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 return $class->_init($args);
204 # Tied array constructor method, called by Perl's tie() function.
208 if (scalar(@_) > 1) { $args = {@_}; }
209 #XXX This use of ref() is bad and is a bug
210 elsif (ref($_[0])) { $args = $_[0]; }
211 else { $args = { file => shift }; }
213 return $class->_init($args);
218 # Class deconstructor. Close file handle if there are no more refs.
220 my $self = _get_self($_[0]);
223 $self->root->{links}--;
225 if (!$self->root->{links}) {
232 # Open a FileHandle to the database, create if nonexistent.
233 # Make sure file signature matches DeepDB spec.
235 my $self = _get_self($_[0]);
237 if (defined($self->fh)) { $self->close(); }
239 if (!(-e $self->root->{file}) && $self->root->{mode} eq 'r+') {
240 my $temp = FileHandle->new( $self->root->{file}, 'w' );
244 #XXX Convert to set_fh()
245 $self->root->{fh} = FileHandle->new( $self->root->{file}, $self->root->{mode} );
246 if (! defined($self->fh)) {
247 return $self->throw_error("Cannot open file: " . $self->root->{file} . ": $!");
250 binmode $self->fh; # for win32
251 if ($self->root->{autoflush}) {
252 $self->fh->autoflush();
256 seek($self->fh, 0, 0);
257 my $bytes_read = $self->fh->read($signature, length(SIG_FILE));
260 # File is empty -- write signature and master index
263 seek($self->fh, 0, 0);
264 $self->fh->print(SIG_FILE);
265 $self->root->{end} = length(SIG_FILE);
266 $self->create_tag($self->base_offset, $self->type, chr(0) x $INDEX_SIZE);
268 my $plain_key = "[base]";
269 $self->fh->print( pack($DATA_LENGTH_PACK, length($plain_key)) . $plain_key );
270 $self->root->{end} += $DATA_LENGTH_SIZE + length($plain_key);
277 # Check signature was valid
279 unless ($signature eq SIG_FILE) {
281 return $self->throw_error("Signature not found -- file is not a Deep DB");
284 $self->root->{end} = (stat($self->fh))[7];
287 # Get our type from master index signature
289 my $tag = $self->load_tag($self->base_offset);
290 #XXX This is a problem - need to verify type, not override it!
291 #XXX We probably also want to store the hash algorithm name, not assume anything
292 #XXX Convert to set_type() when one is written
293 $self->{type} = $tag->{signature};
300 # Close database FileHandle
302 my $self = _get_self($_[0]);
303 undef $self->root->{fh};
308 # Given offset, signature and content, create tag and write to disk
310 my ($self, $offset, $sig, $content) = @_;
311 my $size = length($content);
313 seek($self->fh, $offset, 0);
314 $self->fh->print( $sig . pack($DATA_LENGTH_PACK, $size) . $content );
316 if ($offset == $self->root->{end}) {
317 $self->root->{end} += SIG_SIZE + $DATA_LENGTH_SIZE + $size;
323 offset => $offset + SIG_SIZE + $DATA_LENGTH_SIZE,
330 # Given offset, load single tag and return signature, size and data
335 seek($self->fh, $offset, 0);
336 if ($self->fh->eof()) { return; }
339 $self->fh->read($sig, SIG_SIZE);
342 $self->fh->read($size, $DATA_LENGTH_SIZE);
343 $size = unpack($DATA_LENGTH_PACK, $size);
346 $self->fh->read($buffer, $size);
351 offset => $offset + SIG_SIZE + $DATA_LENGTH_SIZE,
358 # Given index tag, lookup single entry in index and return .
361 my ($tag, $index) = @_;
363 my $location = unpack($LONG_PACK, substr($tag->{content}, $index * $LONG_SIZE, $LONG_SIZE) );
364 if (!$location) { return; }
366 return $self->load_tag( $location );
371 # Adds one key/value pair to bucket list, given offset, MD5 digest of key,
372 # plain (undigested) key and value.
375 my ($tag, $md5, $plain_key, $value) = @_;
376 my $keys = $tag->{content};
380 my $is_dbm_deep = eval { $value->isa( 'DBM::Deep' ) };
381 my $internal_ref = $is_dbm_deep && ($value->root eq $self->root);
384 # Iterate through buckets, seeing if this is a new entry or a replace.
386 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
387 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
388 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
391 # Found empty bucket (end of list). Populate and exit loop.
395 if ($internal_ref) { $location = $value->base_offset; }
396 else { $location = $self->root->{end}; }
398 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
399 $self->fh->print( $md5 . pack($LONG_PACK, $location) );
402 elsif ($md5 eq $key) {
404 # Found existing bucket with same key. Replace with new value.
409 $location = $value->base_offset;
410 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
411 $self->fh->print( $md5 . pack($LONG_PACK, $location) );
414 seek($self->fh, $subloc + SIG_SIZE, 0);
416 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
419 # If value is a hash, array, or raw value with equal or less size, we can
420 # reuse the same content area of the database. Otherwise, we have to create
421 # a new content area at the EOF.
424 my $r = Scalar::Util::reftype( $value ) || '';
425 if ( $r eq 'HASH' || $r eq 'ARRAY' ) { $actual_length = $INDEX_SIZE; }
426 else { $actual_length = length($value); }
428 if ($actual_length <= $size) {
432 $location = $self->root->{end};
433 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE) + $HASH_SIZE, 0);
434 $self->fh->print( pack($LONG_PACK, $location) );
442 # If this is an internal reference, return now.
443 # No need to write value or plain key
450 # If bucket didn't fit into list, split into a new index level
453 seek($self->fh, $tag->{ref_loc}, 0);
454 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
456 my $index_tag = $self->create_tag($self->root->{end}, SIG_INDEX, chr(0) x $INDEX_SIZE);
459 $keys .= $md5 . pack($LONG_PACK, 0);
461 for (my $i=0; $i<=$MAX_BUCKETS; $i++) {
462 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
464 my $old_subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
465 my $num = ord(substr($key, $tag->{ch} + 1, 1));
467 if ($offsets[$num]) {
468 my $offset = $offsets[$num] + SIG_SIZE + $DATA_LENGTH_SIZE;
469 seek($self->fh, $offset, 0);
471 $self->fh->read($subkeys, $BUCKET_LIST_SIZE);
473 for (my $k=0; $k<$MAX_BUCKETS; $k++) {
474 my $subloc = unpack($LONG_PACK, substr($subkeys, ($k * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
476 seek($self->fh, $offset + ($k * $BUCKET_SIZE), 0);
477 $self->fh->print( $key . pack($LONG_PACK, $old_subloc || $self->root->{end}) );
483 $offsets[$num] = $self->root->{end};
484 seek($self->fh, $index_tag->{offset} + ($num * $LONG_SIZE), 0);
485 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
487 my $blist_tag = $self->create_tag($self->root->{end}, SIG_BLIST, chr(0) x $BUCKET_LIST_SIZE);
489 seek($self->fh, $blist_tag->{offset}, 0);
490 $self->fh->print( $key . pack($LONG_PACK, $old_subloc || $self->root->{end}) );
495 $location ||= $self->root->{end};
496 } # re-index bucket list
499 # Seek to content area and store signature, value and plaintext key
503 seek($self->fh, $location, 0);
506 # Write signature based on content type, set content length and write actual value.
508 my $r = Scalar::Util::reftype($value) || '';
510 $self->fh->print( TYPE_HASH );
511 $self->fh->print( pack($DATA_LENGTH_PACK, $INDEX_SIZE) . chr(0) x $INDEX_SIZE );
512 $content_length = $INDEX_SIZE;
514 elsif ($r eq 'ARRAY') {
515 $self->fh->print( TYPE_ARRAY );
516 $self->fh->print( pack($DATA_LENGTH_PACK, $INDEX_SIZE) . chr(0) x $INDEX_SIZE );
517 $content_length = $INDEX_SIZE;
519 elsif (!defined($value)) {
520 $self->fh->print( SIG_NULL );
521 $self->fh->print( pack($DATA_LENGTH_PACK, 0) );
525 $self->fh->print( SIG_DATA );
526 $self->fh->print( pack($DATA_LENGTH_PACK, length($value)) . $value );
527 $content_length = length($value);
531 # Plain key is stored AFTER value, as keys are typically fetched less often.
533 $self->fh->print( pack($DATA_LENGTH_PACK, length($plain_key)) . $plain_key );
536 # If value is blessed, preserve class name
538 my $value_class = Scalar::Util::blessed($value);
539 if ($self->root->{autobless} && defined $value_class && $value_class ne 'DBM::Deep' ) {
541 # Blessed ref -- will restore later
543 $self->fh->print( chr(1) );
544 $self->fh->print( pack($DATA_LENGTH_PACK, length($value_class)) . $value_class );
545 $content_length += 1;
546 $content_length += $DATA_LENGTH_SIZE + length($value_class);
550 # If this is a new content area, advance EOF counter
552 if ($location == $self->root->{end}) {
553 $self->root->{end} += SIG_SIZE;
554 $self->root->{end} += $DATA_LENGTH_SIZE + $content_length;
555 $self->root->{end} += $DATA_LENGTH_SIZE + length($plain_key);
559 # If content is a hash or array, create new child DeepDB object and
560 # pass each key or element to it.
