Commit | Line | Data |
ffed8b01 |
1 | package DBM::Deep; |
2 | |
3 | ## |
4 | # DBM::Deep |
5 | # |
6 | # Description: |
d0b74c17 |
7 | # Multi-level database module for storing hash trees, arrays and simple |
8 | # key/value pairs into FTP-able, cross-platform binary database files. |
ffed8b01 |
9 | # |
d0b74c17 |
10 | # Type `perldoc DBM::Deep` for complete documentation. |
ffed8b01 |
11 | # |
12 | # Usage Examples: |
d0b74c17 |
13 | # my %db; |
14 | # tie %db, 'DBM::Deep', 'my_database.db'; # standard tie() method |
ffed8b01 |
15 | # |
d0b74c17 |
16 | # my $db = new DBM::Deep( 'my_database.db' ); # preferred OO method |
17 | # |
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"; |
ffed8b01 |
25 | # |
26 | # Copyright: |
d0b74c17 |
27 | # (c) 2002-2006 Joseph Huckaby. All Rights Reserved. |
28 | # This program is free software; you can redistribute it and/or |
29 | # modify it under the same terms as Perl itself. |
ffed8b01 |
30 | ## |
31 | |
460b1067 |
32 | use 5.6.0; |
33 | |
ffed8b01 |
34 | use strict; |
460b1067 |
35 | use warnings; |
8b957036 |
36 | |
86867f3a |
37 | our $VERSION = q(0.99_01); |
38 | |
596e9574 |
39 | use Fcntl qw( :DEFAULT :flock :seek ); |
ffed8b01 |
40 | use Digest::MD5 (); |
41 | use Scalar::Util (); |
ffed8b01 |
42 | |
95967a5e |
43 | use DBM::Deep::Engine; |
460b1067 |
44 | use DBM::Deep::File; |
95967a5e |
45 | |
ffed8b01 |
46 | ## |
47 | # Setup constants for users to pass to new() |
48 | ## |
86867f3a |
49 | sub TYPE_HASH () { DBM::Deep::Engine->SIG_HASH } |
50 | sub TYPE_ARRAY () { DBM::Deep::Engine->SIG_ARRAY } |
ffed8b01 |
51 | |
0ca7ea98 |
52 | sub _get_args { |
53 | my $proto = shift; |
54 | |
55 | my $args; |
56 | if (scalar(@_) > 1) { |
57 | if ( @_ % 2 ) { |
58 | $proto->_throw_error( "Odd number of parameters to " . (caller(1))[2] ); |
59 | } |
60 | $args = {@_}; |
61 | } |
d0b74c17 |
62 | elsif ( ref $_[0] ) { |
4d35d856 |
63 | unless ( eval { local $SIG{'__DIE__'}; %{$_[0]} || 1 } ) { |
0ca7ea98 |
64 | $proto->_throw_error( "Not a hashref in args to " . (caller(1))[2] ); |
65 | } |
66 | $args = $_[0]; |
67 | } |
d0b74c17 |
68 | else { |
0ca7ea98 |
69 | $args = { file => shift }; |
70 | } |
71 | |
72 | return $args; |
73 | } |
74 | |
ffed8b01 |
75 | sub new { |
d0b74c17 |
76 | ## |
77 | # Class constructor method for Perl OO interface. |
78 | # Calls tie() and returns blessed reference to tied hash or array, |
79 | # providing a hybrid OO/tie interface. |
80 | ## |
81 | my $class = shift; |
82 | my $args = $class->_get_args( @_ ); |
83 | |
84 | ## |
85 | # Check if we want a tied hash or array. |
86 | ## |
87 | my $self; |
88 | if (defined($args->{type}) && $args->{type} eq TYPE_ARRAY) { |
6fe26b29 |
89 | $class = 'DBM::Deep::Array'; |
90 | require DBM::Deep::Array; |
d0b74c17 |
91 | tie @$self, $class, %$args; |
92 | } |
93 | else { |
6fe26b29 |
94 | $class = 'DBM::Deep::Hash'; |
95 | require DBM::Deep::Hash; |
d0b74c17 |
96 | tie %$self, $class, %$args; |
97 | } |
ffed8b01 |
98 | |
d0b74c17 |
99 | return bless $self, $class; |
ffed8b01 |
100 | } |
101 | |
96041a25 |
102 | # This initializer is called from the various TIE* methods. new() calls tie(), |
103 | # which allows for a single point of entry. |
0795f290 |
104 | sub _init { |
0795f290 |
105 | my $class = shift; |
994ccd8e |
106 | my ($args) = @_; |
0795f290 |
107 | |
460b1067 |
108 | $args->{fileobj} = DBM::Deep::File->new( $args ) |
109 | unless exists $args->{fileobj}; |
110 | |
111 | # locking implicitly enables autoflush |
112 | if ($args->{locking}) { $args->{autoflush} = 1; } |
113 | |
0795f290 |
114 | # These are the defaults to be optionally overridden below |
115 | my $self = bless { |
95967a5e |
116 | type => TYPE_HASH, |
e06824f8 |
117 | base_offset => undef, |
359a01ac |
118 | |
119 | parent => undef, |
120 | parent_key => undef, |
121 | |
460b1067 |
122 | fileobj => undef, |
0795f290 |
123 | }, $class; |
359a01ac |
124 | $self->{engine} = DBM::Deep::Engine->new( { %{$args}, obj => $self } ); |
8db25060 |
125 | |
fde3db1a |
126 | # Grab the parameters we want to use |
0795f290 |
127 | foreach my $param ( keys %$self ) { |
128 | next unless exists $args->{$param}; |
3e9498a1 |
129 | $self->{$param} = $args->{$param}; |
ffed8b01 |
130 | } |
d0b74c17 |
131 | |
15d754dd |
132 | $self->_engine->setup_fh( $self ); |
0795f290 |
133 | |
359a01ac |
134 | $self->{fileobj}->set_db( $self ); |
135 | |
0795f290 |
136 | return $self; |
ffed8b01 |
137 | } |
138 | |
ffed8b01 |
139 | sub TIEHASH { |
6fe26b29 |
140 | shift; |
141 | require DBM::Deep::Hash; |
142 | return DBM::Deep::Hash->TIEHASH( @_ ); |
ffed8b01 |
143 | } |
144 | |
145 | sub TIEARRAY { |
6fe26b29 |
146 | shift; |
147 | require DBM::Deep::Array; |
148 | return DBM::Deep::Array->TIEARRAY( @_ ); |
ffed8b01 |
149 | } |
150 | |
ffed8b01 |
151 | sub lock { |
994ccd8e |
152 | my $self = shift->_get_self; |
15ba72cc |
153 | return $self->_fileobj->lock( $self, @_ ); |
ffed8b01 |
154 | } |
155 | |
156 | sub unlock { |
994ccd8e |
157 | my $self = shift->_get_self; |
15ba72cc |
158 | return $self->_fileobj->unlock( $self, @_ ); |
ffed8b01 |
159 | } |
160 | |
906c8e01 |
161 | sub _copy_value { |
162 | my $self = shift->_get_self; |
163 | my ($spot, $value) = @_; |
164 | |
165 | if ( !ref $value ) { |
166 | ${$spot} = $value; |
167 | } |
168 | elsif ( eval { local $SIG{__DIE__}; $value->isa( 'DBM::Deep' ) } ) { |
f9c33187 |
169 | ${$spot} = $value->_repr; |
906c8e01 |
170 | $value->_copy_node( ${$spot} ); |
171 | } |
172 | else { |
173 | my $r = Scalar::Util::reftype( $value ); |
174 | my $c = Scalar::Util::blessed( $value ); |
175 | if ( $r eq 'ARRAY' ) { |
176 | ${$spot} = [ @{$value} ]; |
177 | } |
178 | else { |
179 | ${$spot} = { %{$value} }; |
180 | } |
95bbd935 |
181 | ${$spot} = bless ${$spot}, $c |
906c8e01 |
182 | if defined $c; |
183 | } |
184 | |
185 | return 1; |
186 | } |
187 | |
261d1296 |
188 | sub _copy_node { |
f9c33187 |
189 | die "Must be implemented in a child class\n"; |
190 | } |
906c8e01 |
191 | |
f9c33187 |
192 | sub _repr { |
193 | die "Must be implemented in a child class\n"; |
ffed8b01 |
194 | } |
195 | |
196 | sub export { |
d0b74c17 |
197 | ## |
198 | # Recursively export into standard Perl hashes and arrays. |
199 | ## |
994ccd8e |
200 | my $self = shift->_get_self; |
d0b74c17 |
201 | |
f9c33187 |
202 | my $temp = $self->_repr; |
d0b74c17 |
203 | |
204 | $self->lock(); |
205 | $self->_copy_node( $temp ); |
206 | $self->unlock(); |
207 | |
208 | return $temp; |
ffed8b01 |
209 | } |
210 | |
211 | sub import { |
d0b74c17 |
212 | ## |
213 | # Recursively import Perl hash/array structure |
214 | ## |
d0b74c17 |
215 | if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore |
216 | |
994ccd8e |
217 | my $self = shift->_get_self; |
218 | my ($struct) = @_; |
d0b74c17 |
219 | |
c9cec40e |
220 | # struct is not a reference, so just import based on our type |
d0b74c17 |
221 | if (!ref($struct)) { |
f9c33187 |
222 | $struct = $self->_repr( @_ ); |
d0b74c17 |
223 | } |
224 | |
f9c33187 |
225 | return $self->_import( $struct ); |
ffed8b01 |
226 | } |
227 | |
228 | sub optimize { |
d0b74c17 |
229 | ## |
230 | # Rebuild entire database into new file, then move |
231 | # it back on top of original. |
232 | ## |
994ccd8e |
233 | my $self = shift->_get_self; |
cc4bef86 |
234 | |
235 | #XXX Need to create a new test for this |
460b1067 |
236 | # if ($self->_fileobj->{links} > 1) { |
1400a48e |
237 | # $self->_throw_error("Cannot optimize: reference count is greater than 1"); |
d0b74c17 |
238 | # } |
239 | |
240 | my $db_temp = DBM::Deep->new( |
460b1067 |
241 | file => $self->_fileobj->{file} . '.tmp', |
d0b74c17 |
242 | type => $self->_type |
243 | ); |
d0b74c17 |
244 | |
245 | $self->lock(); |
246 | $self->_copy_node( $db_temp ); |
247 | undef $db_temp; |
248 | |
249 | ## |
250 | # Attempt to copy user, group and permissions over to new file |
251 | ## |
252 | my @stats = stat($self->_fh); |
253 | my $perms = $stats[2] & 07777; |
254 | my $uid = $stats[4]; |
255 | my $gid = $stats[5]; |
460b1067 |
256 | chown( $uid, $gid, $self->_fileobj->{file} . '.tmp' ); |
257 | chmod( $perms, $self->_fileobj->{file} . '.tmp' ); |
d0b74c17 |
258 | |
ffed8b01 |
259 | # q.v. perlport for more information on this variable |
90f93b43 |
260 | if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) { |
d0b74c17 |
261 | ## |
262 | # Potential race condition when optmizing on Win32 with locking. |
263 | # The Windows filesystem requires that the filehandle be closed |
264 | # before it is overwritten with rename(). This could be redone |
265 | # with a soft copy. |
266 | ## |
267 | $self->unlock(); |
460b1067 |
268 | $self->_fileobj->close; |
d0b74c17 |
269 | } |
270 | |
460b1067 |
271 | if (!rename $self->_fileobj->{file} . '.tmp', $self->_fileobj->{file}) { |
272 | unlink $self->_fileobj->{file} . '.tmp'; |
d0b74c17 |
273 | $self->unlock(); |
1400a48e |
274 | $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!"); |
d0b74c17 |
275 | } |
276 | |
277 | $self->unlock(); |
460b1067 |
278 | $self->_fileobj->close; |
279 | $self->_fileobj->open; |
15d754dd |
280 | $self->_engine->setup_fh( $self ); |
70b55428 |
281 | |
d0b74c17 |
282 | return 1; |
ffed8b01 |
283 | } |
284 | |
285 | sub clone { |
d0b74c17 |
286 | ## |
287 | # Make copy of object and return |
288 | ## |
994ccd8e |
289 | my $self = shift->_get_self; |
d0b74c17 |
290 | |
291 | return DBM::Deep->new( |
292 | type => $self->_type, |
293 | base_offset => $self->_base_offset, |
460b1067 |
294 | fileobj => $self->_fileobj, |
d0b74c17 |
295 | ); |
ffed8b01 |
296 | } |
297 | |
298 | { |
299 | my %is_legal_filter = map { |
300 | $_ => ~~1, |
301 | } qw( |
302 | store_key store_value |
303 | fetch_key fetch_value |
304 | ); |
305 | |
306 | sub set_filter { |
307 | ## |
308 | # Setup filter function for storing or fetching the key or value |
309 | ## |
994ccd8e |
310 | my $self = shift->_get_self; |
