A raft of minor improvements
[dbsrgits/DBM-Deep.git] / lib / DBM / Deep.pod
CommitLineData
9a63e1f2 1=head1 NAME
2
3DBM::Deep - A pure perl multi-level hash/array DBM that supports transactions
4
5=head1 SYNOPSIS
6
7 use DBM::Deep;
8 my $db = DBM::Deep->new( "foo.db" );
9
10 $db->{key} = 'value';
11 print $db->{key};
12
13 $db->put('key' => 'value');
14 print $db->get('key');
15
16 # true multi-level support
17 $db->{my_complex} = [
18 'hello', { perl => 'rules' },
19 42, 99,
20 ];
21
22 $db->begin_work;
23
24 # Do stuff here
25
26 $db->rollback;
27 $db->commit;
28
29 tie my %db, 'DBM::Deep', 'foo.db';
30 $db{key} = 'value';
31 print $db{key};
32
33 tied(%db)->put('key' => 'value');
34 print tied(%db)->get('key');
35
36=head1 DESCRIPTION
37
b48ae6ec 38A unique flat-file database module, written in pure perl. True multi-level
9a63e1f2 39hash/array support (unlike MLDBM, which is faked), hybrid OO / tie()
40interface, cross-platform FTPable files, ACID transactions, and is quite fast.
41Can handle millions of keys and unlimited levels without significant
b48ae6ec 42slow-down. Written from the ground-up in pure perl -- this is NOT a wrapper
43around a C-based DBM. Out-of-the-box compatibility with Unix, Mac OS X and
9a63e1f2 44Windows.
45
46=head1 VERSION DIFFERENCES
47
c192d0f5 48B<NOTE>: 1.0000 has significant file format differences from prior versions.
49THere is a backwards-compatibility layer at C<utils/upgrade_db.pl>. Files
50created by 1.0000 or higher are B<NOT> compatible with scripts using prior
51versions.
9a63e1f2 52
53=head1 SETUP
54
55Construction can be done OO-style (which is the recommended way), or using
b48ae6ec 56Perl's tie() function. Both are examined here.
9a63e1f2 57
b48ae6ec 58=head2 OO Construction
9a63e1f2 59
60The recommended way to construct a DBM::Deep object is to use the new()
61method, which gets you a blessed I<and> tied hash (or array) reference.
62
63 my $db = DBM::Deep->new( "foo.db" );
64
b48ae6ec 65This opens a new database handle, mapped to the file "foo.db". If this
66file does not exist, it will automatically be created. DB files are
9a63e1f2 67opened in "r+" (read/write) mode, and the type of object returned is a
68hash, unless otherwise specified (see L<OPTIONS> below).
69
70You can pass a number of options to the constructor to specify things like
b48ae6ec 71locking, autoflush, etc. This is done by passing an inline hash (or hashref):
9a63e1f2 72
73 my $db = DBM::Deep->new(
74 file => "foo.db",
75 locking => 1,
76 autoflush => 1
77 );
78
79Notice that the filename is now specified I<inside> the hash with
80the "file" parameter, as opposed to being the sole argument to the
b48ae6ec 81constructor. This is required if any options are specified.
9a63e1f2 82See L<OPTIONS> below for the complete list.
83
b48ae6ec 84You can also start with an array instead of a hash. For this, you must
9a63e1f2 85specify the C<type> parameter:
86
87 my $db = DBM::Deep->new(
88 file => "foo.db",
89 type => DBM::Deep->TYPE_ARRAY
90 );
91
92B<Note:> Specifing the C<type> parameter only takes effect when beginning
b48ae6ec 93a new DB file. If you create a DBM::Deep object with an existing file, the
9a63e1f2 94C<type> will be loaded from the file header, and an error will be thrown if
95the wrong type is passed in.
96
b48ae6ec 97=head2 Tie Construction
9a63e1f2 98
99Alternately, you can create a DBM::Deep handle by using Perl's built-in
b48ae6ec 100tie() function. The object returned from tie() can be used to call methods,
9a63e1f2 101such as lock() and unlock(). (That object can be retrieved from the tied
102variable at any time using tied() - please see L<perltie/> for more info.
103
104 my %hash;
105 my $db = tie %hash, "DBM::Deep", "foo.db";
106
107 my @array;
108 my $db = tie @array, "DBM::Deep", "bar.db";
109
110As with the OO constructor, you can replace the DB filename parameter with
111a hash containing one or more options (see L<OPTIONS> just below for the
112complete list).
113
114 tie %hash, "DBM::Deep", {
115 file => "foo.db",
116 locking => 1,
117 autoflush => 1
118 };
119
b48ae6ec 120=head2 Options
9a63e1f2 121
122There are a number of options that can be passed in when constructing your
b48ae6ec 123DBM::Deep objects. These apply to both the OO- and tie- based approaches.
9a63e1f2 124
125=over
126
127=item * file
128
b48ae6ec 129Filename of the DB file to link the handle to. You can pass a full absolute
9a63e1f2 130filesystem path, partial path, or a plain filename if the file is in the
b48ae6ec 131current working directory. This is a required parameter (though q.v. fh).