563 my $branch = DBM::Deep->new(
565 base_offset => $location,
568 foreach my $key (keys %{$value}) {
569 $branch->{$key} = $value->{$key};
572 elsif ($r eq 'ARRAY') {
573 my $branch = DBM::Deep->new(
575 base_offset => $location,
579 foreach my $element (@{$value}) {
580 $branch->[$index] = $element;
588 return $self->throw_error("Fatal error: indexing failed -- possibly due to corruption in file");
591 sub get_bucket_value {
593 # Fetch single value given tag and MD5 digested key.
596 my ($tag, $md5) = @_;
597 my $keys = $tag->{content};
600 # Iterate through buckets, looking for a key match
603 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
604 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
605 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
609 # Hit end of list, no match
614 if ( $md5 ne $key ) {
619 # Found match -- seek to offset and read signature
622 seek($self->fh, $subloc, 0);
623 $self->fh->read($signature, SIG_SIZE);
626 # If value is a hash or array, return new DeepDB object with correct offset
628 if (($signature eq TYPE_HASH) || ($signature eq TYPE_ARRAY)) {
629 my $obj = DBM::Deep->new(
631 base_offset => $subloc,
635 if ($self->root->{autobless}) {
637 # Skip over value and plain key to see if object needs
640 seek($self->fh, $DATA_LENGTH_SIZE + $INDEX_SIZE, 1);
643 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
644 if ($size) { seek($self->fh, $size, 1); }
647 $self->fh->read($bless_bit, 1);
648 if (ord($bless_bit)) {
650 # Yes, object needs to be re-blessed
653 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
654 if ($size) { $self->fh->read($class_name, $size); }
655 if ($class_name) { $obj = bless( $obj, $class_name ); }
663 # Otherwise return actual value
665 elsif ($signature eq SIG_DATA) {
668 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
669 if ($size) { $self->fh->read($value, $size); }
674 # Key exists, but content is null
684 # Delete single key/value pair given tag and MD5 digested key.
687 my ($tag, $md5) = @_;
688 my $keys = $tag->{content};
691 # Iterate through buckets, looking for a key match
694 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
695 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
696 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
700 # Hit end of list, no match
705 if ( $md5 ne $key ) {
710 # Matched key -- delete bucket and return
712 seek($self->fh, $tag->{offset} + ($i * $BUCKET_SIZE), 0);
713 $self->fh->print( substr($keys, ($i+1) * $BUCKET_SIZE ) );
714 $self->fh->print( chr(0) x $BUCKET_SIZE );
724 # Check existence of single key given tag and MD5 digested key.
727 my ($tag, $md5) = @_;
728 my $keys = $tag->{content};
731 # Iterate through buckets, looking for a key match
734 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
735 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
736 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
740 # Hit end of list, no match
745 if ( $md5 ne $key ) {
750 # Matched key -- return true
758 sub find_bucket_list {
760 # Locate offset for bucket list, given digested key
766 # Locate offset for bucket list using digest index system
769 my $tag = $self->load_tag($self->base_offset);
770 if (!$tag) { return; }
772 while ($tag->{signature} ne SIG_BLIST) {
773 $tag = $self->index_lookup($tag, ord(substr($md5, $ch, 1)));
774 if (!$tag) { return; }
783 # Scan index and recursively step into deeper levels, looking for next key.
785 my ($self, $offset, $ch, $force_return_next) = @_;
786 $force_return_next = undef unless $force_return_next;
788 my $tag = $self->load_tag( $offset );
790 if ($tag->{signature} ne SIG_BLIST) {
791 my $content = $tag->{content};
793 if ($self->{return_next}) { $start = 0; }
794 else { $start = ord(substr($self->{prev_md5}, $ch, 1)); }
796 for (my $index = $start; $index < 256; $index++) {
797 my $subloc = unpack($LONG_PACK, substr($content, $index * $LONG_SIZE, $LONG_SIZE) );
799 my $result = $self->traverse_index( $subloc, $ch + 1, $force_return_next );
800 if (defined($result)) { return $result; }
804 $self->{return_next} = 1;
807 elsif ($tag->{signature} eq SIG_BLIST) {
808 my $keys = $tag->{content};
809 if ($force_return_next) { $self->{return_next} = 1; }
812 # Iterate through buckets, looking for a key match
814 for (my $i=0; $i<$MAX_BUCKETS; $i++) {
815 my $key = substr($keys, $i * $BUCKET_SIZE, $HASH_SIZE);
816 my $subloc = unpack($LONG_PACK, substr($keys, ($i * $BUCKET_SIZE) + $HASH_SIZE, $LONG_SIZE));
820 # End of bucket list -- return to outer loop
822 $self->{return_next} = 1;
825 elsif ($key eq $self->{prev_md5}) {
827 # Located previous key -- return next one found
829 $self->{return_next} = 1;
832 elsif ($self->{return_next}) {
834 # Seek to bucket location and skip over signature
836 seek($self->fh, $subloc + SIG_SIZE, 0);
839 # Skip over value to get to plain key
842 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
843 if ($size) { seek($self->fh, $size, 1); }
846 # Read in plain key and return as scalar
849 $self->fh->read($size, $DATA_LENGTH_SIZE); $size = unpack($DATA_LENGTH_PACK, $size);
850 if ($size) { $self->fh->read($plain_key, $size); }
856 $self->{return_next} = 1;
857 } # tag is a bucket list
864 # Locate next key, given digested previous one
866 my $self = _get_self($_[0]);
868 $self->{prev_md5} = $_[1] ? $_[1] : undef;
869 $self->{return_next} = 0;
872 # If the previous key was not specifed, start at the top and
873 # return the first one found.
875 if (!$self->{prev_md5}) {
876 $self->{prev_md5} = chr(0) x $HASH_SIZE;
877 $self->{return_next} = 1;
880 return $self->traverse_index( $self->base_offset, 0 );
885 # If db locking is set, flock() the db file. If called multiple
886 # times before unlock(), then the same number of unlocks() must
887 # be called before the lock is released.
889 my $self = _get_self($_[0]);
891 $type = LOCK_EX unless defined $type;
893 if ($self->root->{locking}) {
894 if (!$self->root->{locked}) { flock($self->fh, $type); }
895 $self->root->{locked}++;
901 # If db locking is set, unlock the db file. See note in lock()
902 # regarding calling lock() multiple times.
904 my $self = _get_self($_[0]);
906 if ($self->root->{locking} && $self->root->{locked} > 0) {
907 $self->root->{locked}--;
908 if (!$self->root->{locked}) { flock($self->fh, LOCK_UN); }
912 #XXX These uses of ref() need verified
915 # Copy single level of keys or elements to new DB handle.
916 # Recurse for nested structures
918 my $self = _get_self($_[0]);
921 if ($self->type eq TYPE_HASH) {
922 my $key = $self->first_key();
924 my $value = $self->get($key);
925 if (!ref($value)) { $db_temp->{$key} = $value; }
927 my $type = $value->type;
928 if ($type eq TYPE_HASH) { $db_temp->{$key} = {}; }
929 else { $db_temp->{$key} = []; }
930 $value->copy_node( $db_temp->{$key} );
932 $key = $self->next_key($key);
936 my $length = $self->length();
937 for (my $index = 0; $index < $length; $index++) {
938 my $value = $self->get($index);
939 if (!ref($value)) { $db_temp->[$index] = $value; }
940 #XXX NO tests for this code
942 my $type = $value->type;
943 if ($type eq TYPE_HASH) { $db_temp->[$index] = {}; }
944 else { $db_temp->[$index] = []; }
945 $value->copy_node( $db_temp->[$index] );
953 # Recursively export into standard Perl hashes and arrays.
955 my $self = _get_self($_[0]);
958 if ($self->type eq TYPE_HASH) { $temp = {}; }
959 elsif ($self->type eq TYPE_ARRAY) { $temp = []; }
962 $self->copy_node( $temp );
970 # Recursively import Perl hash/array structure
972 #XXX This use of ref() seems to be ok
973 if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore
975 my $self = _get_self($_[0]);
978 #XXX This use of ref() seems to be ok
981 # struct is not a reference, so just import based on our type
985 if ($self->type eq TYPE_HASH) { $struct = {@_}; }
986 elsif ($self->type eq TYPE_ARRAY) { $struct = [@_]; }
989 my $r = Scalar::Util::reftype($struct) || '';
990 if ($r eq "HASH" && $self->type eq TYPE_HASH) {
991 foreach my $key (keys %$struct) { $self->put($key, $struct->{$key}); }
993 elsif ($r eq "ARRAY" && $self->type eq TYPE_ARRAY) {
994 $self->push( @$struct );
997 return $self->throw_error("Cannot import: type mismatch");
1005 # Rebuild entire database into new file, then move
1006 # it back on top of original.