311 | my $type = lc shift; |
312 | my $func = shift; |
d0b74c17 |
313 | |
ffed8b01 |
314 | if ( $is_legal_filter{$type} ) { |
460b1067 |
315 | $self->_fileobj->{"filter_$type"} = $func; |
ffed8b01 |
316 | return 1; |
317 | } |
318 | |
319 | return; |
320 | } |
321 | } |
322 | |
fee0243f |
323 | sub begin_work { |
324 | my $self = shift->_get_self; |
28394a1a |
325 | $self->_fileobj->begin_transaction; |
326 | return 1; |
fee0243f |
327 | } |
328 | |
329 | sub rollback { |
330 | my $self = shift->_get_self; |
28394a1a |
331 | $self->_fileobj->end_transaction; |
332 | return 1; |
fee0243f |
333 | } |
334 | |
359a01ac |
335 | sub commit { |
336 | my $self = shift->_get_self; |
25c7c8d6 |
337 | $self->_fileobj->commit_transaction; |
359a01ac |
338 | return 1; |
339 | } |
fee0243f |
340 | |
ffed8b01 |
341 | ## |
342 | # Accessor methods |
343 | ## |
344 | |
15d754dd |
345 | sub _engine { |
346 | my $self = $_[0]->_get_self; |
347 | return $self->{engine}; |
348 | } |
349 | |
460b1067 |
350 | sub _fileobj { |
2ac02042 |
351 | my $self = $_[0]->_get_self; |
460b1067 |
352 | return $self->{fileobj}; |
ffed8b01 |
353 | } |
354 | |
4d35d856 |
355 | sub _type { |
2ac02042 |
356 | my $self = $_[0]->_get_self; |
d0b74c17 |
357 | return $self->{type}; |
ffed8b01 |
358 | } |
359 | |
4d35d856 |
360 | sub _base_offset { |
2ac02042 |
361 | my $self = $_[0]->_get_self; |
d0b74c17 |
362 | return $self->{base_offset}; |
ffed8b01 |
363 | } |
364 | |
994ccd8e |
365 | sub _fh { |
994ccd8e |
366 | my $self = $_[0]->_get_self; |
460b1067 |
367 | return $self->_fileobj->{fh}; |
994ccd8e |
368 | } |
369 | |
ffed8b01 |
370 | ## |
371 | # Utility methods |
372 | ## |
373 | |
261d1296 |
374 | sub _throw_error { |
95967a5e |
375 | die "DBM::Deep: $_[1]\n"; |
ffed8b01 |
376 | } |
377 | |
acd4faf2 |
378 | sub _is_writable { |
379 | my $fh = shift; |
380 | (O_WRONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0); |
381 | } |
382 | |
9be51a89 |
383 | #sub _is_readable { |
384 | # my $fh = shift; |
385 | # (O_RDONLY | O_RDWR) & fcntl( $fh, F_GETFL, my $slush = 0); |
386 | #} |
acd4faf2 |
387 | |
359a01ac |
388 | sub _find_parent { |
389 | my $self = shift; |
cfd97a7f |
390 | |
391 | my $base = ''; |
633df1fd |
392 | #XXX This if() is redundant |
cfd97a7f |
393 | if ( my $parent = $self->{parent} ) { |
394 | my $child = $self; |
25c7c8d6 |
395 | while ( $parent->{parent} ) { |
cfd97a7f |
396 | $base = ( |
397 | $parent->_type eq TYPE_HASH |
398 | ? "\{$child->{parent_key}\}" |
399 | : "\[$child->{parent_key}\]" |
400 | ) . $base; |
401 | |
402 | $child = $parent; |
403 | $parent = $parent->{parent}; |
25c7c8d6 |
404 | } |
405 | if ( $base ) { |
406 | $base = "\$db->get( '$child->{parent_key}' )->" . $base; |
407 | } |
408 | else { |
409 | $base = "\$db->get( '$child->{parent_key}' )"; |
359a01ac |
410 | } |
359a01ac |
411 | } |
25c7c8d6 |
412 | return $base; |
359a01ac |
413 | } |
414 | |
ffed8b01 |
415 | sub STORE { |
d0b74c17 |
416 | ## |
417 | # Store single hash key/value or array element in database. |
418 | ## |
419 | my $self = shift->_get_self; |
359a01ac |
420 | my ($key, $value, $orig_key) = @_; |
81d3d316 |
421 | |
aa83bc1e |
422 | |
9e4f83a0 |
423 | if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) { |
acd4faf2 |
424 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
425 | } |
d0b74c17 |
426 | |
504185fb |
427 | #XXX The second condition needs to disappear |
428 | if ( defined $orig_key && !( $self->_type eq TYPE_ARRAY && $orig_key eq 'length') ) { |
4768a580 |
429 | my $rhs; |
430 | |
431 | my $r = Scalar::Util::reftype( $value ) || ''; |
432 | if ( $r eq 'HASH' ) { |
433 | $rhs = '{}'; |
434 | } |
435 | elsif ( $r eq 'ARRAY' ) { |
436 | $rhs = '[]'; |
437 | } |
438 | elsif ( defined $value ) { |
439 | $rhs = "'$value'"; |
440 | } |
441 | else { |
442 | $rhs = "undef"; |
443 | } |
444 | |
445 | if ( my $c = Scalar::Util::blessed( $value ) ) { |
446 | $rhs = "bless $rhs, '$c'"; |
447 | } |
448 | |
25c7c8d6 |
449 | my $lhs = $self->_find_parent; |
450 | if ( $lhs ) { |
451 | if ( $self->_type eq TYPE_HASH ) { |
452 | $lhs .= "->\{$orig_key\}"; |
453 | } |
454 | else { |
455 | $lhs .= "->\[$orig_key\]"; |
456 | } |
457 | |
458 | $lhs .= "=$rhs;"; |
459 | } |
460 | else { |
461 | $lhs = "\$db->put('$orig_key',$rhs);"; |
462 | } |
463 | |
25c7c8d6 |
464 | $self->_fileobj->audit($lhs); |
4768a580 |
465 | } |
359a01ac |
466 | |
d0b74c17 |
467 | ## |
468 | # Request exclusive lock for writing |
469 | ## |
470 | $self->lock( LOCK_EX ); |
471 | |
15d754dd |
472 | my $md5 = $self->_engine->{digest}->($key); |
d0b74c17 |
473 | |
15d754dd |
474 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5, { create => 1 } ); |
d0b74c17 |
475 | |
476 | # User may be storing a hash, in which case we do not want it run |
477 | # through the filtering system |
460b1067 |
478 | if ( !ref($value) && $self->_fileobj->{filter_store_value} ) { |
479 | $value = $self->_fileobj->{filter_store_value}->( $value ); |
d0b74c17 |
480 | } |
481 | |
482 | ## |
483 | # Add key/value to bucket list |
484 | ## |
15d754dd |
485 | $self->_engine->add_bucket( $tag, $md5, $key, $value, undef, $orig_key ); |
d0b74c17 |
486 | |
487 | $self->unlock(); |
488 | |
86867f3a |
489 | return 1; |
ffed8b01 |
490 | } |
491 | |
492 | sub FETCH { |
d0b74c17 |
493 | ## |
494 | # Fetch single value or element given plain key or array index |
495 | ## |
cb79ec85 |
496 | my $self = shift->_get_self; |
a97c8f67 |
497 | my ($key, $orig_key) = @_; |
ffed8b01 |
498 | |
15d754dd |
499 | my $md5 = $self->_engine->{digest}->($key); |
d0b74c17 |
500 | |
501 | ## |
502 | # Request shared lock for reading |
503 | ## |
504 | $self->lock( LOCK_SH ); |
505 | |
15d754dd |
506 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 );#, { create => 1 } ); |
94e8af14 |
507 | #XXX This needs to autovivify |
d0b74c17 |
508 | if (!$tag) { |
509 | $self->unlock(); |
510 | return; |
511 | } |
512 | |
513 | ## |
514 | # Get value from bucket list |
515 | ## |
15d754dd |
516 | my $result = $self->_engine->get_bucket_value( $tag, $md5, $orig_key ); |
d0b74c17 |
517 | |
518 | $self->unlock(); |
519 | |
a86430bd |
520 | # Filters only apply to scalar values, so the ref check is making |
521 | # sure the fetched bucket is a scalar, not a child hash or array. |
460b1067 |
522 | return ($result && !ref($result) && $self->_fileobj->{filter_fetch_value}) |
523 | ? $self->_fileobj->{filter_fetch_value}->($result) |
cb79ec85 |
524 | : $result; |
ffed8b01 |
525 | } |
526 | |
527 | sub DELETE { |
d0b74c17 |
528 | ## |
529 | # Delete single key/value pair or element given plain key or array index |
530 | ## |
a97c8f67 |
531 | my $self = shift->_get_self; |
532 | my ($key, $orig_key) = @_; |
d0b74c17 |
533 | |
9e4f83a0 |
534 | if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) { |
a86430bd |
535 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
536 | } |
d0b74c17 |
537 | |
4768a580 |
538 | if ( defined $orig_key ) { |
539 | my $lhs = $self->_find_parent; |
25c7c8d6 |
540 | if ( $lhs ) { |
541 | $self->_fileobj->audit( "delete $lhs;" ); |
a97c8f67 |
542 | } |
4768a580 |
543 | else { |
25c7c8d6 |
544 | $self->_fileobj->audit( "\$db->delete('$orig_key');" ); |
4768a580 |
545 | } |
a97c8f67 |
546 | } |
547 | |
d0b74c17 |
548 | ## |
549 | # Request exclusive lock for writing |
550 | ## |
551 | $self->lock( LOCK_EX ); |
552 | |
15d754dd |
553 | my $md5 = $self->_engine->{digest}->($key); |
a86430bd |
554 | |
15d754dd |
555 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 ); |
d0b74c17 |
556 | if (!$tag) { |
557 | $self->unlock(); |
558 | return; |
559 | } |
560 | |
561 | ## |
562 | # Delete bucket |
563 | ## |
15d754dd |
564 | my $value = $self->_engine->get_bucket_value( $tag, $md5 ); |
a86430bd |
565 | |
460b1067 |
566 | if (defined $value && !ref($value) && $self->_fileobj->{filter_fetch_value}) { |
567 | $value = $self->_fileobj->{filter_fetch_value}->($value); |
3b6a5056 |
568 | } |
569 | |
15d754dd |
570 | my $result = $self->_engine->delete_bucket( $tag, $md5, $orig_key ); |
d0b74c17 |
571 | |
572 | ## |
573 | # If this object is an array and the key deleted was on the end of the stack, |
574 | # decrement the length variable. |
575 | ## |
576 | |
577 | $self->unlock(); |
578 | |
579 | return $value; |
ffed8b01 |
580 | } |
581 | |
582 | sub EXISTS { |
d0b74c17 |
583 | ## |
584 | # Check if a single key or element exists given plain key or array index |
585 | ## |
a97c8f67 |
586 | my $self = shift->_get_self; |
587 | my ($key) = @_; |
d0b74c17 |
588 | |
15d754dd |
589 | my $md5 = $self->_engine->{digest}->($key); |
d0b74c17 |
590 | |
591 | ## |
592 | # Request shared lock for reading |
593 | ## |
594 | $self->lock( LOCK_SH ); |
595 | |
15d754dd |
596 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 ); |
d0b74c17 |
597 | if (!$tag) { |
598 | $self->unlock(); |
599 | |
600 | ## |
601 | # For some reason, the built-in exists() function returns '' for false |
602 | ## |
603 | return ''; |
604 | } |
605 | |
606 | ## |
607 | # Check if bucket exists and return 1 or '' |
608 | ## |
15d754dd |
609 | my $result = $self->_engine->bucket_exists( $tag, $md5 ) || ''; |
d0b74c17 |
610 | |
611 | $self->unlock(); |
612 | |
613 | return $result; |
ffed8b01 |
614 | } |
615 | |
616 | sub CLEAR { |
d0b74c17 |
617 | ## |
618 | # Clear all keys from hash, or all elements from array. |
619 | ## |
a97c8f67 |
620 | my $self = shift->_get_self; |
ffed8b01 |
621 | |
9e4f83a0 |
622 | if ( $^O ne 'MSWin32' && !_is_writable( $self->_fh ) ) { |
a86430bd |
623 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
624 | } |
625 | |
4768a580 |
626 | { |
a97c8f67 |
627 | my $lhs = $self->_find_parent; |
628 | |
a97c8f67 |
629 | if ( $self->_type eq TYPE_HASH ) { |
e82621dd |
630 | $lhs = '%{' . $lhs . '}'; |
a97c8f67 |
631 | } |
632 | else { |
e82621dd |
633 | $lhs = '@{' . $lhs . '}'; |