9a63e1f2 132
133=item * fh
134
135If you want, you can pass in the fh instead of the file. This is most useful for doing
136something like:
137
138 my $db = DBM::Deep->new( { fh => \*DATA } );
139
140You are responsible for making sure that the fh has been opened appropriately for your
141needs. If you open it read-only and attempt to write, an exception will be thrown. If you
142open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
143needs to read from the fh.
144
145=item * file_offset
146
147This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
148not need to set this. However, it's there if you want it.
149
150If you pass in fh and do not set this, it will be set appropriately.
151
152=item * type
153
b48ae6ec 154This parameter specifies what type of object to create, a hash or array. Use
9a63e1f2 155one of these two constants:
156
157=over 4
158
159=item * C<DBM::Deep-E<gt>TYPE_HASH>
160
161=item * C<DBM::Deep-E<gt>TYPE_ARRAY>.
162
163=back
164
b48ae6ec 165This only takes effect when beginning a new file. This is an optional
9a63e1f2 166parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.
167
168=item * locking
169
b48ae6ec 170Specifies whether locking is to be enabled. DBM::Deep uses Perl's flock()
9a63e1f2 171function to lock the database in exclusive mode for writes, and shared mode
b48ae6ec 172for reads. Pass any true value to enable. This affects the base DB handle
173I<and any child hashes or arrays> that use the same DB file. This is an
174optional parameter, and defaults to 1 (enabled). See L<LOCKING> below for
9a63e1f2 175more.
176
177=item * autoflush
178
179Specifies whether autoflush is to be enabled on the underlying filehandle.
180This obviously slows down write operations, but is required if you may have
181multiple processes accessing the same DB file (also consider enable I<locking>).
b48ae6ec 182Pass any true value to enable. This is an optional parameter, and defaults to 1
9a63e1f2 183(enabled).
184
185=item * filter_*
186
187See L</FILTERS> below.
188
189=back
190
191The following parameters may be specified in the constructor the first time the
192datafile is created. However, they will be stored in the header of the file and
193cannot be overridden by subsequent openings of the file - the values will be set
194from the values stored in the datafile's header.
195
196=over 4
197
198=item * num_txns
199
f72b2dfb 200This is the number of transactions that can be running at one time. The
201default is one - the HEAD. The minimum is one and the maximum is 255. The more
202transactions, the larger and quicker the datafile grows.
9a63e1f2 203
204See L</TRANSACTIONS> below.
205
206=item * max_buckets
207
208This is the number of entries that can be added before a reindexing. The larger
209this number is made, the larger a file gets, but the better performance you will
f72b2dfb 210have. The default and minimum number this can be is 16. The maximum is 256, but
211more than 64 isn't recommended.
212
213=item * data_sector_size
214
215This is the size in bytes of a given data sector. Data sectors will chain, so
216a value of any size can be stored. However, chaining is expensive in terms of
217time. Setting this value to something close to the expected common length of
218your scalars will improve your performance. If it is too small, your file will
219have a lot of chaining. If it is too large, your file will have a lot of dead
220space in it.
221
222The default for this is 64 bytes. The minimum value is 32 and the maximum is
223256 bytes.
224
225B<Note:> There are between 6 and 10 bytes taken up in each data sector for
226bookkeeping. (It's 4 + the number of bytes in your L</pack_size>.) This is
227included within the data_sector_size, thus the effective value is 6-10 bytes
228less than what you specified.
9a63e1f2 229
230=item * pack_size
231
232This is the size of the file pointer used throughout the file. The valid values
233are:
234
235=over 4
236
237=item * small
238
f72b2dfb 239This uses 2-byte offsets, allowing for a maximum file size of 65 KB.
9a63e1f2 240
241=item * medium (default)
242
f72b2dfb 243This uses 4-byte offsets, allowing for a maximum file size of 4 GB.
9a63e1f2 244
245=item * large
246
f72b2dfb 247This uses 8-byte offsets, allowing for a maximum file size of 16 XB
248(exabytes). This can only be enabled if your Perl is compiled for 64-bit.
9a63e1f2 249
250=back
251
252See L</LARGEFILE SUPPORT> for more information.
253
254=back
255
256=head1 TIE INTERFACE
257
258With DBM::Deep you can access your databases using Perl's standard hash/array
b48ae6ec 259syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
260treat them as such. DBM::Deep will intercept all reads/writes and direct them
261to the right place -- the DB file. This has nothing to do with the
262L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
9a63e1f2 263using regular hashes and arrays, rather than calling functions like C<get()>
b48ae6ec 264and C<put()> (although those work too). It is entirely up to you how to want
9a63e1f2 265to access your databases.