1008 my $self = _get_self($_[0]);
1009 if ($self->root->{links} > 1) {
1010 return $self->throw_error("Cannot optimize: reference count is greater than 1");
1013 my $db_temp = DBM::Deep->new(
1014 file => $self->root->{file} . '.tmp',
1018 return $self->throw_error("Cannot optimize: failed to open temp file: $!");
1022 $self->copy_node( $db_temp );
1026 # Attempt to copy user, group and permissions over to new file
1028 my @stats = stat($self->fh);
1029 my $perms = $stats[2] & 07777;
1030 my $uid = $stats[4];
1031 my $gid = $stats[5];
1032 chown( $uid, $gid, $self->root->{file} . '.tmp' );
1033 chmod( $perms, $self->root->{file} . '.tmp' );
1035 # q.v. perlport for more information on this variable
1036 if ( $^O eq 'MSWin32' ) {
1038 # Potential race condition when optmizing on Win32 with locking.
1039 # The Windows filesystem requires that the filehandle be closed
1040 # before it is overwritten with rename(). This could be redone
1047 if (!rename $self->root->{file} . '.tmp', $self->root->{file}) {
1048 unlink $self->root->{file} . '.tmp';
1050 return $self->throw_error("Optimize failed: Cannot copy temp file over original: $!");
1062 # Make copy of object and return
1064 my $self = _get_self($_[0]);
1066 return DBM::Deep->new(
1067 type => $self->type,
1068 base_offset => $self->base_offset,
1074 my %is_legal_filter = map {
1077 store_key store_value
1078 fetch_key fetch_value
1083 # Setup filter function for storing or fetching the key or value
1085 my $self = _get_self($_[0]);
1086 my $type = lc $_[1];
1087 my $func = $_[2] ? $_[2] : undef;
1089 if ( $is_legal_filter{$type} ) {
1090 $self->root->{"filter_$type"} = $func;
1104 # Get access to the root structure
1106 my $self = _get_self($_[0]);
1107 return $self->{root};
1112 # Get access to the raw FileHandle
1114 #XXX It will be useful, though, when we split out HASH and ARRAY
1115 my $self = _get_self($_[0]);
1116 return $self->root->{fh};
1121 # Get type of current node (TYPE_HASH or TYPE_ARRAY)
1123 my $self = _get_self($_[0]);
1124 return $self->{type};
1129 # Get base_offset of current node (TYPE_HASH or TYPE_ARRAY)
1131 my $self = _get_self($_[0]);
1132 return $self->{base_offset};
1137 # Get last error string, or undef if no error
1140 ? ( _get_self($_[0])->{root}->{error} or undef )
1150 # Store error string in self
1152 my $self = _get_self($_[0]);
1153 my $error_text = $_[1];
1155 $self->root->{error} = $error_text;
1157 unless ($self->root->{debug}) {
1158 die "DBM::Deep: $error_text\n";
1161 warn "DBM::Deep: $error_text\n";
1169 my $self = _get_self($_[0]);
1171 undef $self->root->{error};
1176 # Precalculate index, bucket and bucket list sizes
1179 #XXX I don't like this ...
1180 set_pack() unless defined $LONG_SIZE;
1182 $INDEX_SIZE = 256 * $LONG_SIZE;
1183 $BUCKET_SIZE = $HASH_SIZE + $LONG_SIZE;
1184 $BUCKET_LIST_SIZE = $MAX_BUCKETS * $BUCKET_SIZE;
1189 # Set pack/unpack modes (see file header for more)
1191 my ($long_s, $long_p, $data_s, $data_p) = @_;
1193 $LONG_SIZE = $long_s ? $long_s : 4;
1194 $LONG_PACK = $long_p ? $long_p : 'N';
1196 $DATA_LENGTH_SIZE = $data_s ? $data_s : 4;
1197 $DATA_LENGTH_PACK = $data_p ? $data_p : 'N';
1204 # Set key digest function (default is MD5)
1206 my ($digest_func, $hash_size) = @_;
1208 $DIGEST_FUNC = $digest_func ? $digest_func : \&Digest::MD5::md5;
1209 $HASH_SIZE = $hash_size ? $hash_size : 16;
1215 # tie() methods (hashes and arrays)
1220 # Store single hash key/value or array element in database.
1222 my $self = _get_self($_[0]);
1223 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1224 #XXX What is ref() checking here?
1225 my $value = ($self->root->{filter_store_value} && !ref($_[2])) ? $self->root->{filter_store_value}->($_[2]) : $_[2];
1227 my $unpacked_key = $key;
1228 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1229 my $md5 = $DIGEST_FUNC->($key);
1232 # Make sure file is open
1234 if (!defined($self->fh) && !$self->_open()) {
1239 # Request exclusive lock for writing
1241 $self->lock( LOCK_EX );
1244 # If locking is enabled, set 'end' parameter again, in case another
1245 # DB instance appended to our file while we were unlocked.
1247 if ($self->root->{locking} || $self->root->{volatile}) {
1248 $self->root->{end} = (stat($self->fh))[7];
1252 # Locate offset for bucket list using digest index system
1254 my $tag = $self->load_tag($self->base_offset);
1256 $tag = $self->create_tag($self->base_offset, SIG_INDEX, chr(0) x $INDEX_SIZE);
1260 while ($tag->{signature} ne SIG_BLIST) {
1261 my $num = ord(substr($md5, $ch, 1));
1262 my $new_tag = $self->index_lookup($tag, $num);
1264 my $ref_loc = $tag->{offset} + ($num * $LONG_SIZE);
1265 seek($self->fh, $ref_loc, 0);
1266 $self->fh->print( pack($LONG_PACK, $self->root->{end}) );
1268 $tag = $self->create_tag($self->root->{end}, SIG_BLIST, chr(0) x $BUCKET_LIST_SIZE);
1269 $tag->{ref_loc} = $ref_loc;
1274 my $ref_loc = $tag->{offset} + ($num * $LONG_SIZE);
1276 $tag->{ref_loc} = $ref_loc;
1283 # Add key/value to bucket list
1285 my $result = $self->add_bucket( $tag, $md5, $key, $value );
1288 # If this object is an array, and bucket was not a replace, and key is numerical,
1289 # and index is equal or greater than current length, advance length variable.
1291 if (($result == 2) && ($self->type eq TYPE_ARRAY) && ($unpacked_key =~ /^\d+$/) && ($unpacked_key >= $self->FETCHSIZE())) {
1292 $self->STORESIZE( $unpacked_key + 1 );
1302 # Fetch single value or element given plain key or array index
1304 my $self = _get_self($_[0]);
1307 if ( $self->type eq TYPE_HASH ) {
1308 if ( my $filter = $self->root->{filter_store_key} ) {
1309 $key = $filter->( $key );
1312 elsif ( $self->type eq TYPE_ARRAY ) {
1313 if ( $key =~ /^\d+$/ ) {
1314 $key = pack($LONG_PACK, $key);
1318 my $md5 = $DIGEST_FUNC->($key);
1321 # Make sure file is open
1323 if (!defined($self->fh)) { $self->_open(); }
1326 # Request shared lock for reading
1328 $self->lock( LOCK_SH );
1330 my $tag = $self->find_bucket_list( $md5 );
1337 # Get value from bucket list
1339 my $result = $self->get_bucket_value( $tag, $md5 );
1343 #XXX What is ref() checking here?
1344 return ($result && !ref($result) && $self->root->{filter_fetch_value}) ? $self->root->{filter_fetch_value}->($result) : $result;
1349 # Delete single key/value pair or element given plain key or array index
1351 my $self = _get_self($_[0]);
1352 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1354 my $unpacked_key = $key;
1355 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1356 my $md5 = $DIGEST_FUNC->($key);
1359 # Make sure file is open
1361 if (!defined($self->fh)) { $self->_open(); }
1364 # Request exclusive lock for writing
1366 $self->lock( LOCK_EX );
1368 my $tag = $self->find_bucket_list( $md5 );
1377 my $result = $self->delete_bucket( $tag, $md5 );
1380 # If this object is an array and the key deleted was on the end of the stack,
1381 # decrement the length variable.
1383 if ($result && ($self->type eq TYPE_ARRAY) && ($unpacked_key == $self->FETCHSIZE() - 1)) {
1384 $self->STORESIZE( $unpacked_key );
1394 # Check if a single key or element exists given plain key or array index
1396 my $self = _get_self($_[0]);
1397 my $key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1399 if (($self->type eq TYPE_ARRAY) && ($key =~ /^\d+$/)) { $key = pack($LONG_PACK, $key); }
1400 my $md5 = $DIGEST_FUNC->($key);
1403 # Make sure file is open
1405 if (!defined($self->fh)) { $self->_open(); }
1408 # Request shared lock for reading
1410 $self->lock( LOCK_SH );
1412 my $tag = $self->find_bucket_list( $md5 );
1415 # For some reason, the built-in exists() function returns '' for false
1423 # Check if bucket exists and return 1 or ''
1425 my $result = $self->bucket_exists( $tag, $md5 ) || '';
1434 # Clear all keys from hash, or all elements from array.
