh)) { $self->_open(); }
1344 # Request shared lock for reading
1346 $self->lock( LOCK_SH );
1348 my $tag = $self->_find_bucket_list( $md5 );
1355 # Get value from bucket list
1357 my $result = $self->_get_bucket_value( $tag, $md5 );
1361 #XXX What is ref() checking here?
1362 return ($result && !ref($result) && $self->root->{filter_fetch_value}) ? $self->root->{filter_fetch_value}->($result) : $result;
1367 # Delete single key/value pair or element given plain key or array index
1369 my $self = _get_self($_[0]);
1370 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1372 my $unpacked_key = $key;
1373 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1374 my $md5 = $DIGEST_FUNC->($key);
1377 # Make sure file is open
1379 if (!defined($self->fh)) { $self->_open(); }
1382 # Request exclusive lock for writing
1384 $self->lock( LOCK_EX );
1386 my $tag = $self->_find_bucket_list( $md5 );
1395 my $result = $self->_delete_bucket( $tag, $md5 );
1398 # If this object is an array and the key deleted was on the end of the stack,
1399 # decrement the length variable.
1401 if ($result && ($self->type eq TYPE_ARRAY) && ($unpacked_key == $self->FETCHSIZE() - 1)) {
1402 $self->STORESIZE( $unpacked_key );
1412 # Check if a single key or element exists given plain key or array index
1414 my $self = _get_self($_[0]);
1415 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1417 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1418 my $md5 = $DIGEST_FUNC->($key);
1421 # Make sure file is open
1423 if (!defined($self->fh)) { $self->_open(); }
1426 # Request shared lock for reading
1428 $self->lock( LOCK_SH );
1430 my $tag = $self->_find_bucket_list( $md5 );
1433 # For some reason, the built-in exists() function returns '' for false
1441 # Check if bucket exists and return 1 or ''
1443 my $result = $self->_bucket_exists( $tag, $md5 ) || '';
1452 # Clear all keys from hash, or all elements from array.
1454 my $self = _get_self($_[0]);
1457 # Make sure file is open
1459 if (!defined($self->fh)) { $self->_open(); }
1462 # Request exclusive lock for writing
1464 $self->lock( LOCK_EX );
1466 seek($self->fh, $self->base_offset, 0);
1467 if ($self->fh->eof()) {
1472 $self->_create_tag($self->base_offset, $self->type, chr(0) x $INDEX_SIZE);
1481 # Locate and return first key (in no particular order)
1483 my $self = _get_self($_[0]);
1484 if ($self->type ne TYPE_HASH) {
1485 return $self->_throw_error("FIRSTKEY method only supported for hashes");
1489 # Make sure file is open
1491 if (!defined($self->fh)) { $self->_open(); }
1494 # Request shared lock for reading
1496 $self->lock( LOCK_SH );
1498 my $result = $self->_get_next_key();
1502 return ($result && $self->root->{filter_fetch_key}) ? $self->root->{filter_fetch_key}->($result) : $result;
1507 # Return next key (in no particular order), given previous one
1509 my $self = _get_self($_[0]);
1510 if ($self->type ne TYPE_HASH) {
1511 return $self->_throw_error("NEXTKEY method only supported for hashes");
1513 my $prev_key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1514 my $prev_md5 = $DIGEST_FUNC->($prev_key);
1517 # Make sure file is open
1519 if (!defined($self->fh)) { $self->_open(); }
1522 # Request shared lock for reading
1524 $self->lock( LOCK_SH );
1526 my $result = $self->_get_next_key( $prev_md5 );
1530 return ($result && $self->root->{filter_fetch_key}) ? $self->root->{filter_fetch_key}->($result) : $result;
1534 # The following methods are for arrays only
1539 # Return the length of the array
1541 my $self = _get_self($_[0]);
1542 if ($self->type ne TYPE_ARRAY) {
1543 return $self->_throw_error("FETCHSIZE method only supported for arrays");
1546 my $SAVE_FILTER = $self->root->{filter_fetch_value};
1547 $self->root->{filter_fetch_value} = undef;
1549 my $packed_size = $self->FETCH('length');
1551 $self->root->{filter_fetch_value} = $SAVE_FILTER;
1553 if ($packed_size) { return int(unpack($LONG_PACK, $packed_size)); }
1559 # Set the length of the array
1561 my $self = _get_self($_[0]);
1562 if ($self->type ne TYPE_ARRAY) {
1563 return $self->_throw_error("STORESIZE method only supported for arrays");
1565 my $new_length = $_[1];
1567 my $SAVE_FILTER = $self->root->{filter_store_value};
1568 $self->root->{filter_store_value} = undef;
1570 my $result = $self->STORE('length', pack($LONG_PACK, $new_length));
1572 $self->root->{filter_store_value} = $SAVE_FILTER;
1579 # Remove and return the last element on the array
1581 my $self = _get_self($_[0]);
1582 if ($self->type ne TYPE_ARRAY) {
1583 return $self->_throw_error("POP method only supported for arrays");
1585 my $length = $self->FETCHSIZE();
1588 my $content = $self->FETCH( $length - 1 );
1589 $self->DELETE( $length - 1 );
1599 # Add new element(s) to the end of the array
1601 my $self = _get_self(shift);
1602 if ($self->type ne TYPE_ARRAY) {
1603 return $self->_throw_error("PUSH method only supported for arrays");
1605 my $length = $self->FETCHSIZE();
1607 while (my $content = shift @_) {
1608 $self->STORE( $length, $content );
1615 # Remove and return first element on the array.
1616 # Shift over remaining elements to take up space.
1618 my $self = _get_self($_[0]);
1619 if ($self->type ne TYPE_ARRAY) {
1620 return $self->_throw_error("SHIFT method only supported for arrays");
1622 my $length = $self->FETCHSIZE();
1625 my $content = $self->FETCH( 0 );
1628 # Shift elements over and remove last one.