266
b48ae6ec 267=head2 Hashes
9a63e1f2 268
b48ae6ec 269You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
9a63e1f2 270or even nested hashes (or arrays) using standard Perl syntax:
271
272 my $db = DBM::Deep->new( "foo.db" );
273
274 $db->{mykey} = "myvalue";
275 $db->{myhash} = {};
276 $db->{myhash}->{subkey} = "subvalue";
277
278 print $db->{myhash}->{subkey} . "\n";
279
280You can even step through hash keys using the normal Perl C<keys()> function:
281
282 foreach my $key (keys %$db) {
283 print "$key: " . $db->{$key} . "\n";
284 }
285
286Remember that Perl's C<keys()> function extracts I<every> key from the hash and
b48ae6ec 287pushes them onto an array, all before the loop even begins. If you have an
288extremely large hash, this may exhaust Perl's memory. Instead, consider using
9a63e1f2 289Perl's C<each()> function, which pulls keys/values one at a time, using very
290little memory:
291
292 while (my ($key, $value) = each %$db) {
293 print "$key: $value\n";
294 }
295
296Please note that when using C<each()>, you should always pass a direct
b48ae6ec 297hash reference, not a lookup. Meaning, you should B<never> do this:
9a63e1f2 298
299 # NEVER DO THIS
300 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
301
302This causes an infinite loop, because for each iteration, Perl is calling
303FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
304it effectively keeps returning the first key over and over again. Instead,
305assign a temporary variable to C<$db->{foo}>, then pass that to each().
306
b48ae6ec 307=head2 Arrays
9a63e1f2 308
309As with hashes, you can treat any DBM::Deep object like a normal Perl array
b48ae6ec 310reference. This includes inserting, removing and manipulating elements,
9a63e1f2 311and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
312The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
b48ae6ec 313or simply be a nested array reference inside a hash. Example:
9a63e1f2 314
315 my $db = DBM::Deep->new(
316 file => "foo-array.db",
317 type => DBM::Deep->TYPE_ARRAY
318 );
319
320 $db->[0] = "foo";
321 push @$db, "bar", "baz";
322 unshift @$db, "bah";
323
324 my $last_elem = pop @$db; # baz
325 my $first_elem = shift @$db; # bah
326 my $second_elem = $db->[1]; # bar
327
328 my $num_elements = scalar @$db;
329
330=head1 OO INTERFACE
331
332In addition to the I<tie()> interface, you can also use a standard OO interface
b48ae6ec 333to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
9a63e1f2 334array) has its own methods, but both types share the following common methods:
335C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>. C<fetch()> and
336C<store(> are aliases to C<put()> and C<get()>, respectively.
337
338=over
339
340=item * new() / clone()
341
342These are the constructor and copy-functions.
343
344=item * put() / store()
345
b48ae6ec 346Stores a new hash key/value pair, or sets an array element value. Takes two
347arguments, the hash key or array index, and the new value. The value can be
348a scalar, hash ref or array ref. Returns true on success, false on failure.
9a63e1f2 349
350 $db->put("foo", "bar"); # for hashes
351 $db->put(1, "bar"); # for arrays
352
353=item * get() / fetch()
354
b48ae6ec 355Fetches the value of a hash key or array element. Takes one argument: the hash
356key or array index. Returns a scalar, hash ref or array ref, depending on the
9a63e1f2 357data type stored.
358
359 my $value = $db->get("foo"); # for hashes
360 my $value = $db->get(1); # for arrays
361
362=item * exists()
363
b48ae6ec 364Checks if a hash key or array index exists. Takes one argument: the hash key
365or array index. Returns true if it exists, false if not.
9a63e1f2 366
367 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
368 if ($db->exists(1)) { print "yay!\n"; } # for arrays
369
370=item * delete()
371
b48ae6ec 372Deletes one hash key/value pair or array element. Takes one argument: the hash
373key or array index. Returns true on success, false if not found. For arrays,
9a63e1f2 374the remaining elements located after the deleted element are NOT moved over.
375The deleted element is essentially just undefined, which is exactly how Perl's
b48ae6ec 376internal arrays work.
9a63e1f2 377
378 $db->delete("foo"); # for hashes
379 $db->delete(1); # for arrays
380
381=item * clear()
382
b48ae6ec 383Deletes B<all> hash keys or array elements. Takes no arguments. No return
384value.
9a63e1f2 385
386 $db->clear(); # hashes or arrays
387
388=item * lock() / unlock()
389
390q.v. Locking.
391
392=item * optimize()
393
394Recover lost disk space. This is important to do, especially if you use
395transactions.
396
397=item * import() / export()
398
399Data going in and out.
400
b48ae6ec 401=item * begin_work() / commit() / rollback()
402
403These are the transactional functions. L</TRANSACTIONS> for more information.
404
9a63e1f2 405=back
406
b48ae6ec 407=head2 Hashes
9a63e1f2 408
409For hashes, DBM::Deep supports all the common methods described above, and the
410following additional methods: C<first_key()> and C<next_key()>.
411
412=over
413
414=item * first_key()
415
b48ae6ec 416Returns the "first" key in the hash. As with built-in Perl hashes, keys are
417fetched in an undefined order (which appears random). Takes no arguments,
9a63e1f2 418returns the key as a scalar value.
419
420 my $key = $db->first_key();
421
422=item * next_key()
423
424Returns the "next" key in the hash, given the previous one as the sole argument.
425Returns undef if there are no more keys to be fetched.