1436 my $self = _get_self($_[0]);
1439 # Make sure file is open
1441 if (!defined($self->fh)) { $self->_open(); }
1444 # Request exclusive lock for writing
1446 $self->lock( LOCK_EX );
1448 seek($self->fh, $self->base_offset, 0);
1449 if ($self->fh->eof()) {
1454 $self->create_tag($self->base_offset, $self->type, chr(0) x $INDEX_SIZE);
1463 # Locate and return first key (in no particular order)
1465 my $self = _get_self($_[0]);
1466 if ($self->type ne TYPE_HASH) {
1467 return $self->throw_error("FIRSTKEY method only supported for hashes");
1471 # Make sure file is open
1473 if (!defined($self->fh)) { $self->_open(); }
1476 # Request shared lock for reading
1478 $self->lock( LOCK_SH );
1480 my $result = $self->get_next_key();
1484 return ($result && $self->root->{filter_fetch_key}) ? $self->root->{filter_fetch_key}->($result) : $result;
1489 # Return next key (in no particular order), given previous one
1491 my $self = _get_self($_[0]);
1492 if ($self->type ne TYPE_HASH) {
1493 return $self->throw_error("NEXTKEY method only supported for hashes");
1495 my $prev_key = ($self->root->{filter_store_key} && $self->type eq TYPE_HASH) ? $self->root->{filter_store_key}->($_[1]) : $_[1];
1496 my $prev_md5 = $DIGEST_FUNC->($prev_key);
1499 # Make sure file is open
1501 if (!defined($self->fh)) { $self->_open(); }
1504 # Request shared lock for reading
1506 $self->lock( LOCK_SH );
1508 my $result = $self->get_next_key( $prev_md5 );
1512 return ($result && $self->root->{filter_fetch_key}) ? $self->root->{filter_fetch_key}->($result) : $result;
1516 # The following methods are for arrays only
1521 # Return the length of the array
1523 my $self = _get_self($_[0]);
1524 if ($self->type ne TYPE_ARRAY) {
1525 return $self->throw_error("FETCHSIZE method only supported for arrays");
1528 my $SAVE_FILTER = $self->root->{filter_fetch_value};
1529 $self->root->{filter_fetch_value} = undef;
1531 my $packed_size = $self->FETCH('length');
1533 $self->root->{filter_fetch_value} = $SAVE_FILTER;
1535 if ($packed_size) { return int(unpack($LONG_PACK, $packed_size)); }
1541 # Set the length of the array
1543 my $self = _get_self($_[0]);
1544 if ($self->type ne TYPE_ARRAY) {
1545 return $self->throw_error("STORESIZE method only supported for arrays");
1547 my $new_length = $_[1];
1549 my $SAVE_FILTER = $self->root->{filter_store_value};
1550 $self->root->{filter_store_value} = undef;
1552 my $result = $self->STORE('length', pack($LONG_PACK, $new_length));
1554 $self->root->{filter_store_value} = $SAVE_FILTER;
1561 # Remove and return the last element on the array
1563 my $self = _get_self($_[0]);
1564 if ($self->type ne TYPE_ARRAY) {
1565 return $self->throw_error("POP method only supported for arrays");
1567 my $length = $self->FETCHSIZE();
1570 my $content = $self->FETCH( $length - 1 );
1571 $self->DELETE( $length - 1 );
1581 # Add new element(s) to the end of the array
1583 my $self = _get_self(shift);
1584 if ($self->type ne TYPE_ARRAY) {
1585 return $self->throw_error("PUSH method only supported for arrays");
1587 my $length = $self->FETCHSIZE();
1589 while (my $content = shift @_) {
1590 $self->STORE( $length, $content );
1597 # Remove and return first element on the array.
1598 # Shift over remaining elements to take up space.
1600 my $self = _get_self($_[0]);
1601 if ($self->type ne TYPE_ARRAY) {
1602 return $self->throw_error("SHIFT method only supported for arrays");
1604 my $length = $self->FETCHSIZE();
1607 my $content = $self->FETCH( 0 );
1610 # Shift elements over and remove last one.
1612 for (my $i = 0; $i < $length - 1; $i++) {
1613 $self->STORE( $i, $self->FETCH($i + 1) );
1615 $self->DELETE( $length - 1 );
1626 # Insert new element(s) at beginning of array.
1627 # Shift over other elements to make space.
1629 my $self = _get_self($_[0]);shift @_;
1630 if ($self->type ne TYPE_ARRAY) {
1631 return $self->throw_error("UNSHIFT method only supported for arrays");
1633 my @new_elements = @_;
1634 my $length = $self->FETCHSIZE();
1635 my $new_size = scalar @new_elements;
1638 for (my $i = $length - 1; $i >= 0; $i--) {
1639 $self->STORE( $i + $new_size, $self->FETCH($i) );
1643 for (my $i = 0; $i < $new_size; $i++) {
1644 $self->STORE( $i, $new_elements[$i] );
1650 # Splices section of array with optional new section.
1651 # Returns deleted section, or last element deleted in scalar context.
1653 my $self = _get_self($_[0]);shift @_;
1654 if ($self->type ne TYPE_ARRAY) {
1655 return $self->throw_error("SPLICE method only supported for arrays");
1657 my $length = $self->FETCHSIZE();
1660 # Calculate offset and length of splice
1662 my $offset = shift || 0;
1663 if ($offset < 0) { $offset += $length; }
1666 if (scalar @_) { $splice_length = shift; }
1667 else { $splice_length = $length - $offset; }
1668 if ($splice_length < 0) { $splice_length += ($length - $offset); }
1671 # Setup array with new elements, and copy out old elements for return
1673 my @new_elements = @_;
1674 my $new_size = scalar @new_elements;
1676 my @old_elements = ();
1677 for (my $i = $offset; $i < $offset + $splice_length; $i++) {
1678 push @old_elements, $self->FETCH( $i );
1682 # Adjust array length, and shift elements to accomodate new section.
1684 if ( $new_size != $splice_length ) {
1685 if ($new_size > $splice_length) {
1686 for (my $i = $length - 1; $i >= $offset + $splice_length; $i--) {
1687 $self->STORE( $i + ($new_size - $splice_length), $self->FETCH($i) );
1691 for (my $i = $offset + $splice_length; $i < $length; $i++) {
1692 $self->STORE( $i + ($new_size - $splice_length), $self->FETCH($i) );
1694 for (my $i = 0; $i < $splice_length - $new_size; $i++) {
1695 $self->DELETE( $length - 1 );
1702 # Insert new elements into array
1704 for (my $i = $offset; $i < $offset + $new_size; $i++) {
1705 $self->STORE( $i, shift @new_elements );
1709 # Return deleted section, or last element in scalar context.
1711 return wantarray ? @old_elements : $old_elements[-1];
1714 #XXX We don't need to define it.
1715 #XXX It will be useful, though, when we split out HASH and ARRAY
1718 # Perl will call EXTEND() when the array is likely to grow.
1719 # We don't care, but include it for compatibility.
1724 # Public method aliases
1726 *put = *store = *STORE;
1727 *get = *fetch = *FETCH;
1731 *first_key = *FIRSTKEY;
1732 *next_key = *NEXTKEY;
1733 *length = *FETCHSIZE;
1737 *unshift = *UNSHIFT;
1746 DBM::Deep - A pure perl multi-level hash/array DBM
1751 my $db = DBM::Deep->new( "foo.db" );
1753 $db->{key} = 'value'; # tie() style
1756 $db->put('key', 'value'); # OO style
1757 print $db->get('key');
1759 # true multi-level support
1760 $db->{my_complex} = [
1761 'hello', { perl => 'rules' },
1766 A unique flat-file database module, written in pure perl. True
1767 multi-level hash/array support (unlike MLDBM, which is faked), hybrid
1768 OO / tie() interface, cross-platform FTPable files, and quite fast. Can
1769 handle millions of keys and unlimited hash levels without significant
1770 slow-down. Written from the ground-up in pure perl -- this is NOT a
1771 wrapper around a C-based DBM. Out-of-the-box compatibility with Unix,
1772 Mac OS X and Windows.
1776 Hopefully you are using CPAN's excellent Perl module, which will download
1777 and install the module for you. If not, get the tarball, and run these
1789 Construction can be done OO-style (which is the recommended way), or using
1790 Perl's tie() function. Both are examined here.
1792 =head2 OO CONSTRUCTION
1794 The recommended way to construct a DBM::Deep object is to use the new()
1795 method, which gets you a blessed, tied hash or array reference.
1797 my $db = DBM::Deep->new( "foo.db" );
1799 This opens a new database handle, mapped to the file "foo.db". If this
1800 file does not exist, it will automatically be created. DB files are
1801 opened in "r+" (read/write) mode, and the type of object returned is a
1802 hash, unless otherwise specified (see L<OPTIONS> below).