1630 for (my $i = 0; $i < $length - 1; $i++) {
1631 $self->STORE( $i, $self->FETCH($i + 1) );
1633 $self->DELETE( $length - 1 );
1644 # Insert new element(s) at beginning of array.
1645 # Shift over other elements to make space.
1647 my $self = _get_self($_[0]);shift @_;
1648 if ($self->type ne TYPE_ARRAY) {
1649 return $self->_throw_error("UNSHIFT method only supported for arrays");
1651 my @new_elements = @_;
1652 my $length = $self->FETCHSIZE();
1653 my $new_size = scalar @new_elements;
1656 for (my $i = $length - 1; $i >= 0; $i--) {
1657 $self->STORE( $i + $new_size, $self->FETCH($i) );
1661 for (my $i = 0; $i < $new_size; $i++) {
1662 $self->STORE( $i, $new_elements[$i] );
1668 # Splices section of array with optional new section.
1669 # Returns deleted section, or last element deleted in scalar context.
1671 my $self = _get_self($_[0]);shift @_;
1672 if ($self->type ne TYPE_ARRAY) {
1673 return $self->_throw_error("SPLICE method only supported for arrays");
1675 my $length = $self->FETCHSIZE();
1678 # Calculate offset and length of splice
1680 my $offset = shift || 0;
1681 if ($offset < 0) { $offset += $length; }
1684 if (scalar @_) { $splice_length = shift; }
1685 else { $splice_length = $length - $offset; }
1686 if ($splice_length < 0) { $splice_length += ($length - $offset); }
1689 # Setup array with new elements, and copy out old elements for return
1691 my @new_elements = @_;
1692 my $new_size = scalar @new_elements;
1694 my @old_elements = ();
1695 for (my $i = $offset; $i < $offset + $splice_length; $i++) {
1696 push @old_elements, $self->FETCH( $i );
1700 # Adjust array length, and shift elements to accomodate new section.
1702 if ( $new_size != $splice_length ) {
1703 if ($new_size > $splice_length) {
1704 for (my $i = $length - 1; $i >= $offset + $splice_length; $i--) {
1705 $self->STORE( $i + ($new_size - $splice_length), $self->FETCH($i) );
1709 for (my $i = $offset + $splice_length; $i < $length; $i++) {
1710 $self->STORE( $i + ($new_size - $splice_length), $self->FETCH($i) );
1712 for (my $i = 0; $i < $splice_length - $new_size; $i++) {
1713 $self->DELETE( $length - 1 );
1720 # Insert new elements into array
1722 for (my $i = $offset; $i < $offset + $new_size; $i++) {
1723 $self->STORE( $i, shift @new_elements );
1727 # Return deleted section, or last element in scalar context.
1729 return wantarray ? @old_elements : $old_elements[-1];
1732 #XXX We don't need to define it.
1733 #XXX It will be useful, though, when we split out HASH and ARRAY
1736 # Perl will call EXTEND() when the array is likely to grow.
1737 # We don't care, but include it for compatibility.
1742 # Public method aliases
1744 *put = *store = *STORE;
1745 *get = *fetch = *FETCH;
1749 *first_key = *FIRSTKEY;
1750 *next_key = *NEXTKEY;
1751 *length = *FETCHSIZE;
1755 *unshift = *UNSHIFT;
1764 DBM::Deep - A pure perl multi-level hash/array DBM
1769 my $db = DBM::Deep->new( "foo.db" );
1771 $db->{key} = 'value'; # tie() style
1774 $db->put('key', 'value'); # OO style
1775 print $db->get('key');
1777 # true multi-level support
1778 $db->{my_complex} = [
1779 'hello', { perl => 'rules' },
1784 A unique flat-file database module, written in pure perl. True
1785 multi-level hash/array support (unlike MLDBM, which is faked), hybrid
1786 OO / tie() interface, cross-platform FTPable files, and quite fast. Can
1787 handle millions of keys and unlimited hash levels without significant
1788 slow-down. Written from the ground-up in pure perl -- this is NOT a
1789 wrapper around a C-based DBM. Out-of-the-box compatibility with Unix,
1790 Mac OS X and Windows.
1794 Hopefully you are using CPAN's excellent Perl module, which will download
1795 and install the module for you. If not, get the tarball, and run these
1807 Construction can be done OO-style (which is the recommended way), or using
1808 Perl's tie() function. Both are examined here.
1810 =head2 OO CONSTRUCTION
1812 The recommended way to construct a DBM::Deep object is to use the new()
1813 method, which gets you a blessed, tied hash or array reference.
1815 my $db = DBM::Deep->new( "foo.db" );
1817 This opens a new database handle, mapped to the file "foo.db". If this
1818 file does not exist, it will automatically be created. DB files are
1819 opened in "r+" (read/write) mode, and the type of object returned is a
1820 hash, unless otherwise specified (see L<OPTIONS> below).
1824 You can pass a number of options to the constructor to specify things like
1825 locking, autoflush, etc. This is done by passing an inline hash:
1827 my $db = DBM::Deep->new(
1833 Notice that the filename is now specified I<inside> the hash with
1834 the "file" parameter, as opposed to being the sole argument to the
1835 constructor. This is required if any options are specified.
1836 See L<OPTIONS> below for the complete list.
1840 You can also start with an array instead of a hash. For this, you must
1841 specify the C<type> parameter:
1843 my $db = DBM::Deep->new(
1845 type => DBM::Deep->TYPE_ARRAY
1848 B<Note:> Specifing the C<type> parameter only takes effect when beginning
1849 a new DB file. If you create a DBM::Deep object with an existing file, the
1850 C<type> will be loaded from the file header, and ignored if it is passed
1853 =head2 TIE CONSTRUCTION
1855 Alternatively, you can create a DBM::Deep handle by using Perl's built-in
1856 tie() function. This is not ideal, because you get only a basic, tied hash
1857 (or array) which is not blessed, so you can't call any functions on it.