426
427 $key = $db->next_key($key);
428
429=back
430
431Here are some examples of using hashes:
432
433 my $db = DBM::Deep->new( "foo.db" );
434
435 $db->put("foo", "bar");
436 print "foo: " . $db->get("foo") . "\n";
437
438 $db->put("baz", {}); # new child hash ref
439 $db->get("baz")->put("buz", "biz");
440 print "buz: " . $db->get("baz")->get("buz") . "\n";
441
442 my $key = $db->first_key();
443 while ($key) {
444 print "$key: " . $db->get($key) . "\n";
445 $key = $db->next_key($key);
446 }
447
448 if ($db->exists("foo")) { $db->delete("foo"); }
449
b48ae6ec 450=head2 Arrays
9a63e1f2 451
452For arrays, DBM::Deep supports all the common methods described above, and the
453following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
454C<unshift()> and C<splice()>.
455
456=over
457
458=item * length()
459
b48ae6ec 460Returns the number of elements in the array. Takes no arguments.
9a63e1f2 461
462 my $len = $db->length();
463
464=item * push()
465
b48ae6ec 466Adds one or more elements onto the end of the array. Accepts scalars, hash
467refs or array refs. No return value.
9a63e1f2 468
469 $db->push("foo", "bar", {});
470
471=item * pop()
472
b48ae6ec 473Fetches the last element in the array, and deletes it. Takes no arguments.
474Returns undef if array is empty. Returns the element value.
9a63e1f2 475
476 my $elem = $db->pop();
477
478=item * shift()
479
480Fetches the first element in the array, deletes it, then shifts all the
b48ae6ec 481remaining elements over to take up the space. Returns the element value. This
9a63e1f2 482method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
483details.
484
485 my $elem = $db->shift();
486
487=item * unshift()
488
489Inserts one or more elements onto the beginning of the array, shifting all
b48ae6ec 490existing elements over to make room. Accepts scalars, hash refs or array refs.
491No return value. This method is not recommended with large arrays -- see
9a63e1f2 492<LARGE ARRAYS> below for details.
493
494 $db->unshift("foo", "bar", {});
495
496=item * splice()
497
b48ae6ec 498Performs exactly like Perl's built-in function of the same name. See L<perldoc
499-f splice> for usage -- it is too complicated to document here. This method is
9a63e1f2 500not recommended with large arrays -- see L<LARGE ARRAYS> below for details.
501
502=back
503
504Here are some examples of using arrays:
505
506 my $db = DBM::Deep->new(
507 file => "foo.db",
508 type => DBM::Deep->TYPE_ARRAY
509 );
510
511 $db->push("bar", "baz");
512 $db->unshift("foo");
513 $db->put(3, "buz");
514
515 my $len = $db->length();
516 print "length: $len\n"; # 4
517
518 for (my $k=0; $k<$len; $k++) {
519 print "$k: " . $db->get($k) . "\n";
520 }
521
522 $db->splice(1, 2, "biz", "baf");
523
524 while (my $elem = shift @$db) {
525 print "shifted: $elem\n";
526 }
527
528=head1 LOCKING
529
530Enable or disable automatic file locking by passing a boolean value to the
531C<locking> parameter when constructing your DBM::Deep object (see L<SETUP>
5a4e70c0 532above).
9a63e1f2 533
534 my $db = DBM::Deep->new(
535 file => "foo.db",
536 locking => 1
537 );
538
539This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
b48ae6ec 540mode for writes, and shared mode for reads. This is required if you have
9a63e1f2 541multiple processes accessing the same database file, to avoid file corruption.
b48ae6ec 542Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
9a63e1f2 543NFS> below for more.
544
b48ae6ec 545=head2 Explicit Locking
9a63e1f2 546
547You can explicitly lock a database, so it remains locked for multiple
b48ae6ec 548actions. This is done by calling the C<lock()> method, and passing an
549optional lock mode argument (defaults to exclusive mode). This is particularly
9a63e1f2 550useful for things like counters, where the current value needs to be fetched,
551then incremented, then stored again.
552
553 $db->lock();
554 my $counter = $db->get("counter");
555 $counter++;
556 $db->put("counter", $counter);
557 $db->unlock();
558
559 # or...
560
561 $db->lock();
562 $db->{counter}++;
563 $db->unlock();
564
565You can pass C<lock()> an optional argument, which specifies which mode to use
b48ae6ec 566(exclusive or shared). Use one of these two constants:
567C<DBM::Deep-E<gt>LOCK_EX> or C<DBM::Deep-E<gt>LOCK_SH>. These are passed
9a63e1f2 568directly to C<flock()>, and are the same as the constants defined in Perl's
569L<Fcntl/> module.
570
571 $db->lock( $db->LOCK_SH );
572 # something here
573 $db->unlock();
574
575=head1 IMPORTING/EXPORTING
576
577You can import existing complex structures by calling the C<import()> method,
578and export an entire database into an in-memory structure using the C<export()>
b48ae6ec 579method. Both are examined here.