1806 You can pass a number of options to the constructor to specify things like
1807 locking, autoflush, etc. This is done by passing an inline hash:
1809 my $db = DBM::Deep->new(
1815 Notice that the filename is now specified I<inside> the hash with
1816 the "file" parameter, as opposed to being the sole argument to the
1817 constructor. This is required if any options are specified.
1818 See L<OPTIONS> below for the complete list.
1822 You can also start with an array instead of a hash. For this, you must
1823 specify the C<type> parameter:
1825 my $db = DBM::Deep->new(
1827 type => DBM::Deep->TYPE_ARRAY
1830 B<Note:> Specifing the C<type> parameter only takes effect when beginning
1831 a new DB file. If you create a DBM::Deep object with an existing file, the
1832 C<type> will be loaded from the file header, and ignored if it is passed
1835 =head2 TIE CONSTRUCTION
1837 Alternatively, you can create a DBM::Deep handle by using Perl's built-in
1838 tie() function. This is not ideal, because you get only a basic, tied hash
1839 (or array) which is not blessed, so you can't call any functions on it.
1842 tie %hash, "DBM::Deep", "foo.db";
1845 tie @array, "DBM::Deep", "bar.db";
1847 As with the OO constructor, you can replace the DB filename parameter with
1848 a hash containing one or more options (see L<OPTIONS> just below for the
1851 tie %hash, "DBM::Deep", {
1859 There are a number of options that can be passed in when constructing your
1860 DBM::Deep objects. These apply to both the OO- and tie- based approaches.
1866 Filename of the DB file to link the handle to. You can pass a full absolute
1867 filesystem path, partial path, or a plain filename if the file is in the
1868 current working directory. This is a required parameter.
1872 File open mode (read-only, read-write, etc.) string passed to Perl's FileHandle
1873 module. This is an optional parameter, and defaults to "r+" (read/write).
1874 B<Note:> If the default (r+) mode is selected, the file will also be auto-
1875 created if it doesn't exist.
1879 This parameter specifies what type of object to create, a hash or array. Use
1880 one of these two constants: C<DBM::Deep-E<gt>TYPE_HASH> or C<DBM::Deep-E<gt>TYPE_ARRAY>.
1881 This only takes effect when beginning a new file. This is an optional
1882 parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
1886 Specifies whether locking is to be enabled. DBM::Deep uses Perl's Fnctl flock()
1887 function to lock the database in exclusive mode for writes, and shared mode for
1888 reads. Pass any true value to enable. This affects the base DB handle I<and
1889 any child hashes or arrays> that use the same DB file. This is an optional
1890 parameter, and defaults to 0 (disabled). See L<LOCKING> below for more.
1894 Specifies whether autoflush is to be enabled on the underlying FileHandle.
1895 This obviously slows down write operations, but is required if you may have
1896 multiple processes accessing the same DB file (also consider enable I<locking>
1897 or at least I<volatile>). Pass any true value to enable. This is an optional
1898 parameter, and defaults to 0 (disabled).
1902 If I<volatile> mode is enabled, DBM::Deep will stat() the DB file before each
1903 STORE() operation. This is required if an outside force may change the size of
1904 the file between transactions. Locking also implicitly enables volatile. This
1905 is useful if you want to use a different locking system or write your own. Pass
1906 any true value to enable. This is an optional parameter, and defaults to 0
1911 If I<autobless> mode is enabled, DBM::Deep will preserve blessed hashes, and
1912 restore them when fetched. This is an B<experimental> feature, and does have
1913 side-effects. Basically, when hashes are re-blessed into their original
1914 classes, they are no longer blessed into the DBM::Deep class! So you won't be
1915 able to call any DBM::Deep methods on them. You have been warned.
1916 This is an optional parameter, and defaults to 0 (disabled).
1920 See L<FILTERS> below.
1924 Setting I<debug> mode will make all errors non-fatal, dump them out to
1925 STDERR, and continue on. This is for debugging purposes only, and probably
1926 not what you want. This is an optional parameter, and defaults to 0 (disabled).
1930 Instead of passing a file path, you can instead pass a handle to an pre-opened
1931 filehandle. Note: Beware of using the magick *DATA handle, as this actually
1932 contains your entire Perl script, as well as the data following the __DATA__
1933 marker. This will not work, because DBM::Deep uses absolute seek()s into the
1934 file. Instead, consider reading *DATA into an IO::Scalar handle, then passing
1939 =head1 TIE INTERFACE
1941 With DBM::Deep you can access your databases using Perl's standard hash/array
1942 syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can treat
1943 them as such. DBM::Deep will intercept all reads/writes and direct them to the right
1944 place -- the DB file. This has nothing to do with the L<TIE CONSTRUCTION>
1945 section above. This simply tells you how to use DBM::Deep using regular hashes
1946 and arrays, rather than calling functions like C<get()> and C<put()> (although those
1947 work too). It is entirely up to you how to want to access your databases.
1951 You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
1952 or even nested hashes (or arrays) using standard Perl syntax:
1954 my $db = DBM::Deep->new( "foo.db" );
1956 $db->{mykey} = "myvalue";
1958 $db->{myhash}->{subkey} = "subvalue";
1960 print $db->{myhash}->{subkey} . "\n";
1962 You can even step through hash keys using the normal Perl C<keys()> function:
1964 foreach my $key (keys %$db) {
1965 print "$key: " . $db->{$key} . "\n";
1968 Remember that Perl's C<keys()> function extracts I<every> key from the hash and
1969 pushes them onto an array, all before the loop even begins. If you have an
1970 extra large hash, this may exhaust Perl's memory. Instead, consider using
1971 Perl's C<each()> function, which pulls keys/values one at a time, using very
1974 while (my ($key, $value) = each %$db) {
1975 print "$key: $value\n";
1978 Please note that when using C<each()>, you should always pass a direct
1979 hash reference, not a lookup. Meaning, you should B<never> do this:
1982 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
1984 This causes an infinite loop, because for each iteration, Perl is calling
1985 FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
1986 it effectively keeps returning the first key over and over again. Instead,
1987 assign a temporary variable to C<$db->{foo}>, then pass that to each().
1991 As with hashes, you can treat any DBM::Deep object like a normal Perl array
1992 reference. This includes inserting, removing and manipulating elements,
1993 and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
1994 The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
1995 or simply be a nested array reference inside a hash. Example:
1997 my $db = DBM::Deep->new(
1998 file => "foo-array.db",
1999 type => DBM::Deep->TYPE_ARRAY
2003 push @$db, "bar", "baz";
2004 unshift @$db, "bah";
2006 my $last_elem = pop @$db; # baz
2007 my $first_elem = shift @$db; # bah
2008 my $second_elem = $db->[1]; # bar
2010 my $num_elements = scalar @$db;
2014 In addition to the I<tie()> interface, you can also use a standard OO interface
2015 to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
2016 array) has its own methods, but both types share the following common methods:
2017 C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>.
2023 Stores a new hash key/value pair, or sets an array element value. Takes two
2024 arguments, the hash key or array index, and the new value. The value can be
2025 a scalar, hash ref or array ref. Returns true on success, false on failure.
2027 $db->put("foo", "bar"); # for hashes
2028 $db->put(1, "bar"); # for arrays
2032 Fetches the value of a hash key or array element. Takes one argument: the hash
2033 key or array index. Returns a scalar, hash ref or array ref, depending on the
2036 my $value = $db->get("foo"); # for hashes
2037 my $value = $db->get(1); # for arrays
2041 Checks if a hash key or array index exists. Takes one argument: the hash key
2042 or array index. Returns true if it exists, false if not.
2044 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
2045 if ($db->exists(1)) { print "yay!\n"; } # for arrays
2049 Deletes one hash key/value pair or array element. Takes one argument: the hash
2050 key or array index. Returns true on success, false if not found. For arrays,
2051 the remaining elements located after the deleted element are NOT moved over.
2052 The deleted element is essentially just undefined, which is exactly how Perl's
2053 internal arrays work. Please note that the space occupied by the deleted
2054 key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY>
2055 below for details and workarounds.
2057 $db->delete("foo"); # for hashes
2058 $db->delete(1); # for arrays
2062 Deletes B<all> hash keys or array elements. Takes no arguments. No return
2063 value. Please note that the space occupied by the deleted keys/values or
2064 elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for
2065 details and workarounds.
2067 $db->clear(); # hashes or arrays
2073 For hashes, DBM::Deep supports all the common methods described above, and the
2074 following additional methods: C<first_key()> and C<next_key()>.
2080 Returns the "first" key in the hash. As with built-in Perl hashes, keys are
2081 fetched in an undefined order (which appears random). Takes no arguments,
2082 returns the key as a scalar value.