1860 tie %hash, "DBM::Deep", "foo.db";
1863 tie @array, "DBM::Deep", "bar.db";
1865 As with the OO constructor, you can replace the DB filename parameter with
1866 a hash containing one or more options (see L<OPTIONS> just below for the
1869 tie %hash, "DBM::Deep", {
1877 There are a number of options that can be passed in when constructing your
1878 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
1884 Filename of the DB file to link the handle to. You can pass a full absolute
1885 filesystem path, partial path, or a plain filename if the file is in the
1886 current working directory. This is a required parameter.
1890 File open mode (read-only, read-write, etc.) string passed to Perl's FileHandle
1891 module. This is an optional parameter, and defaults to "r+" (read/write).
1892 B<Note:> If the default (r+) mode is selected, the file will also be auto-
1893 created if it doesn't exist.
1897 This parameter specifies what type of object to create, a hash or array. Use
1898 one of these two constants: C<DBM::Deep-E<gt>TYPE_HASH> or C<DBM::Deep-E<gt>TYPE_ARRAY>.
1899 This only takes effect when beginning a new file. This is an optional
1900 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
1904 Specifies whether locking is to be enabled. DBM::Deep uses Perl's Fnctl flock()
1905 function to lock the database in exclusive mode for writes, and shared mode for
1906 reads. Pass any true value to enable. This affects the base DB handle I<and
1907 any child hashes or arrays> that use the same DB file. This is an optional
1908 parameter, and defaults to 0 (disabled). See L<LOCKING> below for more.
1912 Specifies whether autoflush is to be enabled on the underlying FileHandle.
1913 This obviously slows down write operations, but is required if you may have
1914 multiple processes accessing the same DB file (also consider enable I<locking>
1915 or at least I<volatile>). Pass any true value to enable. This is an optional
1916 parameter, and defaults to 0 (disabled).
1920 If I<volatile> mode is enabled, DBM::Deep will stat() the DB file before each
1921 STORE() operation. This is required if an outside force may change the size of
1922 the file between transactions. Locking also implicitly enables volatile. This
1923 is useful if you want to use a different locking system or write your own. Pass
1924 any true value to enable. This is an optional parameter, and defaults to 0
1929 If I<autobless> mode is enabled, DBM::Deep will preserve blessed hashes, and
1930 restore them when fetched. This is an B<experimental> feature, and does have
1931 side-effects. Basically, when hashes are re-blessed into their original
1932 classes, they are no longer blessed into the DBM::Deep class! So you won't be
1933 able to call any DBM::Deep methods on them. You have been warned.
1934 This is an optional parameter, and defaults to 0 (disabled).
1938 See L<FILTERS> below.
1942 Setting I<debug> mode will make all errors non-fatal, dump them out to
1943 STDERR, and continue on. This is for debugging purposes only, and probably
1944 not what you want. This is an optional parameter, and defaults to 0 (disabled).
1948 Instead of passing a file path, you can instead pass a handle to an pre-opened
1949 filehandle. Note: Beware of using the magick *DATA handle, as this actually
1950 contains your entire Perl script, as well as the data following the __DATA__
1951 marker. This will not work, because DBM::Deep uses absolute seek()s into the
1952 file. Instead, consider reading *DATA into an IO::Scalar handle, then passing
1957 =head1 TIE INTERFACE
1959 With DBM::Deep you can access your databases using Perl's standard hash/array
1960 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can treat
1961 them as such. DBM::Deep will intercept all reads/writes and direct them to the right
1962 place -- the DB file. This has nothing to do with the L<TIE CONSTRUCTION>
1963 section above. This simply tells you how to use DBM::Deep using regular hashes
1964 and arrays, rather than calling functions like C<get()> and C<put()> (although those
1965 work too). It is entirely up to you how to want to access your databases.
1969 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
1970 or even nested hashes (or arrays) using standard Perl syntax:
1972 my $db = DBM::Deep->new( "foo.db" );
1974 $db->{mykey} = "myvalue";
1976 $db->{myhash}->{subkey} = "subvalue";
1978 print $db->{myhash}->{subkey} . "\n";
1980 You can even step through hash keys using the normal Perl C<keys()> function:
1982 foreach my $key (keys %$db) {
1983 print "$key: " . $db->{$key} . "\n";
1986 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
1987 pushes them onto an array, all before the loop even begins. If you have an
1988 extra large hash, this may exhaust Perl's memory. Instead, consider using
1989 Perl's C<each()> function, which pulls keys/values one at a time, using very
1992 while (my ($key, $value) = each %$db) {
1993 print "$key: $value\n";
1996 Please note that when using C<each()>, you should always pass a direct
1997 hash reference, not a lookup. Meaning, you should B<never> do this:
2000 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
2002 This causes an infinite loop, because for each iteration, Perl is calling
2003 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
2004 it effectively keeps returning the first key over and over again. Instead,
2005 assign a temporary variable to C<$db->{foo}>, then pass that to each().
2009 As with hashes, you can treat any DBM::Deep object like a normal Perl array
2010 reference. This includes inserting, removing and manipulating elements,
2011 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
2012 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
2013 or simply be a nested array reference inside a hash. Example:
2015 my $db = DBM::Deep->new(
2016 file => "foo-array.db",
2017 type => DBM::Deep->TYPE_ARRAY
2021 push @$db, "bar", "baz";
2022 unshift @$db, "bah";
2024 my $last_elem = pop @$db; # baz
2025 my $first_elem = shift @$db; # bah
2026 my $second_elem = $db->[1]; # bar
2028 my $num_elements = scalar @$db;
2032 In addition to the I<tie()> interface, you can also use a standard OO interface
2033 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
2034 array) has its own methods, but both types share the following common methods:
2035 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>.