9a63e1f2 580
b48ae6ec 581=head2 Importing
9a63e1f2 582
b48ae6ec 583Say you have an existing hash with nested hashes/arrays inside it. Instead of
9a63e1f2 584walking the structure and adding keys/elements to the database as you go,
b48ae6ec 585simply pass a reference to the C<import()> method. This recursively adds
586everything to an existing DBM::Deep object for you. Here is an example:
9a63e1f2 587
588 my $struct = {
589 key1 => "value1",
590 key2 => "value2",
591 array1 => [ "elem0", "elem1", "elem2" ],
592 hash1 => {
593 subkey1 => "subvalue1",
594 subkey2 => "subvalue2"
595 }
596 };
597
598 my $db = DBM::Deep->new( "foo.db" );
599 $db->import( $struct );
600
601 print $db->{key1} . "\n"; # prints "value1"
602
603This recursively imports the entire C<$struct> object into C<$db>, including
b48ae6ec 604all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
9a63e1f2 605keys are merged with the existing ones, replacing if they already exist.
606The C<import()> method can be called on any database level (not just the base
607level), and works with both hash and array DB types.
608
609B<Note:> Make sure your existing structure has no circular references in it.
610These will cause an infinite loop when importing. There are plans to fix this
611in a later release.
612
b48ae6ec 613=head2 Exporting
9a63e1f2 614
615Calling the C<export()> method on an existing DBM::Deep object will return
b48ae6ec 616a reference to a new in-memory copy of the database. The export is done
9a63e1f2 617recursively, so all nested hashes/arrays are all exported to standard Perl
b48ae6ec 618objects. Here is an example:
9a63e1f2 619
620 my $db = DBM::Deep->new( "foo.db" );
621
622 $db->{key1} = "value1";
623 $db->{key2} = "value2";
624 $db->{hash1} = {};
625 $db->{hash1}->{subkey1} = "subvalue1";
626 $db->{hash1}->{subkey2} = "subvalue2";
627
628 my $struct = $db->export();
629
630 print $struct->{key1} . "\n"; # prints "value1"
631
632This makes a complete copy of the database in memory, and returns a reference
b48ae6ec 633to it. The C<export()> method can be called on any database level (not just
634the base level), and works with both hash and array DB types. Be careful of
9a63e1f2 635large databases -- you can store a lot more data in a DBM::Deep object than an
636in-memory Perl structure.
637
638B<Note:> Make sure your database has no circular references in it.
639These will cause an infinite loop when exporting. There are plans to fix this
640in a later release.
641
642=head1 FILTERS
643
644DBM::Deep has a number of hooks where you can specify your own Perl function
b48ae6ec 645to perform filtering on incoming or outgoing data. This is a perfect
9a63e1f2 646way to extend the engine, and implement things like real-time compression or
b48ae6ec 647encryption. Filtering applies to the base DB level, and all child hashes /
648arrays. Filter hooks can be specified when your DBM::Deep object is first
649constructed, or by calling the C<set_filter()> method at any time. There are
da1f1300 650four available filter hooks.
651
652=head2 set_filter()
653
654This method takes two paramters - the filter type and the filter subreference.
655The four types are:
9a63e1f2 656
657=over
658
659=item * filter_store_key
660
b48ae6ec 661This filter is called whenever a hash key is stored. It
9a63e1f2 662is passed the incoming key, and expected to return a transformed key.
663
664=item * filter_store_value
665
b48ae6ec 666This filter is called whenever a hash key or array element is stored. It
9a63e1f2 667is passed the incoming value, and expected to return a transformed value.
668
669=item * filter_fetch_key
670
671This filter is called whenever a hash key is fetched (i.e. via
b48ae6ec 672C<first_key()> or C<next_key()>). It is passed the transformed key,
9a63e1f2 673and expected to return the plain key.
674
675=item * filter_fetch_value
676
677This filter is called whenever a hash key or array element is fetched.
678It is passed the transformed value, and expected to return the plain value.
679
680=back
681
682Here are the two ways to setup a filter hook:
683
684 my $db = DBM::Deep->new(
685 file => "foo.db",
686 filter_store_value => \&my_filter_store,
687 filter_fetch_value => \&my_filter_fetch
688 );
689
690 # or...
691
692 $db->set_filter( "filter_store_value", \&my_filter_store );
693 $db->set_filter( "filter_fetch_value", \&my_filter_fetch );
694
695Your filter function will be called only when dealing with SCALAR keys or
b48ae6ec 696values. When nested hashes and arrays are being stored/fetched, filtering
697is bypassed. Filters are called as static functions, passed a single SCALAR
698argument, and expected to return a single SCALAR value. If you want to
9a63e1f2 699remove a filter, set the function reference to C<undef>:
700
701 $db->set_filter( "filter_store_value", undef );
702
f4593de8 703=head2 Examples
9a63e1f2 704
f4593de8 705Please read L<DBM::Deep::Manual/> for examples of filters.
9a63e1f2 706
9a63e1f2 707=head1 ERROR HANDLING
708
709Most DBM::Deep methods return a true value for success, and call die() on
b48ae6ec 710failure. You can wrap calls in an eval block to catch the die.