a97c8f67 |
634 | } |
635 | |
71a941fd |
636 | $self->_fileobj->audit( "$lhs = ();" ); |
a97c8f67 |
637 | } |
638 | |
d0b74c17 |
639 | ## |
640 | # Request exclusive lock for writing |
641 | ## |
642 | $self->lock( LOCK_EX ); |
643 | |
15d754dd |
644 | if ( $self->_type eq TYPE_HASH ) { |
645 | my $key = $self->first_key; |
646 | while ( $key ) { |
647 | my $next_key = $self->next_key( $key ); |
648 | my $md5 = $self->_engine->{digest}->($key); |
649 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 ); |
650 | $self->_engine->delete_bucket( $tag, $md5, $key ); |
651 | $key = $next_key; |
652 | } |
653 | } |
654 | else { |
655 | my $size = $self->FETCHSIZE; |
656 | for my $key ( map { pack ( $self->_engine->{long_pack}, $_ ) } 0 .. $size - 1 ) { |
657 | my $md5 = $self->_engine->{digest}->($key); |
658 | my $tag = $self->_engine->find_blist( $self->_base_offset, $md5 ); |
659 | $self->_engine->delete_bucket( $tag, $md5, $key ); |
660 | } |
661 | $self->STORESIZE( 0 ); |
662 | } |
f9c33187 |
663 | #XXX This needs updating to use _release_space |
15d754dd |
664 | # $self->_engine->write_tag( |
665 | # $self->_base_offset, $self->_type, |
666 | # chr(0)x$self->_engine->{index_size}, |
667 | # ); |
d0b74c17 |
668 | |
669 | $self->unlock(); |
670 | |
671 | return 1; |
ffed8b01 |
672 | } |
673 | |
ffed8b01 |
674 | ## |
675 | # Public method aliases |
676 | ## |
7f441181 |
677 | sub put { (shift)->STORE( @_ ) } |
678 | sub store { (shift)->STORE( @_ ) } |
679 | sub get { (shift)->FETCH( @_ ) } |
680 | sub fetch { (shift)->FETCH( @_ ) } |
baa27ab6 |
681 | sub delete { (shift)->DELETE( @_ ) } |
682 | sub exists { (shift)->EXISTS( @_ ) } |
683 | sub clear { (shift)->CLEAR( @_ ) } |
ffed8b01 |
684 | |
685 | 1; |
ffed8b01 |
686 | __END__ |
687 | |
688 | =head1 NAME |
689 | |
690 | DBM::Deep - A pure perl multi-level hash/array DBM |
691 | |
692 | =head1 SYNOPSIS |
693 | |
694 | use DBM::Deep; |
695 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
696 | |
15d754dd |
697 | $db->{key} = 'value'; |
ffed8b01 |
698 | print $db->{key}; |
d0b74c17 |
699 | |
15d754dd |
700 | $db->put('key' => 'value'); |
ffed8b01 |
701 | print $db->get('key'); |
d0b74c17 |
702 | |
ffed8b01 |
703 | # true multi-level support |
704 | $db->{my_complex} = [ |
d0b74c17 |
705 | 'hello', { perl => 'rules' }, |
706 | 42, 99, |
90f93b43 |
707 | ]; |
ffed8b01 |
708 | |
15d754dd |
709 | tie my %db, 'DBM::Deep', 'foo.db'; |
710 | $db{key} = 'value'; |
711 | print $db{key}; |
ffed8b01 |
712 | |
15d754dd |
713 | tied(%db)->put('key' => 'value'); |
714 | print tied(%db)->get('key'); |
ffed8b01 |
715 | |
15d754dd |
716 | =head1 DESCRIPTION |
8db25060 |
717 | |
15d754dd |
718 | A unique flat-file database module, written in pure perl. True multi-level |
719 | hash/array support (unlike MLDBM, which is faked), hybrid OO / tie() |
720 | interface, cross-platform FTPable files, ACID transactions, and is quite fast. |
721 | Can handle millions of keys and unlimited levels without significant |
722 | slow-down. Written from the ground-up in pure perl -- this is NOT a wrapper |
723 | around a C-based DBM. Out-of-the-box compatibility with Unix, Mac OS X and |
724 | Windows. |
ffed8b01 |
725 | |
15d754dd |
726 | =head1 VERSION DIFFERENCES |
ffed8b01 |
727 | |
15d754dd |
728 | B<NOTE>: 0.99_01 and above have significant file format differences from 0.983 and |
729 | before. There will be a backwards-compatibility layer in 1.00, but that is |
730 | slated for a later 0.99_x release. This version is B<NOT> backwards compatible |
731 | with 0.983 and before. |
ffed8b01 |
732 | |
733 | =head1 SETUP |
734 | |
d0b74c17 |
735 | Construction can be done OO-style (which is the recommended way), or using |
ffed8b01 |
736 | Perl's tie() function. Both are examined here. |
737 | |
738 | =head2 OO CONSTRUCTION |
739 | |
740 | The recommended way to construct a DBM::Deep object is to use the new() |
15d754dd |
741 | method, which gets you a blessed I<and> tied hash (or array) reference. |
ffed8b01 |
742 | |
d0b74c17 |
743 | my $db = DBM::Deep->new( "foo.db" ); |
ffed8b01 |
744 | |
745 | This opens a new database handle, mapped to the file "foo.db". If this |
d0b74c17 |
746 | file does not exist, it will automatically be created. DB files are |
ffed8b01 |
747 | opened in "r+" (read/write) mode, and the type of object returned is a |
748 | hash, unless otherwise specified (see L<OPTIONS> below). |
749 | |
ffed8b01 |
750 | You can pass a number of options to the constructor to specify things like |
15d754dd |
751 | locking, autoflush, etc. This is done by passing an inline hash (or hashref): |
ffed8b01 |
752 | |
d0b74c17 |
753 | my $db = DBM::Deep->new( |
15d754dd |
754 | file => "foo.db", |
755 | locking => 1, |
d0b74c17 |
756 | autoflush => 1 |
757 | ); |
ffed8b01 |
758 | |
759 | Notice that the filename is now specified I<inside> the hash with |
d0b74c17 |
760 | the "file" parameter, as opposed to being the sole argument to the |
ffed8b01 |
761 | constructor. This is required if any options are specified. |
762 | See L<OPTIONS> below for the complete list. |
763 | |
ffed8b01 |
764 | You can also start with an array instead of a hash. For this, you must |
765 | specify the C<type> parameter: |
766 | |
d0b74c17 |
767 | my $db = DBM::Deep->new( |
768 | file => "foo.db", |
769 | type => DBM::Deep->TYPE_ARRAY |
770 | ); |
ffed8b01 |
771 | |
772 | B<Note:> Specifing the C<type> parameter only takes effect when beginning |
773 | a new DB file. If you create a DBM::Deep object with an existing file, the |
90f93b43 |
774 | C<type> will be loaded from the file header, and an error will be thrown if |
775 | the wrong type is passed in. |
ffed8b01 |
776 | |
777 | =head2 TIE CONSTRUCTION |
778 | |
90f93b43 |
779 | Alternately, you can create a DBM::Deep handle by using Perl's built-in |
780 | tie() function. The object returned from tie() can be used to call methods, |
15d754dd |
781 | such as lock() and unlock(). (That object can be retrieved from the tied |
782 | variable at any time using tied() - please see L<perltie/> for more info. |
ffed8b01 |
783 | |
d0b74c17 |
784 | my %hash; |
785 | my $db = tie %hash, "DBM::Deep", "foo.db"; |
786 | |
787 | my @array; |
788 | my $db = tie @array, "DBM::Deep", "bar.db"; |
ffed8b01 |
789 | |
790 | As with the OO constructor, you can replace the DB filename parameter with |
791 | a hash containing one or more options (see L<OPTIONS> just below for the |
792 | complete list). |
793 | |
d0b74c17 |
794 | tie %hash, "DBM::Deep", { |
795 | file => "foo.db", |
796 | locking => 1, |
797 | autoflush => 1 |
798 | }; |
ffed8b01 |
799 | |
800 | =head2 OPTIONS |
801 | |
802 | There are a number of options that can be passed in when constructing your |
803 | DBM::Deep objects. These apply to both the OO- and tie- based approaches. |
804 | |
805 | =over |
806 | |
807 | =item * file |
808 | |
809 | Filename of the DB file to link the handle to. You can pass a full absolute |
d0b74c17 |
810 | filesystem path, partial path, or a plain filename if the file is in the |
714618f0 |
811 | current working directory. This is a required parameter (though q.v. fh). |
812 | |
813 | =item * fh |
814 | |
815 | If you want, you can pass in the fh instead of the file. This is most useful for doing |
816 | something like: |
817 | |
818 | my $db = DBM::Deep->new( { fh => \*DATA } ); |
819 | |
820 | You are responsible for making sure that the fh has been opened appropriately for your |
821 | needs. If you open it read-only and attempt to write, an exception will be thrown. If you |
822 | open it write-only or append-only, an exception will be thrown immediately as DBM::Deep |
823 | needs to read from the fh. |
824 | |
15d754dd |
825 | =item * audit_file / audit_fh |
826 | |
827 | These are just like file/fh, except for auditing. Please see L</AUDITING> for |
828 | more information. |
829 | |
714618f0 |
830 | =item * file_offset |
831 | |
832 | This is the offset within the file that the DBM::Deep db starts. Most of the time, you will |
833 | not need to set this. However, it's there if you want it. |
834 | |
835 | If you pass in fh and do not set this, it will be set appropriately. |
ffed8b01 |
836 | |
ffed8b01 |
837 | =item * type |
838 | |
839 | This parameter specifies what type of object to create, a hash or array. Use |
359a01ac |
840 | one of these two constants: |
841 | |
842 | =over 4 |
843 | |
844 | =item * C<DBM::Deep-E<gt>TYPE_HASH> |
845 | |
846 | =item * C<DBM::Deep-E<gt>TYPE_ARRAY>. |
847 | |
848 | =back |
849 | |
d0b74c17 |
850 | This only takes effect when beginning a new file. This is an optional |
ffed8b01 |
851 | parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>. |
852 | |
853 | =item * locking |
854 | |
15d754dd |
855 | Specifies whether locking is to be enabled. DBM::Deep uses Perl's flock() |
856 | function to lock the database in exclusive mode for writes, and shared mode |
857 | for reads. Pass any true value to enable. This affects the base DB handle |
858 | I<and any child hashes or arrays> that use the same DB file. This is an |
859 | optional parameter, and defaults to 0 (disabled). See L<LOCKING> below for |
860 | more. |
ffed8b01 |
861 | |
862 | =item * autoflush |
863 | |
d0b74c17 |
864 | Specifies whether autoflush is to be enabled on the underlying filehandle. |
865 | This obviously slows down write operations, but is required if you may have |
866 | multiple processes accessing the same DB file (also consider enable I<locking>). |
867 | Pass any true value to enable. This is an optional parameter, and defaults to 0 |
ffed8b01 |
868 | (disabled). |
869 | |
870 | =item * autobless |
871 | |
359a01ac |
872 | If I<autobless> mode is enabled, DBM::Deep will preserve the class something |
873 | is blessed into, and restores it when fetched. This is an optional parameter, and defaults to 1 (enabled). |
874 | |