2041 Stores a new hash key/value pair, or sets an array element value. Takes two
2042 arguments, the hash key or array index, and the new value. The value can be
2043 a scalar, hash ref or array ref. Returns true on success, false on failure.
2045 $db->put("foo", "bar"); # for hashes
2046 $db->put(1, "bar"); # for arrays
2050 Fetches the value of a hash key or array element. Takes one argument: the hash
2051 key or array index. Returns a scalar, hash ref or array ref, depending on the
2054 my $value = $db->get("foo"); # for hashes
2055 my $value = $db->get(1); # for arrays
2059 Checks if a hash key or array index exists. Takes one argument: the hash key
2060 or array index. Returns true if it exists, false if not.
2062 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
2063 if ($db->exists(1)) { print "yay!\n"; } # for arrays
2067 Deletes one hash key/value pair or array element. Takes one argument: the hash
2068 key or array index. Returns true on success, false if not found. For arrays,
2069 the remaining elements located after the deleted element are NOT moved over.
2070 The deleted element is essentially just undefined, which is exactly how Perl's
2071 internal arrays work. Please note that the space occupied by the deleted
2072 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
2073 below for details and workarounds.
2075 $db->delete("foo"); # for hashes
2076 $db->delete(1); # for arrays
2080 Deletes B<all> hash keys or array elements. Takes no arguments. No return
2081 value. Please note that the space occupied by the deleted keys/values or
2082 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
2083 details and workarounds.
2085 $db->clear(); # hashes or arrays
2091 For hashes, DBM::Deep supports all the common methods described above, and the
2092 following additional methods: C<first_key()> and C<next_key()>.
2098 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
2099 fetched in an undefined order (which appears random). Takes no arguments,
2100 returns the key as a scalar value.
2102 my $key = $db->first_key();
2106 Returns the "next" key in the hash, given the previous one as the sole argument.
2107 Returns undef if there are no more keys to be fetched.
2109 $key = $db->next_key($key);
2113 Here are some examples of using hashes:
2115 my $db = DBM::Deep->new( "foo.db" );
2117 $db->put("foo", "bar");
2118 print "foo: " . $db->get("foo") . "\n";
2120 $db->put("baz", {}); # new child hash ref
2121 $db->get("baz")->put("buz", "biz");
2122 print "buz: " . $db->get("baz")->get("buz") . "\n";
2124 my $key = $db->first_key();
2126 print "$key: " . $db->get($key) . "\n";
2127 $key = $db->next_key($key);
2130 if ($db->exists("foo")) { $db->delete("foo"); }
2134 For arrays, DBM::Deep supports all the common methods described above, and the
2135 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
2136 C<unshift()> and C<splice()>.
2142 Returns the number of elements in the array. Takes no arguments.
2144 my $len = $db->length();
2148 Adds one or more elements onto the end of the array. Accepts scalars, hash
2149 refs or array refs. No return value.
2151 $db->push("foo", "bar", {});
2155 Fetches the last element in the array, and deletes it. Takes no arguments.
2156 Returns undef if array is empty. Returns the element value.
2158 my $elem = $db->pop();
2162 Fetches the first element in the array, deletes it, then shifts all the
2163 remaining elements over to take up the space. Returns the element value. This
2164 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
2167 my $elem = $db->shift();
2171 Inserts one or more elements onto the beginning of the array, shifting all
2172 existing elements over to make room. Accepts scalars, hash refs or array refs.
2173 No return value. This method is not recommended with large arrays -- see
2174 <LARGE ARRAYS> below for details.
2176 $db->unshift("foo", "bar", {});
2180 Performs exactly like Perl's built-in function of the same name. See L<perldoc
2181 -f splice> for usage -- it is too complicated to document here. This method is
2182 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
2186 Here are some examples of using arrays:
2188 my $db = DBM::Deep->new(
2190 type => DBM::Deep->TYPE_ARRAY
2193 $db->push("bar", "baz");
2194 $db->unshift("foo");
2197 my $len = $db->length();
2198 print "length: $len\n"; # 4
2200 for (my $k=0; $k<$len; $k++) {
2201 print "$k: " . $db->get($k) . "\n";
2204 $db->splice(1, 2, "biz", "baf");
2206 while (my $elem = shift @$db) {
2207 print "shifted: $elem\n";
2212 Enable automatic file locking by passing a true value to the C<locking>
2213 parameter when constructing your DBM::Deep object (see L<SETUP> above).
2215 my $db = DBM::Deep->new(
2220 This causes DBM::Deep to C<flock()> the underlying FileHandle object with exclusive
2221 mode for writes, and shared mode for reads. This is required if you have
2222 multiple processes accessing the same database file, to avoid file corruption.
2223 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
2224 NFS> below for more.
2226 =head2 EXPLICIT LOCKING
2228 You can explicitly lock a database, so it remains locked for multiple
2229 transactions. This is done by calling the C<lock()> method, and passing an
2230 optional lock mode argument (defaults to exclusive mode). This is particularly
2231 useful for things like counters, where the current value needs to be fetched,
2232 then incremented, then stored again.
2235 my $counter = $db->get("counter");
2237 $db->put("counter", $counter);
2246 You can pass C<lock()> an optional argument, which specifies which mode to use
2247 (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX>
2248 or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the
2249 same as the constants defined in Perl's C<Fcntl> module.
2251 $db->lock( DBM::Deep->LOCK_SH );
2255 If you want to implement your own file locking scheme, be sure to create your
2256 DBM::Deep objects setting the C<volatile> option to true. This hints to DBM::Deep
2257 that the DB file may change between transactions. See L<LOW-LEVEL ACCESS>
2260 =head1 IMPORTING/EXPORTING
2262 You can import existing complex structures by calling the C<import()> method,
2263 and export an entire database into an in-memory structure using the C<export()>
2264 method. Both are examined here.