2084 my $key = $db->first_key();
2088 Returns the "next" key in the hash, given the previous one as the sole argument.
2089 Returns undef if there are no more keys to be fetched.
2091 $key = $db->next_key($key);
2095 Here are some examples of using hashes:
2097 my $db = DBM::Deep->new( "foo.db" );
2099 $db->put("foo", "bar");
2100 print "foo: " . $db->get("foo") . "\n";
2102 $db->put("baz", {}); # new child hash ref
2103 $db->get("baz")->put("buz", "biz");
2104 print "buz: " . $db->get("baz")->get("buz") . "\n";
2106 my $key = $db->first_key();
2108 print "$key: " . $db->get($key) . "\n";
2109 $key = $db->next_key($key);
2112 if ($db->exists("foo")) { $db->delete("foo"); }
2116 For arrays, DBM::Deep supports all the common methods described above, and the
2117 following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
2118 C<unshift()> and C<splice()>.
2124 Returns the number of elements in the array. Takes no arguments.
2126 my $len = $db->length();
2130 Adds one or more elements onto the end of the array. Accepts scalars, hash
2131 refs or array refs. No return value.
2133 $db->push("foo", "bar", {});
2137 Fetches the last element in the array, and deletes it. Takes no arguments.
2138 Returns undef if array is empty. Returns the element value.
2140 my $elem = $db->pop();
2144 Fetches the first element in the array, deletes it, then shifts all the
2145 remaining elements over to take up the space. Returns the element value. This
2146 method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
2149 my $elem = $db->shift();
2153 Inserts one or more elements onto the beginning of the array, shifting all
2154 existing elements over to make room. Accepts scalars, hash refs or array refs.
2155 No return value. This method is not recommended with large arrays -- see
2156 <LARGE ARRAYS> below for details.
2158 $db->unshift("foo", "bar", {});
2162 Performs exactly like Perl's built-in function of the same name. See L<perldoc
2163 -f splice> for usage -- it is too complicated to document here. This method is
2164 not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
2168 Here are some examples of using arrays:
2170 my $db = DBM::Deep->new(
2172 type => DBM::Deep->TYPE_ARRAY
2175 $db->push("bar", "baz");
2176 $db->unshift("foo");
2179 my $len = $db->length();
2180 print "length: $len\n"; # 4
2182 for (my $k=0; $k<$len; $k++) {
2183 print "$k: " . $db->get($k) . "\n";
2186 $db->splice(1, 2, "biz", "baf");
2188 while (my $elem = shift @$db) {
2189 print "shifted: $elem\n";
2194 Enable automatic file locking by passing a true value to the C<locking>
2195 parameter when constructing your DBM::Deep object (see L<SETUP> above).
2197 my $db = DBM::Deep->new(
2202 This causes DBM::Deep to C<flock()> the underlying FileHandle object with exclusive
2203 mode for writes, and shared mode for reads. This is required if you have
2204 multiple processes accessing the same database file, to avoid file corruption.
2205 Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
2206 NFS> below for more.
2208 =head2 EXPLICIT LOCKING
2210 You can explicitly lock a database, so it remains locked for multiple
2211 transactions. This is done by calling the C<lock()> method, and passing an
2212 optional lock mode argument (defaults to exclusive mode). This is particularly
2213 useful for things like counters, where the current value needs to be fetched,
2214 then incremented, then stored again.
2217 my $counter = $db->get("counter");
2219 $db->put("counter", $counter);
2228 You can pass C<lock()> an optional argument, which specifies which mode to use
2229 (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX>
2230 or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the
2231 same as the constants defined in Perl's C<Fcntl> module.
2233 $db->lock( DBM::Deep->LOCK_SH );
2237 If you want to implement your own file locking scheme, be sure to create your
2238 DBM::Deep objects setting the C<volatile> option to true. This hints to DBM::Deep
2239 that the DB file may change between transactions. See L<LOW-LEVEL ACCESS>
2242 =head1 IMPORTING/EXPORTING
2244 You can import existing complex structures by calling the C<import()> method,
2245 and export an entire database into an in-memory structure using the C<export()>
2246 method. Both are examined here.
2250 Say you have an existing hash with nested hashes/arrays inside it. Instead of
2251 walking the structure and adding keys/elements to the database as you go,
2252 simply pass a reference to the C<import()> method. This recursively adds
2253 everything to an existing DBM::Deep object for you. Here is an example:
2258 array1 => [ "elem0", "elem1", "elem2" ],
2260 subkey1 => "subvalue1",
2261 subkey2 => "subvalue2"
2265 my $db = DBM::Deep->new( "foo.db" );
2266 $db->import( $struct );
2268 print $db->{key1} . "\n"; # prints "value1"
2270 This recursively imports the entire C<$struct> object into C<$db>, including
2271 all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
2272 keys are merged with the existing ones, replacing if they already exist.
2273 The C<import()> method can be called on any database level (not just the base
2274 level), and works with both hash and array DB types.
2278 B<Note:> Make sure your existing structure has no circular references in it.
2279 These will cause an infinite loop when importing.
2283 Calling the C<export()> method on an existing DBM::Deep object will return
2284 a reference to a new in-memory copy of the database. The export is done
2285 recursively, so all nested hashes/arrays are all exported to standard Perl
2286 objects. Here is an example:
2288 my $db = DBM::Deep->new( "foo.db" );
2290 $db->{key1} = "value1";
2291 $db->{key2} = "value2";
2293 $db->{hash1}->{subkey1} = "subvalue1";
2294 $db->{hash1}->{subkey2} = "subvalue2";
2296 my $struct = $db->export();
2298 print $struct->{key1} . "\n"; # prints "value1"
2300 This makes a complete copy of the database in memory, and returns a reference
2301 to it. The C<export()> method can be called on any database level (not just
2302 the base level), and works with both hash and array DB types. Be careful of
2303 large databases -- you can store a lot more data in a DBM::Deep object than an
2304 in-memory Perl structure.
2308 B<Note:> Make sure your database has no circular references in it.
2309 These will cause an infinite loop when exporting.
2313 DBM::Deep has a number of hooks where you can specify your own Perl function
2314 to perform filtering on incoming or outgoing data. This is a perfect
2315 way to extend the engine, and implement things like real-time compression or
2316 encryption. Filtering applies to the base DB level, and all child hashes /
2317 arrays. Filter hooks can be specified when your DBM::Deep object is first
2318 constructed, or by calling the C<set_filter()> method at any time. There are
2319 four available filter hooks, described below:
2323 =item * filter_store_key
2325 This filter is called whenever a hash key is stored. It
2326 is passed the incoming key, and expected to return a transformed key.
2328 =item * filter_store_value
2330 This filter is called whenever a hash key or array element is stored. It
2331 is passed the incoming value, and expected to return a transformed value.
2333 =item * filter_fetch_key
2335 This filter is called whenever a hash key is fetched (i.e. via
2336 C<first_key()> or C<next_key()>). It is passed the transformed key,
2337 and expected to return the plain key.
2339 =item * filter_fetch_value
2341 This filter is called whenever a hash key or array element is fetched.
2342 It is passed the transformed value, and expected to return the plain value.
2346 Here are the two ways to setup a filter hook:
2348 my $db = DBM::Deep->new(
2350 filter_store_value => \&my_filter_store,
2351 filter_fetch_value => \&my_filter_fetch
2356 $db->set_filter( "filter_store_value", \&my_filter_store );
2357 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
2359 Your filter function will be called only when dealing with SCALAR keys or
2360 values. When nested hashes and arrays are being stored/fetched, filtering
2361 is bypassed. Filters are called as static functions, passed a single SCALAR
2362 argument, and expected to return a single SCALAR value. If you want to
2363 remove a filter, set the function reference to C<undef>:
2365 $db->set_filter( "filter_store_value", undef );
2367 =head2 REAL-TIME ENCRYPTION EXAMPLE
2369 Here is a working example that uses the I<Crypt::Blowfish> module to
2370 do real-time encryption / decryption of keys & values with DBM::Deep Filters.
2371 Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more
2372 on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module.
2375 use Crypt::Blowfish;
2378 my $cipher = Crypt::CBC->new({
2379 'key' => 'my secret key',
2380 'cipher' => 'Blowfish',
2382 'regenerate_key' => 0,
2383 'padding' => 'space',
2387 my $db = DBM::Deep->new(
2388 file => "foo-encrypt.db",
2389 filter_store_key => \&my_encrypt,
2390 filter_store_value => \&my_encrypt,
2391 filter_fetch_key => \&my_decrypt,
2392 filter_fetch_value => \&my_decrypt,
2395 $db->{key1} = "value1";
2396 $db->{key2} = "value2";
2397 print "key1: " . $db->{key1} . "\n";
2398 print "key2: " . $db->{key2} . "\n";
2404 return $cipher->encrypt( $_[0] );
2407 return $cipher->decrypt( $_[0] );
2410 =head2 REAL-TIME COMPRESSION EXAMPLE
2412 Here is a working example that uses the I<Compress::Zlib> module to do real-time
2413 compression / decompression of keys & values with DBM::Deep Filters.
2414 Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for
2415 more on I<Compress::Zlib>.