875 | B<Note:> If you use the OO-interface, you will not be able to call any methods |
876 | of DBM::Deep on the blessed item. This is considered to be a feature. |
ffed8b01 |
877 | |
878 | =item * filter_* |
879 | |
359a01ac |
880 | See L</FILTERS> below. |
ffed8b01 |
881 | |
ffed8b01 |
882 | =back |
883 | |
884 | =head1 TIE INTERFACE |
885 | |
886 | With DBM::Deep you can access your databases using Perl's standard hash/array |
90f93b43 |
887 | syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can |
888 | treat them as such. DBM::Deep will intercept all reads/writes and direct them |
889 | to the right place -- the DB file. This has nothing to do with the |
890 | L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep |
891 | using regular hashes and arrays, rather than calling functions like C<get()> |
892 | and C<put()> (although those work too). It is entirely up to you how to want |
893 | to access your databases. |
ffed8b01 |
894 | |
895 | =head2 HASHES |
896 | |
897 | You can treat any DBM::Deep object like a normal Perl hash reference. Add keys, |
898 | or even nested hashes (or arrays) using standard Perl syntax: |
899 | |
d0b74c17 |
900 | my $db = DBM::Deep->new( "foo.db" ); |
901 | |
902 | $db->{mykey} = "myvalue"; |
903 | $db->{myhash} = {}; |
904 | $db->{myhash}->{subkey} = "subvalue"; |
ffed8b01 |
905 | |
d0b74c17 |
906 | print $db->{myhash}->{subkey} . "\n"; |
ffed8b01 |
907 | |
908 | You can even step through hash keys using the normal Perl C<keys()> function: |
909 | |
d0b74c17 |
910 | foreach my $key (keys %$db) { |
911 | print "$key: " . $db->{$key} . "\n"; |
912 | } |
ffed8b01 |
913 | |
914 | Remember that Perl's C<keys()> function extracts I<every> key from the hash and |
d0b74c17 |
915 | pushes them onto an array, all before the loop even begins. If you have an |
15d754dd |
916 | extremely large hash, this may exhaust Perl's memory. Instead, consider using |
d0b74c17 |
917 | Perl's C<each()> function, which pulls keys/values one at a time, using very |
ffed8b01 |
918 | little memory: |
919 | |
d0b74c17 |
920 | while (my ($key, $value) = each %$db) { |
921 | print "$key: $value\n"; |
922 | } |
ffed8b01 |
923 | |
924 | Please note that when using C<each()>, you should always pass a direct |
925 | hash reference, not a lookup. Meaning, you should B<never> do this: |
926 | |
d0b74c17 |
927 | # NEVER DO THIS |
928 | while (my ($key, $value) = each %{$db->{foo}}) { # BAD |
ffed8b01 |
929 | |
930 | This causes an infinite loop, because for each iteration, Perl is calling |
931 | FETCH() on the $db handle, resulting in a "new" hash for foo every time, so |
d0b74c17 |
932 | it effectively keeps returning the first key over and over again. Instead, |
ffed8b01 |
933 | assign a temporary variable to C<$db->{foo}>, then pass that to each(). |
934 | |
935 | =head2 ARRAYS |
936 | |
937 | As with hashes, you can treat any DBM::Deep object like a normal Perl array |
d0b74c17 |
938 | reference. This includes inserting, removing and manipulating elements, |
ffed8b01 |
939 | and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions. |
d0b74c17 |
940 | The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>, |
ffed8b01 |
941 | or simply be a nested array reference inside a hash. Example: |
942 | |
d0b74c17 |
943 | my $db = DBM::Deep->new( |
944 | file => "foo-array.db", |
945 | type => DBM::Deep->TYPE_ARRAY |
946 | ); |
947 | |
948 | $db->[0] = "foo"; |
949 | push @$db, "bar", "baz"; |
950 | unshift @$db, "bah"; |
951 | |
952 | my $last_elem = pop @$db; # baz |
953 | my $first_elem = shift @$db; # bah |
954 | my $second_elem = $db->[1]; # bar |
955 | |
956 | my $num_elements = scalar @$db; |
ffed8b01 |
957 | |
958 | =head1 OO INTERFACE |
959 | |
960 | In addition to the I<tie()> interface, you can also use a standard OO interface |
961 | to manipulate all aspects of DBM::Deep databases. Each type of object (hash or |
d0b74c17 |
962 | array) has its own methods, but both types share the following common methods: |
15d754dd |
963 | C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>. C<fetch()> and |
964 | C<store(> are aliases to C<put()> and C<get()>, respectively. |
ffed8b01 |
965 | |
966 | =over |
967 | |
4d35d856 |
968 | =item * new() / clone() |
969 | |
970 | These are the constructor and copy-functions. |
971 | |
90f93b43 |
972 | =item * put() / store() |
ffed8b01 |
973 | |
974 | Stores a new hash key/value pair, or sets an array element value. Takes two |
975 | arguments, the hash key or array index, and the new value. The value can be |
976 | a scalar, hash ref or array ref. Returns true on success, false on failure. |
977 | |
d0b74c17 |
978 | $db->put("foo", "bar"); # for hashes |
979 | $db->put(1, "bar"); # for arrays |
ffed8b01 |
980 | |
90f93b43 |
981 | =item * get() / fetch() |
ffed8b01 |
982 | |
983 | Fetches the value of a hash key or array element. Takes one argument: the hash |
d0b74c17 |
984 | key or array index. Returns a scalar, hash ref or array ref, depending on the |
ffed8b01 |
985 | data type stored. |
986 | |
d0b74c17 |
987 | my $value = $db->get("foo"); # for hashes |
988 | my $value = $db->get(1); # for arrays |
ffed8b01 |
989 | |
990 | =item * exists() |
991 | |
d0b74c17 |
992 | Checks if a hash key or array index exists. Takes one argument: the hash key |
ffed8b01 |
993 | or array index. Returns true if it exists, false if not. |
994 | |
d0b74c17 |
995 | if ($db->exists("foo")) { print "yay!\n"; } # for hashes |
996 | if ($db->exists(1)) { print "yay!\n"; } # for arrays |
ffed8b01 |
997 | |
998 | =item * delete() |
999 | |
1000 | Deletes one hash key/value pair or array element. Takes one argument: the hash |
1001 | key or array index. Returns true on success, false if not found. For arrays, |
1002 | the remaining elements located after the deleted element are NOT moved over. |
1003 | The deleted element is essentially just undefined, which is exactly how Perl's |
d0b74c17 |
1004 | internal arrays work. Please note that the space occupied by the deleted |
1005 | key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY> |
ffed8b01 |
1006 | below for details and workarounds. |
1007 | |
d0b74c17 |
1008 | $db->delete("foo"); # for hashes |
1009 | $db->delete(1); # for arrays |
ffed8b01 |
1010 | |
1011 | =item * clear() |
1012 | |
d0b74c17 |
1013 | Deletes B<all> hash keys or array elements. Takes no arguments. No return |
1014 | value. Please note that the space occupied by the deleted keys/values or |
1015 | elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for |
ffed8b01 |
1016 | details and workarounds. |
1017 | |
d0b74c17 |
1018 | $db->clear(); # hashes or arrays |
ffed8b01 |
1019 | |
4d35d856 |
1020 | =item * lock() / unlock() |
1021 | |
1022 | q.v. Locking. |
1023 | |
1024 | =item * optimize() |
1025 | |
15d754dd |
1026 | Recover lost disk space. This is important to do, especially if you use |
1027 | transactions. |
4d35d856 |
1028 | |
1029 | =item * import() / export() |
1030 | |
1031 | Data going in and out. |
1032 | |
ffed8b01 |
1033 | =back |
1034 | |
1035 | =head2 HASHES |
1036 | |
d0b74c17 |
1037 | For hashes, DBM::Deep supports all the common methods described above, and the |
ffed8b01 |
1038 | following additional methods: C<first_key()> and C<next_key()>. |
1039 | |
1040 | =over |
1041 | |
1042 | =item * first_key() |
1043 | |
d0b74c17 |
1044 | Returns the "first" key in the hash. As with built-in Perl hashes, keys are |
1045 | fetched in an undefined order (which appears random). Takes no arguments, |
ffed8b01 |
1046 | returns the key as a scalar value. |
1047 | |
d0b74c17 |
1048 | my $key = $db->first_key(); |
ffed8b01 |
1049 | |
1050 | =item * next_key() |
1051 | |
1052 | Returns the "next" key in the hash, given the previous one as the sole argument. |
1053 | Returns undef if there are no more keys to be fetched. |
1054 | |
d0b74c17 |
1055 | $key = $db->next_key($key); |
ffed8b01 |
1056 | |
1057 | =back |
1058 | |
1059 | Here are some examples of using hashes: |
1060 | |
d0b74c17 |
1061 | my $db = DBM::Deep->new( "foo.db" ); |
1062 | |
1063 | $db->put("foo", "bar"); |
1064 | print "foo: " . $db->get("foo") . "\n"; |
1065 | |
1066 | $db->put("baz", {}); # new child hash ref |
1067 | $db->get("baz")->put("buz", "biz"); |
1068 | print "buz: " . $db->get("baz")->get("buz") . "\n"; |
1069 | |
1070 | my $key = $db->first_key(); |
1071 | while ($key) { |
1072 | print "$key: " . $db->get($key) . "\n"; |
1073 | $key = $db->next_key($key); |
1074 | } |
1075 | |
1076 | if ($db->exists("foo")) { $db->delete("foo"); } |
ffed8b01 |
1077 | |
1078 | =head2 ARRAYS |
1079 | |
d0b74c17 |
1080 | For arrays, DBM::Deep supports all the common methods described above, and the |
1081 | following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>, |
ffed8b01 |
1082 | C<unshift()> and C<splice()>. |
1083 | |
1084 | =over |
1085 | |
1086 | =item * length() |
1087 | |
1088 | Returns the number of elements in the array. Takes no arguments. |
1089 | |
d0b74c17 |
1090 | my $len = $db->length(); |
ffed8b01 |
1091 | |
1092 | =item * push() |
1093 | |
d0b74c17 |
1094 | Adds one or more elements onto the end of the array. Accepts scalars, hash |
ffed8b01 |
1095 | refs or array refs. No return value. |
1096 | |
d0b74c17 |
1097 | $db->push("foo", "bar", {}); |
ffed8b01 |
1098 | |
1099 | =item * pop() |
1100 | |
1101 | Fetches the last element in the array, and deletes it. Takes no arguments. |
1102 | Returns undef if array is empty. Returns the element value. |
1103 | |
d0b74c17 |
1104 | my $elem = $db->pop(); |
ffed8b01 |
1105 | |
1106 | =item * shift() |
1107 | |
d0b74c17 |
1108 | Fetches the first element in the array, deletes it, then shifts all the |
1109 | remaining elements over to take up the space. Returns the element value. This |
1110 | method is not recommended with large arrays -- see L<LARGE ARRAYS> below for |
ffed8b01 |
1111 | details. |
1112 | |
d0b74c17 |