2268 Say you have an existing hash with nested hashes/arrays inside it. Instead of
2269 walking the structure and adding keys/elements to the database as you go,
2270 simply pass a reference to the C<import()> method. This recursively adds
2271 everything to an existing DBM::Deep object for you. Here is an example:
2276 array1 => [ "elem0", "elem1", "elem2" ],
2278 subkey1 => "subvalue1",
2279 subkey2 => "subvalue2"
2283 my $db = DBM::Deep->new( "foo.db" );
2284 $db->import( $struct );
2286 print $db->{key1} . "\n"; # prints "value1"
2288 This recursively imports the entire C<$struct> object into C<$db>, including
2289 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
2290 keys are merged with the existing ones, replacing if they already exist.
2291 The C<import()> method can be called on any database level (not just the base
2292 level), and works with both hash and array DB types.
2296 B<Note:> Make sure your existing structure has no circular references in it.
2297 These will cause an infinite loop when importing.
2301 Calling the C<export()> method on an existing DBM::Deep object will return
2302 a reference to a new in-memory copy of the database. The export is done
2303 recursively, so all nested hashes/arrays are all exported to standard Perl
2304 objects. Here is an example:
2306 my $db = DBM::Deep->new( "foo.db" );
2308 $db->{key1} = "value1";
2309 $db->{key2} = "value2";
2311 $db->{hash1}->{subkey1} = "subvalue1";
2312 $db->{hash1}->{subkey2} = "subvalue2";
2314 my $struct = $db->export();
2316 print $struct->{key1} . "\n"; # prints "value1"
2318 This makes a complete copy of the database in memory, and returns a reference
2319 to it. The C<export()> method can be called on any database level (not just
2320 the base level), and works with both hash and array DB types. Be careful of
2321 large databases -- you can store a lot more data in a DBM::Deep object than an
2322 in-memory Perl structure.
2326 B<Note:> Make sure your database has no circular references in it.
2327 These will cause an infinite loop when exporting.
2331 DBM::Deep has a number of hooks where you can specify your own Perl function
2332 to perform filtering on incoming or outgoing data. This is a perfect
2333 way to extend the engine, and implement things like real-time compression or
2334 encryption. Filtering applies to the base DB level, and all child hashes /
2335 arrays. Filter hooks can be specified when your DBM::Deep object is first
2336 constructed, or by calling the C<set_filter()> method at any time. There are
2337 four available filter hooks, described below:
2341 =item * filter_store_key
2343 This filter is called whenever a hash key is stored. It
2344 is passed the incoming key, and expected to return a transformed key.
2346 =item * filter_store_value
2348 This filter is called whenever a hash key or array element is stored. It
2349 is passed the incoming value, and expected to return a transformed value.
2351 =item * filter_fetch_key
2353 This filter is called whenever a hash key is fetched (i.e. via
2354 C<first_key()> or C<next_key()>). It is passed the transformed key,
2355 and expected to return the plain key.
2357 =item * filter_fetch_value
2359 This filter is called whenever a hash key or array element is fetched.
2360 It is passed the transformed value, and expected to return the plain value.
2364 Here are the two ways to setup a filter hook:
2366 my $db = DBM::Deep->new(
2368 filter_store_value => \&my_filter_store,
2369 filter_fetch_value => \&my_filter_fetch
2374 $db->set_filter( "filter_store_value", \&my_filter_store );
2375 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
2377 Your filter function will be called only when dealing with SCALAR keys or
2378 values. When nested hashes and arrays are being stored/fetched, filtering
2379 is bypassed. Filters are called as static functions, passed a single SCALAR
2380 argument, and expected to return a single SCALAR value. If you want to
2381 remove a filter, set the function reference to C<undef>:
2383 $db->set_filter( "filter_store_value", undef );
2385 =head2 REAL-TIME ENCRYPTION EXAMPLE
2387 Here is a working example that uses the I<Crypt::Blowfish> module to
2388 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
2389 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
2390 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
2393 use Crypt::Blowfish;
2396 my $cipher = Crypt::CBC->new({
2397 'key' => 'my secret key',
2398 'cipher' => 'Blowfish',
2400 'regenerate_key' => 0,
2401 'padding' => 'space',
2405 my $db = DBM::Deep->new(
2406 file => "foo-encrypt.db",
2407 filter_store_key => \&my_encrypt,
2408 filter_store_value => \&my_encrypt,
2409 filter_fetch_key => \&my_decrypt,
2410 filter_fetch_value => \&my_decrypt,
2413 $db->{key1} = "value1";
2414 $db->{key2} = "value2";
2415 print "key1: " . $db->{key1} . "\n";
2416 print "key2: " . $db->{key2} . "\n";
2422 return $cipher->encrypt( $_[0] );
2425 return $cipher->decrypt( $_[0] );
2428 =head2 REAL-TIME COMPRESSION EXAMPLE
2430 Here is a working example that uses the I<Compress::Zlib> module to do real-time
2431 compression / decompression of keys & values with DBM::Deep Filters.
2432 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
2433 more on I<Compress::Zlib>.
2438 my $db = DBM::Deep->new(
2439 file => "foo-compress.db",
2440 filter_store_key => \&my_compress,
2441 filter_store_value => \&my_compress,
2442 filter_fetch_key => \&my_decompress,
2443 filter_fetch_value => \&my_decompress,
2446 $db->{key1} = "value1";
2447 $db->{key2} = "value2";
2448 print "key1: " . $db->{key1} . "\n";
2449 print "key2: " . $db->{key2} . "\n";
2455 return Compress::Zlib::memGzip( $_[0] ) ;
2458 return Compress::Zlib::memGunzip( $_[0] ) ;
2461 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
2462 actually numerical index numbers, and are not filtered.