9a63e1f2 711
712 my $db = DBM::Deep->new( "foo.db" ); # create hash
713 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
714
715 print $@; # prints error message
716
717=head1 LARGEFILE SUPPORT
718
719If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
f72b2dfb 720and 64-bit support, you I<may> be able to create databases larger than 4 GB.
9a63e1f2 721DBM::Deep by default uses 32-bit file offset tags, but these can be changed
722by specifying the 'pack_size' parameter when constructing the file.
723
724 DBM::Deep->new(
725 filename => $filename,
726 pack_size => 'large',
727 );
728
729This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
b48ae6ec 730instead of 32-bit longs. After setting these values your DB files have a
9a63e1f2 731theoretical maximum size of 16 XB (exabytes).
732
733You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file
734offsets.
735
736B<Note:> Changing these values will B<NOT> work for existing database files.
737Only change this for new files. Once the value has been set, it is stored in
738the file's header and cannot be changed for the life of the file. These
739parameters are per-file, meaning you can access 32-bit and 64-bit files, as
740you choose.
741
da1f1300 742B<Note:> We have not personally tested files larger than 4 GB -- all our
f4593de8 743systems have only a 32-bit Perl. However, we have received user reports that
f72b2dfb 744this does indeed work.
9a63e1f2 745
746=head1 LOW-LEVEL ACCESS
747
748If you require low-level access to the underlying filehandle that DBM::Deep uses,
749you can call the C<_fh()> method, which returns the handle:
750
751 my $fh = $db->_fh();
752
753This method can be called on the root level of the datbase, or any child
b48ae6ec 754hashes or arrays. All levels share a I<root> structure, which contains things
9a63e1f2 755like the filehandle, a reference counter, and all the options specified
b48ae6ec 756when you created the object. You can get access to this file object by
9a63e1f2 757calling the C<_storage()> method.
758
759 my $file_obj = $db->_storage();
760
761This is useful for changing options after the object has already been created,
b48ae6ec 762such as enabling/disabling locking. You can also store your own temporary user
9a63e1f2 763data in this structure (be wary of name collision), which is then accessible from
764any child hash or array.
765
9a63e1f2 766=head1 CIRCULAR REFERENCES
767
da1f1300 768DBM::Deep has full support for circular references. Meaning you
9a63e1f2 769can have a nested hash key or array element that points to a parent object.
770This relationship is stored in the DB file, and is preserved between sessions.
771Here is an example:
772
773 my $db = DBM::Deep->new( "foo.db" );
774
775 $db->{foo} = "bar";
776 $db->{circle} = $db; # ref to self
777
778 print $db->{foo} . "\n"; # prints "bar"
779 print $db->{circle}->{foo} . "\n"; # prints "bar" again
780
da1f1300 781This also works as expected with array and hash references. So, the following
782works as expected:
783
784 $db->{foo} = [ 1 .. 3 ];
785 $db->{bar} = $db->{foo};
786
787 push @{$db->{foo}}, 42;
788 is( $db->{bar}[-1], 42 ); # Passes
789
790This, however, does I<not> extend to assignments from one DB file to another.
791So, the following will throw an error:
792
793 my $db1 = DBM::Deep->new( "foo.db" );
794 my $db2 = DBM::Deep->new( "bar.db" );
795
796 $db1->{foo} = [];
797 $db2->{foo} = $db1->{foo}; # dies
798
9a63e1f2 799B<Note>: Passing the object to a function that recursively walks the
800object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
801C<export()> methods) will result in an infinite loop. This will be fixed in
da1f1300 802a future release by adding singleton support.
9a63e1f2 803
804=head1 TRANSACTIONS
805
da1f1300 806As of 1.0000, DBM::Deep hass ACID transactions. Every DBM::Deep object is completely
9a63e1f2 807transaction-ready - it is not an option you have to turn on. You do have to
808specify how many transactions may run simultaneously (q.v. L</num_txns>).
809
810Three new methods have been added to support them. They are:
811
812=over 4
813
814=item * begin_work()
815
816This starts a transaction.
817
818=item * commit()
819
820This applies the changes done within the transaction to the mainline and ends
821the transaction.
822
823=item * rollback()
824
825This discards the changes done within the transaction to the mainline and ends
826the transaction.
827
828=back
829
b48ae6ec 830Transactions in DBM::Deep are done using a variant of the MVCC method, the
831same method used by the InnoDB MySQL engine.
832
f72b2dfb 833=head1 MIGRATION
834
835As of 1.0000, the file format has changed. Furthermore, DBM::Deep is now
836designed to potentially change file format between point-releases, if needed to
837support a requested feature. To aid in this, a migration script is provided
838within the CPAN distribution called C<utils/upgrade_db.pl>.
839
840B<NOTE:> This script is not installed onto your system because it carries a copy
841of every version prior to the current version.
842
9a63e1f2 843=head1 TODO
844
845The following are items that are planned to be added in future releases. These
846are separate from the L<CAVEATS, ISSUES & BUGS> below.
847
b48ae6ec 848=head2 Sub-Transactions
9a63e1f2 849
850Right now, you cannot run a transaction within a transaction. Removing this
851restriction is technically straightforward, but the combinatorial explosion of
852possible usecases hurts my head. If this is something you want to see
853immediately, please submit many testcases.