1113 | my $elem = $db->shift(); |
ffed8b01 |
1114 | |
1115 | =item * unshift() |
1116 | |
d0b74c17 |
1117 | Inserts one or more elements onto the beginning of the array, shifting all |
1118 | existing elements over to make room. Accepts scalars, hash refs or array refs. |
1119 | No return value. This method is not recommended with large arrays -- see |
ffed8b01 |
1120 | <LARGE ARRAYS> below for details. |
1121 | |
d0b74c17 |
1122 | $db->unshift("foo", "bar", {}); |
ffed8b01 |
1123 | |
1124 | =item * splice() |
1125 | |
d0b74c17 |
1126 | Performs exactly like Perl's built-in function of the same name. See L<perldoc |
ffed8b01 |
1127 | -f splice> for usage -- it is too complicated to document here. This method is |
1128 | not recommended with large arrays -- see L<LARGE ARRAYS> below for details. |
1129 | |
1130 | =back |
1131 | |
1132 | Here are some examples of using arrays: |
1133 | |
d0b74c17 |
1134 | my $db = DBM::Deep->new( |
1135 | file => "foo.db", |
1136 | type => DBM::Deep->TYPE_ARRAY |
1137 | ); |
1138 | |
1139 | $db->push("bar", "baz"); |
1140 | $db->unshift("foo"); |
1141 | $db->put(3, "buz"); |
1142 | |
1143 | my $len = $db->length(); |
1144 | print "length: $len\n"; # 4 |
1145 | |
1146 | for (my $k=0; $k<$len; $k++) { |
1147 | print "$k: " . $db->get($k) . "\n"; |
1148 | } |
1149 | |
1150 | $db->splice(1, 2, "biz", "baf"); |
1151 | |
1152 | while (my $elem = shift @$db) { |
1153 | print "shifted: $elem\n"; |
1154 | } |
ffed8b01 |
1155 | |
1156 | =head1 LOCKING |
1157 | |
d0b74c17 |
1158 | Enable automatic file locking by passing a true value to the C<locking> |
ffed8b01 |
1159 | parameter when constructing your DBM::Deep object (see L<SETUP> above). |
1160 | |
d0b74c17 |
1161 | my $db = DBM::Deep->new( |
1162 | file => "foo.db", |
1163 | locking => 1 |
1164 | ); |
ffed8b01 |
1165 | |
d0b74c17 |
1166 | This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive |
1167 | mode for writes, and shared mode for reads. This is required if you have |
1168 | multiple processes accessing the same database file, to avoid file corruption. |
1169 | Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER |
ffed8b01 |
1170 | NFS> below for more. |
1171 | |
1172 | =head2 EXPLICIT LOCKING |
1173 | |
d0b74c17 |
1174 | You can explicitly lock a database, so it remains locked for multiple |
1175 | transactions. This is done by calling the C<lock()> method, and passing an |
90f93b43 |
1176 | optional lock mode argument (defaults to exclusive mode). This is particularly |
d0b74c17 |
1177 | useful for things like counters, where the current value needs to be fetched, |
ffed8b01 |
1178 | then incremented, then stored again. |
1179 | |
d0b74c17 |
1180 | $db->lock(); |
1181 | my $counter = $db->get("counter"); |
1182 | $counter++; |
1183 | $db->put("counter", $counter); |
1184 | $db->unlock(); |
1185 | |
1186 | # or... |
ffed8b01 |
1187 | |
d0b74c17 |
1188 | $db->lock(); |
1189 | $db->{counter}++; |
1190 | $db->unlock(); |
ffed8b01 |
1191 | |
1192 | You can pass C<lock()> an optional argument, which specifies which mode to use |
d0b74c17 |
1193 | (exclusive or shared). Use one of these two constants: C<DBM::Deep-E<gt>LOCK_EX> |
1194 | or C<DBM::Deep-E<gt>LOCK_SH>. These are passed directly to C<flock()>, and are the |
ffed8b01 |
1195 | same as the constants defined in Perl's C<Fcntl> module. |
1196 | |
d0b74c17 |
1197 | $db->lock( DBM::Deep->LOCK_SH ); |
1198 | # something here |
1199 | $db->unlock(); |
ffed8b01 |
1200 | |
ffed8b01 |
1201 | =head1 IMPORTING/EXPORTING |
1202 | |
1203 | You can import existing complex structures by calling the C<import()> method, |
1204 | and export an entire database into an in-memory structure using the C<export()> |
1205 | method. Both are examined here. |
1206 | |
1207 | =head2 IMPORTING |
1208 | |
1209 | Say you have an existing hash with nested hashes/arrays inside it. Instead of |
d0b74c17 |
1210 | walking the structure and adding keys/elements to the database as you go, |
1211 | simply pass a reference to the C<import()> method. This recursively adds |
ffed8b01 |
1212 | everything to an existing DBM::Deep object for you. Here is an example: |
1213 | |
d0b74c17 |
1214 | my $struct = { |
1215 | key1 => "value1", |
1216 | key2 => "value2", |
1217 | array1 => [ "elem0", "elem1", "elem2" ], |
1218 | hash1 => { |
1219 | subkey1 => "subvalue1", |
1220 | subkey2 => "subvalue2" |
1221 | } |
1222 | }; |
1223 | |
1224 | my $db = DBM::Deep->new( "foo.db" ); |
1225 | $db->import( $struct ); |
1226 | |
1227 | print $db->{key1} . "\n"; # prints "value1" |
1228 | |
1229 | This recursively imports the entire C<$struct> object into C<$db>, including |
ffed8b01 |
1230 | all nested hashes and arrays. If the DBM::Deep object contains exsiting data, |
d0b74c17 |
1231 | keys are merged with the existing ones, replacing if they already exist. |
1232 | The C<import()> method can be called on any database level (not just the base |
ffed8b01 |
1233 | level), and works with both hash and array DB types. |
1234 | |
ffed8b01 |
1235 | B<Note:> Make sure your existing structure has no circular references in it. |
15d754dd |
1236 | These will cause an infinite loop when importing. There are plans to fix this |
1237 | in a later release. |
ffed8b01 |
1238 | |
1239 | =head2 EXPORTING |
1240 | |
d0b74c17 |
1241 | Calling the C<export()> method on an existing DBM::Deep object will return |
1242 | a reference to a new in-memory copy of the database. The export is done |
ffed8b01 |
1243 | recursively, so all nested hashes/arrays are all exported to standard Perl |
1244 | objects. Here is an example: |
1245 | |
d0b74c17 |
1246 | my $db = DBM::Deep->new( "foo.db" ); |
1247 | |
1248 | $db->{key1} = "value1"; |
1249 | $db->{key2} = "value2"; |
1250 | $db->{hash1} = {}; |
1251 | $db->{hash1}->{subkey1} = "subvalue1"; |
1252 | $db->{hash1}->{subkey2} = "subvalue2"; |
1253 | |
1254 | my $struct = $db->export(); |
1255 | |
1256 | print $struct->{key1} . "\n"; # prints "value1" |
ffed8b01 |
1257 | |
1258 | This makes a complete copy of the database in memory, and returns a reference |
d0b74c17 |
1259 | to it. The C<export()> method can be called on any database level (not just |
1260 | the base level), and works with both hash and array DB types. Be careful of |
1261 | large databases -- you can store a lot more data in a DBM::Deep object than an |
ffed8b01 |
1262 | in-memory Perl structure. |
1263 | |
ffed8b01 |
1264 | B<Note:> Make sure your database has no circular references in it. |
15d754dd |
1265 | These will cause an infinite loop when exporting. There are plans to fix this |
1266 | in a later release. |
ffed8b01 |
1267 | |
1268 | =head1 FILTERS |
1269 | |
1270 | DBM::Deep has a number of hooks where you can specify your own Perl function |
1271 | to perform filtering on incoming or outgoing data. This is a perfect |
1272 | way to extend the engine, and implement things like real-time compression or |
d0b74c17 |
1273 | encryption. Filtering applies to the base DB level, and all child hashes / |
1274 | arrays. Filter hooks can be specified when your DBM::Deep object is first |
1275 | constructed, or by calling the C<set_filter()> method at any time. There are |
ffed8b01 |
1276 | four available filter hooks, described below: |
1277 | |
1278 | =over |
1279 | |
1280 | =item * filter_store_key |
1281 | |
d0b74c17 |
1282 | This filter is called whenever a hash key is stored. It |
ffed8b01 |
1283 | is passed the incoming key, and expected to return a transformed key. |
1284 | |
1285 | =item * filter_store_value |
1286 | |
d0b74c17 |
1287 | This filter is called whenever a hash key or array element is stored. It |
ffed8b01 |
1288 | is passed the incoming value, and expected to return a transformed value. |
1289 | |
1290 | =item * filter_fetch_key |
1291 | |
d0b74c17 |
1292 | This filter is called whenever a hash key is fetched (i.e. via |
ffed8b01 |
1293 | C<first_key()> or C<next_key()>). It is passed the transformed key, |
1294 | and expected to return the plain key. |
1295 | |
1296 | =item * filter_fetch_value |
1297 | |
d0b74c17 |
1298 | This filter is called whenever a hash key or array element is fetched. |
ffed8b01 |
1299 | It is passed the transformed value, and expected to return the plain value. |
1300 | |
1301 | =back |
1302 | |
1303 | Here are the two ways to setup a filter hook: |
1304 | |
d0b74c17 |
1305 | my $db = DBM::Deep->new( |
1306 | file => "foo.db", |
1307 | filter_store_value => \&my_filter_store, |
1308 | filter_fetch_value => \&my_filter_fetch |
1309 | ); |
1310 | |
1311 | # or... |
1312 | |
1313 | $db->set_filter( "filter_store_value", \&my_filter_store ); |
1314 | $db->set_filter( "filter_fetch_value", \&my_filter_fetch ); |
ffed8b01 |
1315 | |
1316 | Your filter function will be called only when dealing with SCALAR keys or |
1317 | values. When nested hashes and arrays are being stored/fetched, filtering |
d0b74c17 |
1318 | is bypassed. Filters are called as static functions, passed a single SCALAR |
ffed8b01 |
1319 | argument, and expected to return a single SCALAR value. If you want to |
1320 | remove a filter, set the function reference to C<undef>: |
1321 | |
d0b74c17 |
1322 | $db->set_filter( "filter_store_value", undef ); |
ffed8b01 |
1323 | |
1324 | =head2 REAL-TIME ENCRYPTION EXAMPLE |
1325 | |
d0b74c17 |
1326 | Here is a working example that uses the I<Crypt::Blowfish> module to |
ffed8b01 |
1327 | do real-time encryption / decryption of keys & values with DBM::Deep Filters. |
d0b74c17 |
1328 | Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more |
ffed8b01 |
1329 | on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module. |
1330 | |
d0b74c17 |
1331 | use DBM::Deep; |
1332 | use Crypt::Blowfish; |
1333 | use Crypt::CBC; |
1334 | |
1335 | my $cipher = Crypt::CBC->new({ |
1336 | 'key' => 'my secret key', |
1337 | 'cipher' => 'Blowfish', |
1338 | 'iv' => '$KJh#(}q', |