2464 =head1 ERROR HANDLING
2466 Most DBM::Deep methods return a true value for success, and call die() on
2467 failure. You can wrap calls in an eval block to catch the die. Also, the
2468 actual error message is stored in an internal scalar, which can be fetched by
2469 calling the C<error()> method.
2471 my $db = DBM::Deep->new( "foo.db" ); # create hash
2472 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
2474 print $db->error(); # prints error message
2476 You can then call C<clear_error()> to clear the current error state.
2480 If you set the C<debug> option to true when creating your DBM::Deep object,
2481 all errors are considered NON-FATAL, and dumped to STDERR. This is only
2482 for debugging purposes.
2484 =head1 LARGEFILE SUPPORT
2486 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
2487 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
2488 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
2489 by calling the static C<set_pack()> method before you do anything else.
2491 DBM::Deep::set_pack(8, 'Q');
2493 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
2494 instead of 32-bit longs. After setting these values your DB files have a
2495 theoretical maximum size of 16 XB (exabytes).
2499 B<Note:> Changing these values will B<NOT> work for existing database files.
2500 Only change this for new files, and make sure it stays set consistently
2501 throughout the file's life. If you do set these values, you can no longer
2502 access 32-bit DB files. You can, however, call C<set_pack(4, 'N')> to change
2503 back to 32-bit mode.
2507 B<Note:> I have not personally tested files > 2 GB -- all my systems have
2508 only a 32-bit Perl. However, I have received user reports that this does
2511 =head1 LOW-LEVEL ACCESS
2513 If you require low-level access to the underlying FileHandle that DBM::Deep uses,
2514 you can call the C<fh()> method, which returns the handle:
2518 This method can be called on the root level of the datbase, or any child
2519 hashes or arrays. All levels share a I<root> structure, which contains things
2520 like the FileHandle, a reference counter, and all your options you specified
2521 when you created the object. You can get access to this root structure by
2522 calling the C<root()> method.
2524 my $root = $db->root();
2526 This is useful for changing options after the object has already been created,
2527 such as enabling/disabling locking, volatile or debug modes. You can also
2528 store your own temporary user data in this structure (be wary of name
2529 collision), which is then accessible from any child hash or array.
2531 =head1 CUSTOM DIGEST ALGORITHM
2533 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
2534 keys. However you can override this, and use another algorithm (such as SHA-256)
2535 or even write your own. But please note that DBM::Deep currently expects zero
2536 collisions, so your algorithm has to be I<perfect>, so to speak.
2537 Collision detection may be introduced in a later version.
2541 You can specify a custom digest algorithm by calling the static C<set_digest()>
2542 function, passing a reference to a subroutine, and the length of the algorithm's
2543 hashes (in bytes). This is a global static function, which affects ALL DBM::Deep
2544 objects. Here is a working example that uses a 256-bit hash from the
2545 I<Digest::SHA256> module. Please see
2546 L<http://search.cpan.org/search?module=Digest::SHA256> for more.
2551 my $context = Digest::SHA256::new(256);
2553 DBM::Deep::set_digest( \&my_digest, 32 );
2555 my $db = DBM::Deep->new( "foo-sha.db" );
2557 $db->{key1} = "value1";
2558 $db->{key2} = "value2";
2559 print "key1: " . $db->{key1} . "\n";
2560 print "key2: " . $db->{key2} . "\n";
2566 return substr( $context->hash($_[0]), 0, 32 );
2569 B<Note:> Your returned digest strings must be B<EXACTLY> the number
2570 of bytes you specify in the C<set_digest()> function (in this case 32).
2572 =head1 CIRCULAR REFERENCES
2574 DBM::Deep has B<experimental> support for circular references. Meaning you
2575 can have a nested hash key or array element that points to a parent object.
2576 This relationship is stored in the DB file, and is preserved between sessions.
2579 my $db = DBM::Deep->new( "foo.db" );
2582 $db->{circle} = $db; # ref to self
2584 print $db->{foo} . "\n"; # prints "foo"
2585 print $db->{circle}->{foo} . "\n"; # prints "foo" again
2587 One catch is, passing the object to a function that recursively walks the
2588 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
2589 C<export()> methods) will result in an infinite loop. The other catch is,
2590 if you fetch the I<key> of a circular reference (i.e. using the C<first_key()>
2591 or C<next_key()> methods), you will get the I<target object's key>, not the
2592 ref's key. This gets even more interesting with the above example, where
2593 the I<circle> key points to the base DB object, which technically doesn't
2594 have a key. So I made DBM::Deep return "[base]" as the key name in that
2597 =head1 CAVEATS / ISSUES / BUGS
2599 This section describes all the known issues with DBM::Deep. It you have found
2600 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
2602 =head2 UNUSED SPACE RECOVERY
2604 One major caveat with DBM::Deep is that space occupied by existing keys and
2605 values is not recovered when they are deleted. Meaning if you keep deleting
2606 and adding new keys, your file will continuously grow. I am working on this,
2607 but in the meantime you can call the built-in C<optimize()> method from time to
2608 time (perhaps in a crontab or something) to recover all your unused space.
2610 $db->optimize(); # returns true on success
2612 This rebuilds the ENTIRE database into a new file, then moves it on top of
2613 the original. The new file will have no unused space, thus it will take up as
2614 little disk space as possible. Please note that this operation can take
2615 a long time for large files, and you need enough disk space to temporarily hold
2616 2 copies of your DB file. The temporary file is created in the same directory
2617 as the original, named with a ".tmp" extension, and is deleted when the
2618 operation completes. Oh, and if locking is enabled, the DB is automatically
2619 locked for the entire duration of the copy.
2623 B<WARNING:> Only call optimize() on the top-level node of the database, and
2624 make sure there are no child references lying around. DBM::Deep keeps a reference
2625 counter, and if it is greater than 1, optimize() will abort and return undef.