2420 my $db = DBM::Deep->new(
2421 file => "foo-compress.db",
2422 filter_store_key => \&my_compress,
2423 filter_store_value => \&my_compress,
2424 filter_fetch_key => \&my_decompress,
2425 filter_fetch_value => \&my_decompress,
2428 $db->{key1} = "value1";
2429 $db->{key2} = "value2";
2430 print "key1: " . $db->{key1} . "\n";
2431 print "key2: " . $db->{key2} . "\n";
2437 return Compress::Zlib::memGzip( $_[0] ) ;
2440 return Compress::Zlib::memGunzip( $_[0] ) ;
2443 B<Note:> Filtering of keys only applies to hashes. Array "keys" are
2444 actually numerical index numbers, and are not filtered.
2446 =head1 ERROR HANDLING
2448 Most DBM::Deep methods return a true value for success, and call die() on
2449 failure. You can wrap calls in an eval block to catch the die. Also, the
2450 actual error message is stored in an internal scalar, which can be fetched by
2451 calling the C<error()> method.
2453 my $db = DBM::Deep->new( "foo.db" ); # create hash
2454 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
2456 print $db->error(); # prints error message
2458 You can then call C<clear_error()> to clear the current error state.
2462 If you set the C<debug> option to true when creating your DBM::Deep object,
2463 all errors are considered NON-FATAL, and dumped to STDERR. This is only
2464 for debugging purposes.
2466 =head1 LARGEFILE SUPPORT
2468 If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
2469 and 64-bit support, you I<may> be able to create databases larger than 2 GB.
2470 DBM::Deep by default uses 32-bit file offset tags, but these can be changed
2471 by calling the static C<set_pack()> method before you do anything else.
2473 DBM::Deep::set_pack(8, 'Q');
2475 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
2476 instead of 32-bit longs. After setting these values your DB files have a
2477 theoretical maximum size of 16 XB (exabytes).
2481 B<Note:> Changing these values will B<NOT> work for existing database files.
2482 Only change this for new files, and make sure it stays set consistently
2483 throughout the file's life. If you do set these values, you can no longer
2484 access 32-bit DB files. You can, however, call C<set_pack(4, 'N')> to change
2485 back to 32-bit mode.
2489 B<Note:> I have not personally tested files > 2 GB -- all my systems have
2490 only a 32-bit Perl. However, I have received user reports that this does
2493 =head1 LOW-LEVEL ACCESS
2495 If you require low-level access to the underlying FileHandle that DBM::Deep uses,
2496 you can call the C<fh()> method, which returns the handle:
2500 This method can be called on the root level of the datbase, or any child
2501 hashes or arrays. All levels share a I<root> structure, which contains things
2502 like the FileHandle, a reference counter, and all your options you specified
2503 when you created the object. You can get access to this root structure by
2504 calling the C<root()> method.
2506 my $root = $db->root();
2508 This is useful for changing options after the object has already been created,
2509 such as enabling/disabling locking, volatile or debug modes. You can also
2510 store your own temporary user data in this structure (be wary of name
2511 collision), which is then accessible from any child hash or array.
2513 =head1 CUSTOM DIGEST ALGORITHM
2515 DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing
2516 keys. However you can override this, and use another algorithm (such as SHA-256)
2517 or even write your own. But please note that DBM::Deep currently expects zero
2518 collisions, so your algorithm has to be I<perfect>, so to speak.
2519 Collision detection may be introduced in a later version.
2523 You can specify a custom digest algorithm by calling the static C<set_digest()>
2524 function, passing a reference to a subroutine, and the length of the algorithm's
2525 hashes (in bytes). This is a global static function, which affects ALL DBM::Deep
2526 objects. Here is a working example that uses a 256-bit hash from the
2527 I<Digest::SHA256> module. Please see
2528 L<http://search.cpan.org/search?module=Digest::SHA256> for more.
2533 my $context = Digest::SHA256::new(256);
2535 DBM::Deep::set_digest( \&my_digest, 32 );
2537 my $db = DBM::Deep->new( "foo-sha.db" );
2539 $db->{key1} = "value1";
2540 $db->{key2} = "value2";
2541 print "key1: " . $db->{key1} . "\n";
2542 print "key2: " . $db->{key2} . "\n";
2548 return substr( $context->hash($_[0]), 0, 32 );
2551 B<Note:> Your returned digest strings must be B<EXACTLY> the number
2552 of bytes you specify in the C<set_digest()> function (in this case 32).
2554 =head1 CIRCULAR REFERENCES
2556 DBM::Deep has B<experimental> support for circular references. Meaning you
2557 can have a nested hash key or array element that points to a parent object.
2558 This relationship is stored in the DB file, and is preserved between sessions.
2561 my $db = DBM::Deep->new( "foo.db" );
2564 $db->{circle} = $db; # ref to self
2566 print $db->{foo} . "\n"; # prints "foo"
2567 print $db->{circle}->{foo} . "\n"; # prints "foo" again
2569 One catch is, passing the object to a function that recursively walks the
2570 object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
2571 C<export()> methods) will result in an infinite loop. The other catch is,
2572 if you fetch the I<key> of a circular reference (i.e. using the C<first_key()>
2573 or C<next_key()> methods), you will get the I<target object's key>, not the
2574 ref's key. This gets even more interesting with the above example, where
2575 the I<circle> key points to the base DB object, which technically doesn't
2576 have a key. So I made DBM::Deep return "[base]" as the key name in that
2579 =head1 CAVEATS / ISSUES / BUGS
2581 This section describes all the known issues with DBM::Deep. It you have found
2582 something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>.
2584 =head2 UNUSED SPACE RECOVERY
2586 One major caveat with DBM::Deep is that space occupied by existing keys and
2587 values is not recovered when they are deleted. Meaning if you keep deleting
2588 and adding new keys, your file will continuously grow. I am working on this,
2589 but in the meantime you can call the built-in C<optimize()> method from time to
2590 time (perhaps in a crontab or something) to recover all your unused space.
2592 $db->optimize(); # returns true on success
2594 This rebuilds the ENTIRE database into a new file, then moves it on top of
2595 the original. The new file will have no unused space, thus it will take up as
2596 little disk space as possible. Please note that this operation can take
2597 a long time for large files, and you need enough disk space to temporarily hold
2598 2 copies of your DB file. The temporary file is created in the same directory
2599 as the original, named with a ".tmp" extension, and is deleted when the
2600 operation completes. Oh, and if locking is enabled, the DB is automatically
2601 locked for the entire duration of the copy.
2605 B<WARNING:> Only call optimize() on the top-level node of the database, and
2606 make sure there are no child references lying around. DBM::Deep keeps a reference
2607 counter, and if it is greater than 1, optimize() will abort and return undef.
2609 =head2 AUTOVIVIFICATION
2611 Unfortunately, autovivification doesn't work with tied hashes. This appears to
2612 be a bug in Perl's tie() system, as I<Jakob Schmidt> encountered the very same
2613 issue with his I<DWH_FIle> module (see L<http://search.cpan.org/search?module=DWH_File>),
2614 and it is also mentioned in the BUGS section for the I<MLDBM> module <see
2615 L<http://search.cpan.org/search?module=MLDBM>). Basically, on a new db file,
2618 $db->{foo}->{bar} = "hello";
2620 Since "foo" doesn't exist, you cannot add "bar" to it. You end up with "foo"
2621 being an empty hash. Try this instead, which works fine:
2623 $db->{foo} = { bar => "hello" };
2625 As of Perl 5.8.7, this bug still exists. I have walked very carefully through
2626 the execution path, and Perl indeed passes an empty hash to the STORE() method.
2627 Probably a bug in Perl.
2629 =head2 FILE CORRUPTION
2631 The current level of error handling in DBM::Deep is minimal. Files I<are> checked
2632 for a 32-bit signature when opened, but other corruption in files can cause
2633 segmentation faults. DBM::Deep may try to seek() past the end of a file, or get
2634 stuck in an infinite loop depending on the level of corruption. File write
2635 operations are not checked for failure (for speed), so if you happen to run
2636 out of disk space, DBM::Deep will probably fail in a bad way. These things will
2637 be addressed in a later version of DBM::Deep.
2641 Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local
2642 filesystems, but will NOT protect you from file corruption over NFS. I've heard
2643 about setting up your NFS server with a locking daemon, then using lockf() to
2644 lock your files, but your milage may vary there as well. From what I
2645 understand, there is no real way to do it. However, if you need access to the
2646 underlying FileHandle in DBM::Deep for using some other kind of locking scheme like
2647 lockf(), see the L<LOW-LEVEL ACCESS> section above.
2649 =head2 COPYING OBJECTS
2651 Beware of copying tied objects in Perl. Very strange things can happen.
2652 Instead, use DBM::Deep's C<clone()> method which safely copies the object and
2653 returns a new, blessed, tied hash or array to the same level in the DB.
2655 my $copy = $db->clone();
2659 Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
2660 These functions cause every element in the array to move, which can be murder
2661 on DBM::Deep, as every element has to be fetched from disk, then stored again in
2662 a different location. This may be addressed in a later version.