1339 | 'regenerate_key' => 0, |
1340 | 'padding' => 'space', |
1341 | 'prepend_iv' => 0 |
1342 | }); |
1343 | |
1344 | my $db = DBM::Deep->new( |
1345 | file => "foo-encrypt.db", |
1346 | filter_store_key => \&my_encrypt, |
1347 | filter_store_value => \&my_encrypt, |
1348 | filter_fetch_key => \&my_decrypt, |
1349 | filter_fetch_value => \&my_decrypt, |
1350 | ); |
1351 | |
1352 | $db->{key1} = "value1"; |
1353 | $db->{key2} = "value2"; |
1354 | print "key1: " . $db->{key1} . "\n"; |
1355 | print "key2: " . $db->{key2} . "\n"; |
1356 | |
1357 | undef $db; |
1358 | exit; |
1359 | |
1360 | sub my_encrypt { |
1361 | return $cipher->encrypt( $_[0] ); |
1362 | } |
1363 | sub my_decrypt { |
1364 | return $cipher->decrypt( $_[0] ); |
1365 | } |
ffed8b01 |
1366 | |
1367 | =head2 REAL-TIME COMPRESSION EXAMPLE |
1368 | |
1369 | Here is a working example that uses the I<Compress::Zlib> module to do real-time |
1370 | compression / decompression of keys & values with DBM::Deep Filters. |
d0b74c17 |
1371 | Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for |
ffed8b01 |
1372 | more on I<Compress::Zlib>. |
1373 | |
d0b74c17 |
1374 | use DBM::Deep; |
1375 | use Compress::Zlib; |
1376 | |
1377 | my $db = DBM::Deep->new( |
1378 | file => "foo-compress.db", |
1379 | filter_store_key => \&my_compress, |
1380 | filter_store_value => \&my_compress, |
1381 | filter_fetch_key => \&my_decompress, |
1382 | filter_fetch_value => \&my_decompress, |
1383 | ); |
1384 | |
1385 | $db->{key1} = "value1"; |
1386 | $db->{key2} = "value2"; |
1387 | print "key1: " . $db->{key1} . "\n"; |
1388 | print "key2: " . $db->{key2} . "\n"; |
1389 | |
1390 | undef $db; |
1391 | exit; |
1392 | |
1393 | sub my_compress { |
1394 | return Compress::Zlib::memGzip( $_[0] ) ; |
1395 | } |
1396 | sub my_decompress { |
1397 | return Compress::Zlib::memGunzip( $_[0] ) ; |
1398 | } |
ffed8b01 |
1399 | |
1400 | B<Note:> Filtering of keys only applies to hashes. Array "keys" are |
1401 | actually numerical index numbers, and are not filtered. |
1402 | |
1403 | =head1 ERROR HANDLING |
1404 | |
1405 | Most DBM::Deep methods return a true value for success, and call die() on |
95967a5e |
1406 | failure. You can wrap calls in an eval block to catch the die. |
ffed8b01 |
1407 | |
d0b74c17 |
1408 | my $db = DBM::Deep->new( "foo.db" ); # create hash |
1409 | eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call |
1410 | |
90f93b43 |
1411 | print $@; # prints error message |
429e4192 |
1412 | |
ffed8b01 |
1413 | =head1 LARGEFILE SUPPORT |
1414 | |
1415 | If you have a 64-bit system, and your Perl is compiled with both LARGEFILE |
1416 | and 64-bit support, you I<may> be able to create databases larger than 2 GB. |
1417 | DBM::Deep by default uses 32-bit file offset tags, but these can be changed |
044e6288 |
1418 | by specifying the 'pack_size' parameter when constructing the file. |
ffed8b01 |
1419 | |
044e6288 |
1420 | DBM::Deep->new( |
1421 | filename => $filename, |
1422 | pack_size => 'large', |
1423 | ); |
ffed8b01 |
1424 | |
d0b74c17 |
1425 | This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words |
1426 | instead of 32-bit longs. After setting these values your DB files have a |
ffed8b01 |
1427 | theoretical maximum size of 16 XB (exabytes). |
1428 | |
044e6288 |
1429 | You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file |
1430 | offsets. |
1431 | |
ffed8b01 |
1432 | B<Note:> Changing these values will B<NOT> work for existing database files. |
044e6288 |
1433 | Only change this for new files. Once the value has been set, it is stored in |
1434 | the file's header and cannot be changed for the life of the file. These |
1435 | parameters are per-file, meaning you can access 32-bit and 64-bit files, as |
1436 | you chose. |
ffed8b01 |
1437 | |
044e6288 |
1438 | B<Note:> We have not personally tested files larger than 2 GB -- all my |
1439 | systems have only a 32-bit Perl. However, I have received user reports that |
1440 | this does indeed work! |
ffed8b01 |
1441 | |
1442 | =head1 LOW-LEVEL ACCESS |
1443 | |
90f93b43 |
1444 | If you require low-level access to the underlying filehandle that DBM::Deep uses, |
4d35d856 |
1445 | you can call the C<_fh()> method, which returns the handle: |
ffed8b01 |
1446 | |
d0b74c17 |
1447 | my $fh = $db->_fh(); |
ffed8b01 |
1448 | |
1449 | This method can be called on the root level of the datbase, or any child |
1450 | hashes or arrays. All levels share a I<root> structure, which contains things |
90f93b43 |
1451 | like the filehandle, a reference counter, and all the options specified |
460b1067 |
1452 | when you created the object. You can get access to this file object by |
1453 | calling the C<_fileobj()> method. |
ffed8b01 |
1454 | |
460b1067 |
1455 | my $file_obj = $db->_fileobj(); |
ffed8b01 |
1456 | |
1457 | This is useful for changing options after the object has already been created, |
f5be9b03 |
1458 | such as enabling/disabling locking. You can also store your own temporary user |
1459 | data in this structure (be wary of name collision), which is then accessible from |
1460 | any child hash or array. |
ffed8b01 |
1461 | |
1462 | =head1 CUSTOM DIGEST ALGORITHM |
1463 | |
1464 | DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing |
1465 | keys. However you can override this, and use another algorithm (such as SHA-256) |
d0b74c17 |
1466 | or even write your own. But please note that DBM::Deep currently expects zero |
044e6288 |
1467 | collisions, so your algorithm has to be I<perfect>, so to speak. Collision |
1468 | detection may be introduced in a later version. |
ffed8b01 |
1469 | |
044e6288 |
1470 | You can specify a custom digest algorithm by passing it into the parameter |
1471 | list for new(), passing a reference to a subroutine as the 'digest' parameter, |
1472 | and the length of the algorithm's hashes (in bytes) as the 'hash_size' |
1473 | parameter. Here is a working example that uses a 256-bit hash from the |
d0b74c17 |
1474 | I<Digest::SHA256> module. Please see |
044e6288 |
1475 | L<http://search.cpan.org/search?module=Digest::SHA256> for more information. |
ffed8b01 |
1476 | |
d0b74c17 |
1477 | use DBM::Deep; |
1478 | use Digest::SHA256; |
1479 | |
1480 | my $context = Digest::SHA256::new(256); |
1481 | |
044e6288 |
1482 | my $db = DBM::Deep->new( |
1483 | filename => "foo-sha.db", |
1484 | digest => \&my_digest, |
1485 | hash_size => 32, |
1486 | ); |
d0b74c17 |
1487 | |
1488 | $db->{key1} = "value1"; |
1489 | $db->{key2} = "value2"; |
1490 | print "key1: " . $db->{key1} . "\n"; |
1491 | print "key2: " . $db->{key2} . "\n"; |
1492 | |
1493 | undef $db; |
1494 | exit; |
1495 | |
1496 | sub my_digest { |
1497 | return substr( $context->hash($_[0]), 0, 32 ); |
1498 | } |
ffed8b01 |
1499 | |
1500 | B<Note:> Your returned digest strings must be B<EXACTLY> the number |
044e6288 |
1501 | of bytes you specify in the hash_size parameter (in this case 32). |
ffed8b01 |
1502 | |
260a80b4 |
1503 | B<Note:> If you do choose to use a custom digest algorithm, you must set it |
1504 | every time you access this file. Otherwise, the default (MD5) will be used. |
1505 | |
ffed8b01 |
1506 | =head1 CIRCULAR REFERENCES |
1507 | |
1508 | DBM::Deep has B<experimental> support for circular references. Meaning you |
1509 | can have a nested hash key or array element that points to a parent object. |
1510 | This relationship is stored in the DB file, and is preserved between sessions. |
1511 | Here is an example: |
1512 | |
d0b74c17 |
1513 | my $db = DBM::Deep->new( "foo.db" ); |
1514 | |
1515 | $db->{foo} = "bar"; |
1516 | $db->{circle} = $db; # ref to self |
1517 | |
4b93c86a |
1518 | print $db->{foo} . "\n"; # prints "bar" |
1519 | print $db->{circle}->{foo} . "\n"; # prints "bar" again |
ffed8b01 |
1520 | |
69c94980 |
1521 | B<Note>: Passing the object to a function that recursively walks the |
ffed8b01 |
1522 | object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or |
69c94980 |
1523 | C<export()> methods) will result in an infinite loop. This will be fixed in |
1524 | a future release. |
ffed8b01 |
1525 | |
15d754dd |
1526 | =head1 AUDITING |
1527 | |
1528 | New in 0.99_01 is the ability to audit your databases actions. By passing in |
1529 | audit_file (or audit_fh) to the constructor, all actions will be logged to |
1530 | that file. The format is one that is suitable for eval'ing against the |
1531 | database to replay the actions. Please see t/33_audit_trail.t for an example |
1532 | of how to do this. |
1533 | |
1534 | =head1 TRANSACTIONS |
1535 | |
1536 | New in 0.99_01 is ACID transactions. Every DBM::Deep object is completely |
1537 | transaction-ready - it is not an option you have to turn on. Three new methods |
1538 | have been added to support them. They are: |
1539 | |
1540 | =over 4 |
1541 | |
1542 | =item * begin_work() |
1543 | |
1544 | This starts a transaction. |
1545 | |
1546 | =item * commit() |
1547 | |
1548 | This applies the changes done within the transaction to the mainline and ends |
1549 | the transaction. |
1550 | |
1551 | =item * rollback() |
1552 | |
1553 | This discards the changes done within the transaction to the mainline and ends |
1554 | the transaction. |
1555 | |
1556 | =back |
1557 | |
1558 | Transactions in DBM::Deep are done using the MVCC method, the same method used |
1559 | by the InnoDB MySQL table type. |
1560 | |
ffed8b01 |
1561 | =head1 CAVEATS / ISSUES / BUGS |
1562 | |
1563 | This section describes all the known issues with DBM::Deep. It you have found |
1564 | something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>. |
1565 | |
1566 | =head2 UNUSED SPACE RECOVERY |
1567 | |
14a3acb6 |
1568 | One major caveat with DBM::Deep is that space occupied by existing keys and |
ffed8b01 |
1569 | values is not recovered when they are deleted. Meaning if you keep deleting |