2627 =head2 AUTOVIVIFICATION
2629 Unfortunately, autovivification doesn't work with tied hashes. This appears to
2630 be a bug in Perl's tie() system, as I<Jakob Schmidt> encountered the very same
2631 issue with his I<DWH_FIle> module (see L<http://search.cpan.org/search?module=DWH_File>),
2632 and it is also mentioned in the BUGS section for the I<MLDBM> module <see
2633 L<http://search.cpan.org/search?module=MLDBM>). Basically, on a new db file,
2636 $db->{foo}->{bar} = "hello";
2638 Since "foo" doesn't exist, you cannot add "bar" to it. You end up with "foo"
2639 being an empty hash. Try this instead, which works fine:
2641 $db->{foo} = { bar => "hello" };
2643 As of Perl 5.8.7, this bug still exists. I have walked very carefully through
2644 the execution path, and Perl indeed passes an empty hash to the STORE() method.
2645 Probably a bug in Perl.
2647 =head2 FILE CORRUPTION
2649 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
2650 for a 32-bit signature when opened, but other corruption in files can cause
2651 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
2652 stuck in an infinite loop depending on the level of corruption. File write
2653 operations are not checked for failure (for speed), so if you happen to run
2654 out of disk space, DBM::Deep will probably fail in a bad way. These things will
2655 be addressed in a later version of DBM::Deep.
2659 Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local
2660 filesystems, but will NOT protect you from file corruption over NFS. I've heard
2661 about setting up your NFS server with a locking daemon, then using lockf() to
2662 lock your files, but your milage may vary there as well. From what I
2663 understand, there is no real way to do it. However, if you need access to the
2664 underlying FileHandle in DBM::Deep for using some other kind of locking scheme like
2665 lockf(), see the L<LOW-LEVEL ACCESS> section above.
2667 =head2 COPYING OBJECTS
2669 Beware of copying tied objects in Perl. Very strange things can happen.
2670 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
2671 returns a new, blessed, tied hash or array to the same level in the DB.
2673 my $copy = $db->clone();
2677 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
2678 These functions cause every element in the array to move, which can be murder
2679 on DBM::Deep, as every element has to be fetched from disk, then stored again in
2680 a different location. This may be addressed in a later version.
2684 This section discusses DBM::Deep's speed and memory usage.
2688 Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as
2689 the almighty I<BerkeleyDB>. But it makes up for it in features like true
2690 multi-level hash/array support, and cross-platform FTPable files. Even so,
2691 DBM::Deep is still pretty fast, and the speed stays fairly consistent, even
2692 with huge databases. Here is some test data:
2694 Adding 1,000,000 keys to new DB file...
2696 At 100 keys, avg. speed is 2,703 keys/sec
2697 At 200 keys, avg. speed is 2,642 keys/sec
2698 At 300 keys, avg. speed is 2,598 keys/sec
2699 At 400 keys, avg. speed is 2,578 keys/sec
2700 At 500 keys, avg. speed is 2,722 keys/sec
2701 At 600 keys, avg. speed is 2,628 keys/sec
2702 At 700 keys, avg. speed is 2,700 keys/sec
2703 At 800 keys, avg. speed is 2,607 keys/sec
2704 At 900 keys, avg. speed is 2,190 keys/sec
2705 At 1,000 keys, avg. speed is 2,570 keys/sec
2706 At 2,000 keys, avg. speed is 2,417 keys/sec
2707 At 3,000 keys, avg. speed is 1,982 keys/sec
2708 At 4,000 keys, avg. speed is 1,568 keys/sec
2709 At 5,000 keys, avg. speed is 1,533 keys/sec
2710 At 6,000 keys, avg. speed is 1,787 keys/sec
2711 At 7,000 keys, avg. speed is 1,977 keys/sec
2712 At 8,000 keys, avg. speed is 2,028 keys/sec
2713 At 9,000 keys, avg. speed is 2,077 keys/sec
2714 At 10,000 keys, avg. speed is 2,031 keys/sec
2715 At 20,000 keys, avg. speed is 1,970 keys/sec
2716 At 30,000 keys, avg. speed is 2,050 keys/sec
2717 At 40,000 keys, avg. speed is 2,073 keys/sec
2718 At 50,000 keys, avg. speed is 1,973 keys/sec
2719 At 60,000 keys, avg. speed is 1,914 keys/sec
2720 At 70,000 keys, avg. speed is 2,091 keys/sec
2721 At 80,000 keys, avg. speed is 2,103 keys/sec
2722 At 90,000 keys, avg. speed is 1,886 keys/sec
2723 At 100,000 keys, avg. speed is 1,970 keys/sec
2724 At 200,000 keys, avg. speed is 2,053 keys/sec
2725 At 300,000 keys, avg. speed is 1,697 keys/sec
2726 At 400,000 keys, avg. speed is 1,838 keys/sec
2727 At 500,000 keys, avg. speed is 1,941 keys/sec
2728 At 600,000 keys, avg. speed is 1,930 keys/sec
2729 At 700,000 keys, avg. speed is 1,735 keys/sec
2730 At 800,000 keys, avg. speed is 1,795 keys/sec
2731 At 900,000 keys, avg. speed is 1,221 keys/sec
2732 At 1,000,000 keys, avg. speed is 1,077 keys/sec
2734 This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl
2735 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and
2736 values were between 6 - 12 chars in length. The DB file ended up at 210MB.
2737 Run time was 12 min 3 sec.
2741 One of the great things about DBM::Deep is that it uses very little memory.
2742 Even with huge databases (1,000,000+ keys) you will not see much increased
2743 memory on your process. DBM::Deep relies solely on the filesystem for storing
2744 and fetching data. Here is output from I</usr/bin/top> before even opening a
2747 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
2748 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl
2750 Basically the process is taking 2,716K of memory. And here is the same
2751 process after storing and fetching 1,000,000 keys:
2753 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
2754 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl
2756 Notice the memory usage increased by only 56K. Test was performed on a 700mHz
2757 x86 box running Linux RedHat 7.2 & Perl 5.6.1.