854
b48ae6ec 855=head2 Caching
9a63e1f2 856
857If a user is willing to assert upon opening the file that this process will be
858the only consumer of that datafile, then there are a number of caching
859possibilities that can be taken advantage of. This does, however, mean that
860DBM::Deep is more vulnerable to losing data due to unflushed changes. It also
861means a much larger in-memory footprint. As such, it's not clear exactly how
862this should be done. Suggestions are welcome.
863
b48ae6ec 864=head2 Ram-only
9a63e1f2 865
866The techniques used in DBM::Deep simply require a seekable contiguous
867datastore. This could just as easily be a large string as a file. By using
868substr, the STM capabilities of DBM::Deep could be used within a
869single-process. I have no idea how I'd specify this, though. Suggestions are
870welcome.
871
b48ae6ec 872=head2 Importing using Data::Walker
873
874Right now, importing is done using C<Clone::clone()> to make a complete copy
875in memory, then tying that copy. It would be much better to use
876L<Data::Walker/> to walk the data structure instead, particularly in the case
877of large datastructures.
878
879=head2 Different contention resolution mechanisms
880
881Currently, the only contention resolution mechanism is last-write-wins. This
882is the mechanism used by most RDBMSes and should be good enough for most uses.
883For advanced uses of STM, other contention mechanisms will be needed. If you
884have an idea of how you'd like to see contention resolution in DBM::Deep,
885please let me know.
886
9a63e1f2 887=head1 CAVEATS, ISSUES & BUGS
888
889This section describes all the known issues with DBM::Deep. These are issues
890that are either intractable or depend on some feature within Perl working
891exactly right. It you have found something that is not listed below, please
892send an e-mail to L<rkinyon@cpan.org>. Likewise, if you think you know of a
893way around one of these issues, please let me know.
894
b48ae6ec 895=head2 References
9a63e1f2 896
b48ae6ec 897(The following assumes a high level of Perl understanding, specifically of
898references. Most users can safely skip this section.)
9a63e1f2 899
900Currently, the only references supported are HASH and ARRAY. The other reference
901types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons.
902
903=over 4
904
905=item * GLOB
906
907These are things like filehandles and other sockets. They can't be supported
908because it's completely unclear how DBM::Deep should serialize them.
909
910=item * SCALAR / REF
911
912The discussion here refers to the following type of example:
913
914 my $x = 25;
915 $db->{key1} = \$x;
916
917 $x = 50;
918
919 # In some other process ...
920
921 my $val = ${ $db->{key1} };
922
923 is( $val, 50, "What actually gets stored in the DB file?" );
924
925The problem is one of synchronization. When the variable being referred to
b48ae6ec 926changes value, the reference isn't notified, which is kind of the point of
927references. This means that the new value won't be stored in the datafile for
928other processes to read. There is no TIEREF.
9a63e1f2 929
930It is theoretically possible to store references to values already within a
931DBM::Deep object because everything already is synchronized, but the change to
932the internals would be quite large. Specifically, DBM::Deep would have to tie
933every single value that is stored. This would bloat the RAM footprint of
934DBM::Deep at least twofold (if not more) and be a significant performance drain,
935all to support a feature that has never been requested.
936
937=item * CODE
938
939L<Data::Dump::Streamer/> provides a mechanism for serializing coderefs,
b48ae6ec 940including saving off all closure state. This would allow for DBM::Deep to
941store the code for a subroutine. Then, whenever the subroutine is read, the
942code could be C<eval()>'ed into being. However, just as for SCALAR and REF,
9a63e1f2 943that closure state may change without notifying the DBM::Deep object storing
b48ae6ec 944the reference. Again, this would generally be considered a feature.
9a63e1f2 945
946=back
947
da1f1300 948=head2 Data::Dumper and references
949
950As of 1.0003, support for independent Perl datastructures was added (q.v. L</CIRCULAR REFERENCES>
951for more info). However, because DBM::Deep doesn't properly provide the same
952in-memory data-structure for a given location on disk, Data::Dumper (and
953friends) doesn't properly note this. This will be addressed in a future release.
954
b48ae6ec 955=head2 File corruption
9a63e1f2 956
b48ae6ec 957The current level of error handling in DBM::Deep is minimal. Files I<are> checked
958for a 32-bit signature when opened, but any other form of corruption in the
959datafile can cause segmentation faults. DBM::Deep may try to C<seek()> past
960the end of a file, or get stuck in an infinite loop depending on the level and
961type of corruption. File write operations are not checked for failure (for
962speed), so if you happen to run out of disk space, DBM::Deep will probably fail in
963a bad way. These things will be addressed in a later version of DBM::Deep.