1570 | and adding new keys, your file will continuously grow. I am working on this, |
d0b74c17 |
1571 | but in the meantime you can call the built-in C<optimize()> method from time to |
ffed8b01 |
1572 | time (perhaps in a crontab or something) to recover all your unused space. |
1573 | |
d0b74c17 |
1574 | $db->optimize(); # returns true on success |
ffed8b01 |
1575 | |
1576 | This rebuilds the ENTIRE database into a new file, then moves it on top of |
1577 | the original. The new file will have no unused space, thus it will take up as |
d0b74c17 |
1578 | little disk space as possible. Please note that this operation can take |
1579 | a long time for large files, and you need enough disk space to temporarily hold |
1580 | 2 copies of your DB file. The temporary file is created in the same directory |
1581 | as the original, named with a ".tmp" extension, and is deleted when the |
1582 | operation completes. Oh, and if locking is enabled, the DB is automatically |
ffed8b01 |
1583 | locked for the entire duration of the copy. |
1584 | |
d0b74c17 |
1585 | B<WARNING:> Only call optimize() on the top-level node of the database, and |
1586 | make sure there are no child references lying around. DBM::Deep keeps a reference |
ffed8b01 |
1587 | counter, and if it is greater than 1, optimize() will abort and return undef. |
1588 | |
eea0d863 |
1589 | =head2 REFERENCES |
1590 | |
1591 | (The reasons given assume a high level of Perl understanding, specifically of |
1592 | references. You can safely skip this section.) |
1593 | |
1594 | Currently, the only references supported are HASH and ARRAY. The other reference |
1595 | types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons. |
1596 | |
1597 | =over 4 |
1598 | |
1599 | =item * GLOB |
1600 | |
1601 | These are things like filehandles and other sockets. They can't be supported |
1602 | because it's completely unclear how DBM::Deep should serialize them. |
1603 | |
1604 | =item * SCALAR / REF |
1605 | |
1606 | The discussion here refers to the following type of example: |
1607 | |
1608 | my $x = 25; |
1609 | $db->{key1} = \$x; |
1610 | |
1611 | $x = 50; |
1612 | |
1613 | # In some other process ... |
1614 | |
1615 | my $val = ${ $db->{key1} }; |
1616 | |
1617 | is( $val, 50, "What actually gets stored in the DB file?" ); |
1618 | |
1619 | The problem is one of synchronization. When the variable being referred to |
1620 | changes value, the reference isn't notified. This means that the new value won't |
1621 | be stored in the datafile for other processes to read. There is no TIEREF. |
1622 | |
1623 | It is theoretically possible to store references to values already within a |
1624 | DBM::Deep object because everything already is synchronized, but the change to |
1625 | the internals would be quite large. Specifically, DBM::Deep would have to tie |
1626 | every single value that is stored. This would bloat the RAM footprint of |
1627 | DBM::Deep at least twofold (if not more) and be a significant performance drain, |
1628 | all to support a feature that has never been requested. |
1629 | |
1630 | =item * CODE |
1631 | |
1632 | L<http://search.cpan.org/search?module=Data::Dump::Streamer> provides a |
1633 | mechanism for serializing coderefs, including saving off all closure state. |
1634 | However, just as for SCALAR and REF, that closure state may change without |
1635 | notifying the DBM::Deep object storing the reference. |
1636 | |
1637 | =back |
1638 | |
ffed8b01 |
1639 | =head2 FILE CORRUPTION |
1640 | |
14a3acb6 |
1641 | The current level of error handling in DBM::Deep is minimal. Files I<are> checked |
1642 | for a 32-bit signature when opened, but other corruption in files can cause |
1643 | segmentation faults. DBM::Deep may try to seek() past the end of a file, or get |
ffed8b01 |
1644 | stuck in an infinite loop depending on the level of corruption. File write |
1645 | operations are not checked for failure (for speed), so if you happen to run |
d0b74c17 |
1646 | out of disk space, DBM::Deep will probably fail in a bad way. These things will |
ffed8b01 |
1647 | be addressed in a later version of DBM::Deep. |
1648 | |
1649 | =head2 DB OVER NFS |
1650 | |
14a3acb6 |
1651 | Beware of using DB files over NFS. DBM::Deep uses flock(), which works well on local |
d0b74c17 |
1652 | filesystems, but will NOT protect you from file corruption over NFS. I've heard |
1653 | about setting up your NFS server with a locking daemon, then using lockf() to |
1654 | lock your files, but your mileage may vary there as well. From what I |
1655 | understand, there is no real way to do it. However, if you need access to the |
1656 | underlying filehandle in DBM::Deep for using some other kind of locking scheme like |
ffed8b01 |
1657 | lockf(), see the L<LOW-LEVEL ACCESS> section above. |
1658 | |
1659 | =head2 COPYING OBJECTS |
1660 | |
d0b74c17 |
1661 | Beware of copying tied objects in Perl. Very strange things can happen. |
1662 | Instead, use DBM::Deep's C<clone()> method which safely copies the object and |
ffed8b01 |
1663 | returns a new, blessed, tied hash or array to the same level in the DB. |
1664 | |
d0b74c17 |
1665 | my $copy = $db->clone(); |
ffed8b01 |
1666 | |
90f93b43 |
1667 | B<Note>: Since clone() here is cloning the object, not the database location, any |
1668 | modifications to either $db or $copy will be visible in both. |
1669 | |
ffed8b01 |
1670 | =head2 LARGE ARRAYS |
1671 | |
1672 | Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays. |
1673 | These functions cause every element in the array to move, which can be murder |
1674 | on DBM::Deep, as every element has to be fetched from disk, then stored again in |
90f93b43 |
1675 | a different location. This will be addressed in the forthcoming version 1.00. |
ffed8b01 |
1676 | |
9be51a89 |
1677 | =head2 WRITEONLY FILES |
1678 | |
1679 | If you pass in a filehandle to new(), you may have opened it in either a readonly or |
1680 | writeonly mode. STORE will verify that the filehandle is writable. However, there |
1681 | doesn't seem to be a good way to determine if a filehandle is readable. And, if the |
1682 | filehandle isn't readable, it's not clear what will happen. So, don't do that. |
1683 | |
ffed8b01 |
1684 | =head1 PERFORMANCE |
1685 | |
1686 | This section discusses DBM::Deep's speed and memory usage. |
1687 | |
1688 | =head2 SPEED |
1689 | |
d0b74c17 |
1690 | Obviously, DBM::Deep isn't going to be as fast as some C-based DBMs, such as |
ffed8b01 |
1691 | the almighty I<BerkeleyDB>. But it makes up for it in features like true |
1692 | multi-level hash/array support, and cross-platform FTPable files. Even so, |
1693 | DBM::Deep is still pretty fast, and the speed stays fairly consistent, even |
1694 | with huge databases. Here is some test data: |
d0b74c17 |
1695 | |
1696 | Adding 1,000,000 keys to new DB file... |
1697 | |
1698 | At 100 keys, avg. speed is 2,703 keys/sec |
1699 | At 200 keys, avg. speed is 2,642 keys/sec |
1700 | At 300 keys, avg. speed is 2,598 keys/sec |
1701 | At 400 keys, avg. speed is 2,578 keys/sec |
1702 | At 500 keys, avg. speed is 2,722 keys/sec |
1703 | At 600 keys, avg. speed is 2,628 keys/sec |
1704 | At 700 keys, avg. speed is 2,700 keys/sec |
1705 | At 800 keys, avg. speed is 2,607 keys/sec |
1706 | At 900 keys, avg. speed is 2,190 keys/sec |
1707 | At 1,000 keys, avg. speed is 2,570 keys/sec |
1708 | At 2,000 keys, avg. speed is 2,417 keys/sec |
1709 | At 3,000 keys, avg. speed is 1,982 keys/sec |
1710 | At 4,000 keys, avg. speed is 1,568 keys/sec |
1711 | At 5,000 keys, avg. speed is 1,533 keys/sec |
1712 | At 6,000 keys, avg. speed is 1,787 keys/sec |
1713 | At 7,000 keys, avg. speed is 1,977 keys/sec |
1714 | At 8,000 keys, avg. speed is 2,028 keys/sec |
1715 | At 9,000 keys, avg. speed is 2,077 keys/sec |
1716 | At 10,000 keys, avg. speed is 2,031 keys/sec |
1717 | At 20,000 keys, avg. speed is 1,970 keys/sec |
1718 | At 30,000 keys, avg. speed is 2,050 keys/sec |
1719 | At 40,000 keys, avg. speed is 2,073 keys/sec |
1720 | At 50,000 keys, avg. speed is 1,973 keys/sec |
1721 | At 60,000 keys, avg. speed is 1,914 keys/sec |
1722 | At 70,000 keys, avg. speed is 2,091 keys/sec |
1723 | At 80,000 keys, avg. speed is 2,103 keys/sec |
1724 | At 90,000 keys, avg. speed is 1,886 keys/sec |
1725 | At 100,000 keys, avg. speed is 1,970 keys/sec |
1726 | At 200,000 keys, avg. speed is 2,053 keys/sec |
1727 | At 300,000 keys, avg. speed is 1,697 keys/sec |
1728 | At 400,000 keys, avg. speed is 1,838 keys/sec |
1729 | At 500,000 keys, avg. speed is 1,941 keys/sec |
1730 | At 600,000 keys, avg. speed is 1,930 keys/sec |
1731 | At 700,000 keys, avg. speed is 1,735 keys/sec |
1732 | At 800,000 keys, avg. speed is 1,795 keys/sec |
1733 | At 900,000 keys, avg. speed is 1,221 keys/sec |
1734 | At 1,000,000 keys, avg. speed is 1,077 keys/sec |
1735 | |
1736 | This test was performed on a PowerMac G4 1gHz running Mac OS X 10.3.2 & Perl |
1737 | 5.8.1, with an 80GB Ultra ATA/100 HD spinning at 7200RPM. The hash keys and |
1738 | values were between 6 - 12 chars in length. The DB file ended up at 210MB. |
ffed8b01 |
1739 | Run time was 12 min 3 sec. |
1740 | |
1741 | =head2 MEMORY USAGE |
1742 | |
1743 | One of the great things about DBM::Deep is that it uses very little memory. |
1744 | Even with huge databases (1,000,000+ keys) you will not see much increased |
14a3acb6 |
1745 | memory on your process. DBM::Deep relies solely on the filesystem for storing |
ffed8b01 |
1746 | and fetching data. Here is output from I</usr/bin/top> before even opening a |
1747 | database handle: |
1748 | |
d0b74c17 |
1749 | PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND |
1750 | 22831 root 11 0 2716 2716 1296 R 0.0 0.2 0:07 perl |
ffed8b01 |
1751 | |
d0b74c17 |
1752 | Basically the process is taking 2,716K of memory. And here is the same |
ffed8b01 |
1753 | process after storing and fetching 1,000,000 keys: |