2759 =head1 DB FILE FORMAT
2761 In case you were interested in the underlying DB file format, it is documented
2762 here in this section. You don't need to know this to use the module, it's just
2763 included for reference.
2767 DBM::Deep files always start with a 32-bit signature to identify the file type.
2768 This is at offset 0. The signature is "DPDB" in network byte order. This is
2769 checked when the file is opened.
2773 The DBM::Deep file is in a I<tagged format>, meaning each section of the file
2774 has a standard header containing the type of data, the length of data, and then
2775 the data itself. The type is a single character (1 byte), the length is a
2776 32-bit unsigned long in network byte order, and the data is, well, the data.
2777 Here is how it unfolds:
2781 Immediately after the 32-bit file signature is the I<Master Index> record.
2782 This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048
2783 bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array,
2784 depending on how the DBM::Deep object was constructed.
2788 The index works by looking at a I<MD5 Hash> of the hash key (or array index
2789 number). The first 8-bit char of the MD5 signature is the offset into the
2790 index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the
2791 index element is a file offset of the next tag for the key/element in question,
2792 which is usually a I<Bucket List> tag (see below).
2796 The next tag I<could> be another index, depending on how many keys/elements
2797 exist. See L<RE-INDEXING> below for details.
2801 A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus
2802 file offsets to where the actual data is stored. It starts with a standard
2803 tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or
2804 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus
2805 the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data.
2806 When the list fills up, a I<Re-Index> operation is performed (See
2807 L<RE-INDEXING> below).
2811 A I<Bucket> is a tag containing a key/value pair (in hash mode), or a
2812 index/value pair (in array mode). It starts with a standard tag header with
2813 type I<D> for scalar data (string, binary, etc.), or it could be a nested
2814 hash (type I<H>) or array (type I<A>). The value comes just after the tag
2815 header. The size reported in the tag header is only for the value, but then,
2816 just after the value is another size (32-bit unsigned long) and then the plain
2817 key itself. Since the value is likely to be fetched more often than the plain
2818 key, I figured it would be I<slightly> faster to store the value first.
2822 If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index>
2823 record for the nested structure, where the process begins all over again.
2827 After a I<Bucket List> grows to 16 records, its allocated space in the file is
2828 exhausted. Then, when another key/element comes in, the list is converted to a
2829 new index record. However, this index will look at the next char in the MD5
2830 hash, and arrange new Bucket List pointers accordingly. This process is called
2831 I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all
2832 17 (16 + new one) keys/elements are removed from the old Bucket List and
2833 inserted into the new index. Several new Bucket Lists are created in the
2834 process, as a new MD5 char from the key is being examined (it is unlikely that
2835 the keys will all share the same next char of their MD5s).
2839 Because of the way the I<MD5> algorithm works, it is impossible to tell exactly
2840 when the Bucket Lists will turn into indexes, but the first round tends to
2841 happen right around 4,000 keys. You will see a I<slight> decrease in
2842 performance here, but it picks back up pretty quick (see L<SPEED> above). Then
2843 it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's
2844 right around 900,000 keys. This process can continue nearly indefinitely --
2845 right up until the point the I<MD5> signatures start colliding with each other,
2846 and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND
2847 getting struck by lightning while you are walking to cash in your tickets.
2848 Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to
2849 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe
2850 this is 340 unodecillion, but don't quote me).
2854 When a new key/element is stored, the key (or index number) is first ran through
2855 I<Digest::MD5> to get a 128-bit signature (example, in hex:
2856 b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked
2857 for the first char of the signature (in this case I<b>). If it does not exist,
2858 a new I<Bucket List> is created for our key (and the next 15 future keys that
2859 happen to also have I<b> as their first MD5 char). The entire MD5 is written
2860 to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at
2861 this point, unless we are replacing an existing I<Bucket>), where the actual
2862 data will be stored.
2866 Fetching an existing key/element involves getting a I<Digest::MD5> of the key
2867 (or index number), then walking along the indexes. If there are enough
2868 keys/elements in this DB level, there might be nested indexes, each linked to
2869 a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which
2870 contains up to 16 full MD5 hashes. Each is checked for equality to the key in
2871 question. If we found a match, the I<Bucket> tag is loaded, where the value and
2872 plain key are stored.
2876 Fetching the plain key occurs when calling the I<first_key()> and I<next_key()>
2877 methods. In this process the indexes are walked systematically, and each key
2878 fetched in increasing MD5 order (which is why it appears random). Once the
2879 I<Bucket> is found, the value is skipped the plain key returned instead.
2880 B<Note:> Do not count on keys being fetched as if the MD5 hashes were
2881 alphabetically sorted. This only happens on an index-level -- as soon as the
2882 I<Bucket Lists> are hit, the keys will come out in the order they went in --
2883 so it's pretty much undefined how the keys will come out -- just like Perl's
2886 =head1 CODE COVERAGE
2888 I use B<Devel::Cover> to test the code coverage of my tests, below is the B<Devel::Cover> report on this
2889 module's test suite.
2891 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2892 File stmt bran cond sub pod time total
2893 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2894 blib/lib/DBM/Deep.pm 94.9 84.5 77.8 100.0 11.1 100.0 89.7
2895 Total 94.9 84.5 77.8 100.0 11.1 100.0 89.7
2896 ---------------------------- ------ ------ ------ ------ ------ ------ ------
2900 Joseph Huckaby, L<jhuckaby@cpan.org>
2902 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
2906 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
2907 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
2911 Copyright (c) 2002-2005 Joseph Huckaby. All Rights Reserved.
2912 This is free software, you may use it and distribute it under the
2913 same terms as Perl itself.