9a63e1f2 964
b48ae6ec 965=head2 DB over NFS
9a63e1f2 966
b48ae6ec 967Beware of using DBM::Deep files over NFS. DBM::Deep uses flock(), which works
9a63e1f2 968well on local filesystems, but will NOT protect you from file corruption over
b48ae6ec 969NFS. I've heard about setting up your NFS server with a locking daemon, then
9a63e1f2 970using C<lockf()> to lock your files, but your mileage may vary there as well.
b48ae6ec 971From what I understand, there is no real way to do it. However, if you need
9a63e1f2 972access to the underlying filehandle in DBM::Deep for using some other kind of
973locking scheme like C<lockf()>, see the L<LOW-LEVEL ACCESS> section above.
974
b48ae6ec 975=head2 Copying Objects
9a63e1f2 976
b48ae6ec 977Beware of copying tied objects in Perl. Very strange things can happen.
9a63e1f2 978Instead, use DBM::Deep's C<clone()> method which safely copies the object and
979returns a new, blessed and tied hash or array to the same level in the DB.
980
981 my $copy = $db->clone();
982
983B<Note>: Since clone() here is cloning the object, not the database location, any
984modifications to either $db or $copy will be visible to both.
985
b48ae6ec 986=head2 Large Arrays
9a63e1f2 987
988Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
989These functions cause every element in the array to move, which can be murder
990on DBM::Deep, as every element has to be fetched from disk, then stored again in
b48ae6ec 991a different location. This will be addressed in a future version.
9a63e1f2 992
b48ae6ec 993=head2 Writeonly Files
9a63e1f2 994
995If you pass in a filehandle to new(), you may have opened it in either a readonly or
996writeonly mode. STORE will verify that the filehandle is writable. However, there
997doesn't seem to be a good way to determine if a filehandle is readable. And, if the
998filehandle isn't readable, it's not clear what will happen. So, don't do that.
999
b48ae6ec 1000=head2 Assignments Within Transactions
9a63e1f2 1001
1002The following will I<not> work as one might expect:
1003
1004 my $x = { a => 1 };
1005
1006 $db->begin_work;
1007 $db->{foo} = $x;
1008 $db->rollback;
1009
1010 is( $x->{a}, 1 ); # This will fail!
1011
1012The problem is that the moment a reference used as the rvalue to a DBM::Deep
1013object's lvalue, it becomes tied itself. This is so that future changes to
1014C<$x> can be tracked within the DBM::Deep file and is considered to be a
1015feature. By the time the rollback occurs, there is no knowledge that there had
1016been an C<$x> or what memory location to assign an C<export()> to.
1017
1018B<NOTE:> This does not affect importing because imports do a walk over the
1019reference to be imported in order to explicitly leave it untied.
1020
1021=head1 CODE COVERAGE
1022
1023B<Devel::Cover> is used to test the code coverage of the tests. Below is the
1024B<Devel::Cover> report on this distribution's test suite.
1025
d005c2e8 1026 ------------------------------------------ ------ ------ ------ ------ ------
1027 File stmt bran cond sub total
1028 ------------------------------------------ ------ ------ ------ ------ ------
1029 blib/lib/DBM/Deep.pm 94.5 85.0 90.5 100.0 93.6
1030 blib/lib/DBM/Deep/Array.pm 100.0 94.3 100.0 100.0 98.7
1031 blib/lib/DBM/Deep/Engine.pm 95.9 84.9 81.7 100.0 92.8
1032 blib/lib/DBM/Deep/File.pm 97.2 81.6 66.7 100.0 91.9
1033 blib/lib/DBM/Deep/Hash.pm 100.0 100.0 100.0 100.0 100.0
1034 Total 96.5 86.5 83.5 100.0 94.0
1035 ------------------------------------------ ------ ------ ------ ------ ------
1036
9a63e1f2 1037=head1 MORE INFORMATION
1038
1039Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
1040or send email to L<DBM-Deep@googlegroups.com>. You can also visit #dbm-deep on
1041irc.perl.org
1042
1043The source code repository is at L<http://svn.perl.org/modules/DBM-Deep>
1044
f72b2dfb 1045=head1 MAINTAINERS
9a63e1f2 1046
1047Rob Kinyon, L<rkinyon@cpan.org>
1048
1049Originally written by Joseph Huckaby, L<jhuckaby@cpan.org>
1050
f72b2dfb 1051=head1 SPONSORS
1052
1053Stonehenge Consulting (L<http://www.stonehenge.com/>) sponsored the
1054developement of transactions and freespace management, leading to the 1.0000
1055release. A great debt of gratitude goes out to them for their continuing
1056leadership in and support of the Perl community.
1057
9a63e1f2 1058=head1 CONTRIBUTORS
1059
1060The following have contributed greatly to make DBM::Deep what it is today:
1061
1062=over 4
1063
f72b2dfb 1064=item * Adam Sah and Rich Gaushell for innumerable contributions early on.
9a63e1f2 1065
1066=item * Dan Golden and others at YAPC::NA 2006 for helping me design through transactions.
1067
1068=back
1069
1070=head1 SEE ALSO
1071
1072perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
1073Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)
1074
1075=head1 LICENSE
1076
b48ae6ec 1077Copyright (c) 2007 Rob Kinyon. All Rights Reserved.
f72b2dfb 1078This is free software, you may use it and distribute it under the same terms
1079as Perl itself.
9a63e1f2 1080
1081=cut