1754 | |
d0b74c17 |
1755 | PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND |
1756 | 22831 root 14 0 2772 2772 1328 R 0.0 0.2 13:32 perl |
ffed8b01 |
1757 | |
d0b74c17 |
1758 | Notice the memory usage increased by only 56K. Test was performed on a 700mHz |
ffed8b01 |
1759 | x86 box running Linux RedHat 7.2 & Perl 5.6.1. |
1760 | |
1761 | =head1 DB FILE FORMAT |
1762 | |
1763 | In case you were interested in the underlying DB file format, it is documented |
d0b74c17 |
1764 | here in this section. You don't need to know this to use the module, it's just |
ffed8b01 |
1765 | included for reference. |
1766 | |
1767 | =head2 SIGNATURE |
1768 | |
1769 | DBM::Deep files always start with a 32-bit signature to identify the file type. |
1770 | This is at offset 0. The signature is "DPDB" in network byte order. This is |
90f93b43 |
1771 | checked for when the file is opened and an error will be thrown if it's not found. |
ffed8b01 |
1772 | |
1773 | =head2 TAG |
1774 | |
1775 | The DBM::Deep file is in a I<tagged format>, meaning each section of the file |
d0b74c17 |
1776 | has a standard header containing the type of data, the length of data, and then |
1777 | the data itself. The type is a single character (1 byte), the length is a |
ffed8b01 |
1778 | 32-bit unsigned long in network byte order, and the data is, well, the data. |
1779 | Here is how it unfolds: |
1780 | |
1781 | =head2 MASTER INDEX |
1782 | |
d0b74c17 |
1783 | Immediately after the 32-bit file signature is the I<Master Index> record. |
1784 | This is a standard tag header followed by 1024 bytes (in 32-bit mode) or 2048 |
1785 | bytes (in 64-bit mode) of data. The type is I<H> for hash or I<A> for array, |
ffed8b01 |
1786 | depending on how the DBM::Deep object was constructed. |
1787 | |
d0b74c17 |
1788 | The index works by looking at a I<MD5 Hash> of the hash key (or array index |
1789 | number). The first 8-bit char of the MD5 signature is the offset into the |
1790 | index, multipled by 4 in 32-bit mode, or 8 in 64-bit mode. The value of the |
ffed8b01 |
1791 | index element is a file offset of the next tag for the key/element in question, |
1792 | which is usually a I<Bucket List> tag (see below). |
1793 | |
ffed8b01 |
1794 | The next tag I<could> be another index, depending on how many keys/elements |
1795 | exist. See L<RE-INDEXING> below for details. |
1796 | |
1797 | =head2 BUCKET LIST |
1798 | |
d0b74c17 |
1799 | A I<Bucket List> is a collection of 16 MD5 hashes for keys/elements, plus |
1800 | file offsets to where the actual data is stored. It starts with a standard |
1801 | tag header, with type I<B>, and a data size of 320 bytes in 32-bit mode, or |
ffed8b01 |
1802 | 384 bytes in 64-bit mode. Each MD5 hash is stored in full (16 bytes), plus |
1803 | the 32-bit or 64-bit file offset for the I<Bucket> containing the actual data. |
d0b74c17 |
1804 | When the list fills up, a I<Re-Index> operation is performed (See |
ffed8b01 |
1805 | L<RE-INDEXING> below). |
1806 | |
1807 | =head2 BUCKET |
1808 | |
1809 | A I<Bucket> is a tag containing a key/value pair (in hash mode), or a |
1810 | index/value pair (in array mode). It starts with a standard tag header with |
1811 | type I<D> for scalar data (string, binary, etc.), or it could be a nested |
1812 | hash (type I<H>) or array (type I<A>). The value comes just after the tag |
1813 | header. The size reported in the tag header is only for the value, but then, |
d0b74c17 |
1814 | just after the value is another size (32-bit unsigned long) and then the plain |
1815 | key itself. Since the value is likely to be fetched more often than the plain |
ffed8b01 |
1816 | key, I figured it would be I<slightly> faster to store the value first. |
1817 | |
ffed8b01 |
1818 | If the type is I<H> (hash) or I<A> (array), the value is another I<Master Index> |
1819 | record for the nested structure, where the process begins all over again. |
1820 | |
1821 | =head2 RE-INDEXING |
1822 | |
1823 | After a I<Bucket List> grows to 16 records, its allocated space in the file is |
d0b74c17 |
1824 | exhausted. Then, when another key/element comes in, the list is converted to a |
1825 | new index record. However, this index will look at the next char in the MD5 |
1826 | hash, and arrange new Bucket List pointers accordingly. This process is called |
1827 | I<Re-Indexing>. Basically, a new index tag is created at the file EOF, and all |
1828 | 17 (16 + new one) keys/elements are removed from the old Bucket List and |
1829 | inserted into the new index. Several new Bucket Lists are created in the |
1830 | process, as a new MD5 char from the key is being examined (it is unlikely that |
ffed8b01 |
1831 | the keys will all share the same next char of their MD5s). |
1832 | |
ffed8b01 |
1833 | Because of the way the I<MD5> algorithm works, it is impossible to tell exactly |
d0b74c17 |
1834 | when the Bucket Lists will turn into indexes, but the first round tends to |
1835 | happen right around 4,000 keys. You will see a I<slight> decrease in |
1836 | performance here, but it picks back up pretty quick (see L<SPEED> above). Then |
1837 | it takes B<a lot> more keys to exhaust the next level of Bucket Lists. It's |
1838 | right around 900,000 keys. This process can continue nearly indefinitely -- |
1839 | right up until the point the I<MD5> signatures start colliding with each other, |
1840 | and this is B<EXTREMELY> rare -- like winning the lottery 5 times in a row AND |
1841 | getting struck by lightning while you are walking to cash in your tickets. |
1842 | Theoretically, since I<MD5> hashes are 128-bit values, you I<could> have up to |
1843 | 340,282,366,921,000,000,000,000,000,000,000,000,000 keys/elements (I believe |
ffed8b01 |
1844 | this is 340 unodecillion, but don't quote me). |
1845 | |
1846 | =head2 STORING |
1847 | |
d0b74c17 |
1848 | When a new key/element is stored, the key (or index number) is first run through |
1849 | I<Digest::MD5> to get a 128-bit signature (example, in hex: |
ffed8b01 |
1850 | b05783b0773d894396d475ced9d2f4f6). Then, the I<Master Index> record is checked |
37c5bcf0 |
1851 | for the first char of the signature (in this case I<b0>). If it does not exist, |
d0b74c17 |
1852 | a new I<Bucket List> is created for our key (and the next 15 future keys that |
1853 | happen to also have I<b> as their first MD5 char). The entire MD5 is written |
ffed8b01 |
1854 | to the I<Bucket List> along with the offset of the new I<Bucket> record (EOF at |
d0b74c17 |
1855 | this point, unless we are replacing an existing I<Bucket>), where the actual |
ffed8b01 |
1856 | data will be stored. |
1857 | |
1858 | =head2 FETCHING |
1859 | |
d0b74c17 |
1860 | Fetching an existing key/element involves getting a I<Digest::MD5> of the key |
1861 | (or index number), then walking along the indexes. If there are enough |
1862 | keys/elements in this DB level, there might be nested indexes, each linked to |
1863 | a particular char of the MD5. Finally, a I<Bucket List> is pointed to, which |
1864 | contains up to 16 full MD5 hashes. Each is checked for equality to the key in |
1865 | question. If we found a match, the I<Bucket> tag is loaded, where the value and |
ffed8b01 |
1866 | plain key are stored. |
1867 | |
ffed8b01 |
1868 | Fetching the plain key occurs when calling the I<first_key()> and I<next_key()> |
1869 | methods. In this process the indexes are walked systematically, and each key |
1870 | fetched in increasing MD5 order (which is why it appears random). Once the |
d0b74c17 |
1871 | I<Bucket> is found, the value is skipped and the plain key returned instead. |
1872 | B<Note:> Do not count on keys being fetched as if the MD5 hashes were |
1873 | alphabetically sorted. This only happens on an index-level -- as soon as the |
1874 | I<Bucket Lists> are hit, the keys will come out in the order they went in -- |
1875 | so it's pretty much undefined how the keys will come out -- just like Perl's |
ffed8b01 |
1876 | built-in hashes. |
1877 | |
261d1296 |
1878 | =head1 CODE COVERAGE |
1879 | |
15d754dd |
1880 | B<Devel::Cover> is used to test the code coverage of the tests. Below is the |
1881 | B<Devel::Cover> report on this distribution's test suite. |
7910cf68 |
1882 | |
15d754dd |
1883 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
1884 | File stmt bran cond sub pod time total |
1885 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
1886 | blib/lib/DBM/Deep.pm 96.2 89.0 75.0 95.8 89.5 36.0 92.9 |
1887 | blib/lib/DBM/Deep/Array.pm 96.1 88.3 100.0 96.4 100.0 15.9 94.7 |
1888 | blib/lib/DBM/Deep/Engine.pm 96.6 86.6 89.5 100.0 0.0 20.0 91.0 |
1889 | blib/lib/DBM/Deep/File.pm 99.4 88.3 55.6 100.0 0.0 19.6 89.5 |
1890 | blib/lib/DBM/Deep/Hash.pm 98.5 83.3 100.0 100.0 100.0 8.5 96.3 |
1891 | Total 96.9 87.4 81.2 98.0 38.5 100.0 92.1 |
1892 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
37c5bcf0 |
1893 | |
1894 | =head1 MORE INFORMATION |
1895 | |
1896 | Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep> |
15d754dd |
1897 | or send email to L<DBM-Deep@googlegroups.com>. You can also visit #dbm-deep on |
1898 | irc.perl.org |
261d1296 |
1899 | |
15d754dd |
1900 | =head1 MAINTAINERS |
37c5bcf0 |
1901 | |
aeeb5497 |
1902 | Rob Kinyon, L<rkinyon@cpan.org> |
ffed8b01 |
1903 | |
15d754dd |
1904 | Originally written by Joseph Huckaby, L<jhuckaby@cpan.org> |
1905 | |
ffed8b01 |
1906 | Special thanks to Adam Sah and Rich Gaushell! You know why :-) |
1907 | |
1908 | =head1 SEE ALSO |
1909 | |
1910 | perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5), |
1911 | Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3) |
1912 | |
1913 | =head1 LICENSE |
1914 | |
aeeb5497 |
1915 | Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved. |
ffed8b01 |
1916 | This is free software, you may use it and distribute it under the |
1917 | same terms as Perl itself. |
1918 | |
1919 | =cut |