2666 This section discusses DBM::Deep's speed and memory usage.
2670 Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as
2671 the almighty I<BerkeleyDB>. But it makes up for it in features like true
2672 multi-level hash/array support, and cross-platform FTPable files. Even so,
2673 DBM::Deep is still pretty fast, and the speed stays fairly consistent, even
2674 with huge databases. Here is some test data:
2676 Adding 1,000,000 keys to new DB file...
2678 At 100 keys, avg. speed is 2,703 keys/sec
2679 At 200 keys, avg. speed is 2,642 keys/sec
2680 At 300 keys, avg. speed is 2,598 keys/sec
2681 At 400 keys, avg. speed is 2,578 keys/sec
2682 At 500 keys, avg. speed is 2,722 keys/sec
2683 At 600 keys, avg. speed is 2,628 keys/sec
2684 At 700 keys, avg. speed is 2,700 keys/sec
2685 At 800 keys, avg. speed is 2,607 keys/sec
2686 At 900 keys, avg. speed is 2,190 keys/sec
2687 At 1,000 keys, avg. speed is 2,570 keys/sec
2688 At 2,000 keys, avg. speed is 2,417 keys/sec
2689 At 3,000 keys, avg. speed is 1,982 keys/sec
2690 At 4,000 keys, avg. speed is 1,568 keys/sec
2691 At 5,000 keys, avg. speed is 1,533 keys/sec
2692 At 6,000 keys, avg. speed is 1,787 keys/sec
2693 At 7,000 keys, avg. speed is 1,977 keys/sec
2694 At 8,000 keys, avg. speed is 2,028 keys/sec
2695 At 9,000 keys, avg. speed is 2,077 keys/sec
2696 At 10,000 keys, avg. speed is 2,031 keys/sec
2697 At 20,000 keys, avg. speed is 1,970 keys/sec
2698 At 30,000 keys, avg. speed is 2,050 keys/sec
2699 At 40,000 keys, avg. speed is 2,073 keys/sec
2700 At 50,000 keys, avg. speed is 1,973 keys/sec
2701 At 60,000 keys, avg. speed is 1,914 keys/sec
2702 At 70,000 keys, avg. speed is 2,091 keys/sec
2703 At 80,000 keys, avg. speed is 2,103 keys/sec
2704 At 90,000 keys, avg. speed is 1,886 keys/sec
2705 At 100,000 keys, avg. speed is 1,970 keys/sec
2706 At 200,000 keys, avg. speed is 2,053 keys/sec
2707 At 300,000 keys, avg. speed is 1,697 keys/sec
2708 At 400,000 keys, avg. speed is 1,838 keys/sec
2709 At 500,000 keys, avg. speed is 1,941 keys/sec
2710 At 600,000 keys, avg. speed is 1,930 keys/sec
2711 At 700,000 keys, avg. speed is 1,735 keys/sec
2712 At 800,000 keys, avg. speed is 1,795 keys/sec
2713 At 900,000 keys, avg. speed is 1,221 keys/sec
2714 At 1,000,000 keys, avg. speed is 1,077 keys/sec
2716 This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl
2717 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and
2718 values were between 6 - 12 chars in length. The DB file ended up at 210MB.
2719 Run time was 12 min 3 sec.
2723 One of the great things about DBM::Deep is that it uses very little memory.
2724 Even with huge databases (1,000,000+ keys) you will not see much increased
2725 memory on your process. DBM::Deep relies solely on the filesystem for storing
2726 and fetching data. Here is output from I</usr/bin/top> before even opening a
2729 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
2730 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl
2732 Basically the process is taking 2,716K of memory. And here is the same
2733 process after storing and fetching 1,000,000 keys:
2735 PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
2736 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl
2738 Notice the memory usage increased by only 56K. Test was performed on a 700mHz
2739 x86 box running Linux RedHat 7.2 & Perl 5.6.1.
2741 =head1 DB FILE FORMAT
2743 In case you were interested in the underlying DB file format, it is documented
2744 here in this section. You don't need to know this to use the module, it's just
2745 included for reference.
2749 DBM::Deep files always start with a 32-bit signature to identify the file type.
2750 This is at offset 0. The signature is "DPDB" in network byte order. This is
2751 checked when the file is opened.
2755 The DBM::Deep file is in a I<tagged format>, meaning each section of the file
2756 has a standard header containing the type of data, the length of data, and then
2757 the data itself. The type is a single character (1 byte), the length is a
2758 32-bit unsigned long in network byte order, and the data is, well, the data.
2759 Here is how it unfolds:
2763 Immediately after the 32-bit file signature is the I<Master Index> record.
2764 This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048
2765 bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array,
2766 depending on how the DBM::Deep object was constructed.
2770 The index works by looking at a I<MD5 Hash> of the hash key (or array index
2771 number). The first 8-bit char of the MD5 signature is the offset into the
2772 index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the
2773 index element is a file offset of the next tag for the key/element in question,
2774 which is usually a I<Bucket List> tag (see below).
2778 The next tag I<could> be another index, depending on how many keys/elements
2779 exist. See L<RE-INDEXING> below for details.
2783 A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus
2784 file offsets to where the actual data is stored. It starts with a standard
2785 tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or
2786 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus
2787 the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data.
2788 When the list fills up, a I<Re-Index> operation is performed (See
2789 L<RE-INDEXING> below).
2793 A I<Bucket> is a tag containing a key/value pair (in hash mode), or a
2794 index/value pair (in array mode). It starts with a standard tag header with
2795 type I<D> for scalar data (string, binary, etc.), or it could be a nested
2796 hash (type I<H>) or array (type I<A>). The value comes just after the tag
2797 header. The size reported in the tag header is only for the value, but then,
2798 just after the value is another size (32-bit unsigned long) and then the plain
2799 key itself. Since the value is likely to be fetched more often than the plain
2800 key, I figured it would be I<slightly> faster to store the value first.
2804 If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index>
2805 record for the nested structure, where the process begins all over again.
2809 After a I<Bucket List> grows to 16 records, its allocated space in the file is
2810 exhausted. Then, when another key/element comes in, the list is converted to a
2811 new index record. However, this index will look at the next char in the MD5
2812 hash, and arrange new Bucket List pointers accordingly. This process is called
2813 I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all
2814 17 (16 + new one) keys/elements are removed from the old Bucket List and
2815 inserted into the new index. Several new Bucket Lists are created in the
2816 process, as a new MD5 char from the key is being examined (it is unlikely that
2817 the keys will all share the same next char of their MD5s).
2821 Because of the way the I<MD5> algorithm works, it is impossible to tell exactly
2822 when the Bucket Lists will turn into indexes, but the first round tends to
2823 happen right around 4,000 keys. You will see a I<slight> decrease in
2824 performance here, but it picks back up pretty quick (see L<SPEED> above). Then
2825 it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's
2826 right around 900,000 keys. This process can continue nearly indefinitely --
2827 right up until the point the I<MD5> signatures start colliding with each other,
2828 and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND
2829 getting struck by lightning while you are walking to cash in your tickets.
2830 Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to
2831 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe
2832 this is 340 unodecillion, but don't quote me).
2836 When a new key/element is stored, the key (or index number) is first ran through
2837 I<Digest::MD5> to get a 128-bit signature (example, in hex:
2838 b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked
2839 for the first char of the signature (in this case I<b>). If it does not exist,
2840 a new I<Bucket List> is created for our key (and the next 15 future keys that
2841 happen to also have I<b> as their first MD5 char). The entire MD5 is written
2842 to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at
2843 this point, unless we are replacing an existing I<Bucket>), where the actual
2844 data will be stored.
2848 Fetching an existing key/element involves getting a I<Digest::MD5> of the key
2849 (or index number), then walking along the indexes. If there are enough
2850 keys/elements in this DB level, there might be nested indexes, each linked to
2851 a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which
2852 contains up to 16 full MD5 hashes. Each is checked for equality to the key in
2853 question. If we found a match, the I<Bucket> tag is loaded, where the value and
2854 plain key are stored.
2858 Fetching the plain key occurs when calling the I<first_key()> and I<next_key()>
2859 methods. In this process the indexes are walked systematically, and each key
2860 fetched in increasing MD5 order (which is why it appears random). Once the
2861 I<Bucket> is found, the value is skipped the plain key returned instead.
2862 B<Note:> Do not count on keys being fetched as if the MD5 hashes were
2863 alphabetically sorted. This only happens on an index-level -- as soon as the
2864 I<Bucket Lists> are hit, the keys will come out in the order they went in --
2865 so it's pretty much undefined how the keys will come out -- just like Perl's
2870 Joseph Huckaby, L<jhuckaby@cpan.org>
2872 Special thanks to Adam Sah and Rich Gaushell! You know why :-)
2876 perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
2877 Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
2881 Copyright (c) 2002-2005 Joseph Huckaby. All Rights Reserved.
2882 This is free software, you may use it and distribute it under the
2883 same terms as Perl itself.