1DBM::Deep(3) User Contributed Perl Documentation DBM::Deep(3)
2
6 DBM::Deep - A pure perl multi-level hash/array DBM that supports
7 transactions
8
10 2.0016
11
13 use DBM::Deep;
14 my $db = DBM::Deep->new( "foo.db" );
15 $db->{key} = 'value';
17 print $db->{key};
18 $db->put('key' => 'value');
20 print $db->get('key');
21 # true multi-level support
23 $db->{my_complex} = [
24 'hello', { perl => 'rules' },
25 42, 99,
26 ];
27 $db->begin_work;
29 # Do stuff here
31 $db->rollback;
33 $db->commit;
34 tie my %db, 'DBM::Deep', 'foo.db';
36 $db{key} = 'value';
37 print $db{key};
38 tied(%db)->put('key' => 'value');
40 print tied(%db)->get('key');
41
43 A unique flat-file database module, written in pure perl. True multi-
44 level hash/array support (unlike MLDBM, which is faked), hybrid OO /
45 tie() interface, cross-platform FTPable files, ACID transactions, and
46 is quite fast. Can handle millions of keys and unlimited levels
47 without significant slow-down. Written from the ground-up in pure perl
48 -- this is NOT a wrapper around a C-based DBM. Out-of-the-box
49 compatibility with Unix, Mac OS X and Windows.
50
52 NOTE: 2.0000 introduces Unicode support in the File back end. This
53 necessitates a change in the file format. The version 1.0003 format is
54 still supported, though, so we have added a db_version() method. If you
55 are using a database in the old format, you will have to upgrade it to
56 get Unicode support.
57 NOTE: 1.0020 introduces different engines which are backed by different
59 types of storage. There is the original storage (called 'File') and a
60 database storage (called 'DBI'). q.v. "PLUGINS" for more information.
61 NOTE: 1.0000 has significant file format differences from prior
63 versions. There is a backwards-compatibility layer at
64 "utils/upgrade_db.pl". Files created by 1.0000 or higher are NOT
65 compatible with scripts using prior versions.
66
68 DBM::Deep is a wrapper around different storage engines. These are:
69 File
71 This is the traditional storage engine, storing the data to a custom
72 file format. The parameters accepted are:
73 • file
75 Filename of the DB file to link the handle to. You can pass a full
77 absolute filesystem path, partial path, or a plain filename if the
78 file is in the current working directory. This is a required
79 parameter (though q.v. fh).
80 • fh
82 If you want, you can pass in the fh instead of the file. This is
84 most useful for doing something like:
85 my $db = DBM::Deep->new( { fh => \*DATA } );
87 You are responsible for making sure that the fh has been opened
89 appropriately for your needs. If you open it read-only and attempt
90 to write, an exception will be thrown. If you open it write-only or
91 append-only, an exception will be thrown immediately as DBM::Deep
92 needs to read from the fh.
93 • file_offset
95 This is the offset within the file that the DBM::Deep db starts.
97 Most of the time, you will not need to set this. However, it's
98 there if you want it.
99 If you pass in fh and do not set this, it will be set
101 appropriately.
102 • locking
104 Specifies whether locking is to be enabled. DBM::Deep uses Perl's
106 flock() function to lock the database in exclusive mode for writes,
107 and shared mode for reads. Pass any true value to enable. This
108 affects the base DB handle and any child hashes or arrays that use
109 the same DB file. This is an optional parameter, and defaults to 1
110 (enabled). See "LOCKING" below for more.
111 When you open an existing database file, the version of the database
113 format will stay the same. But if you are creating a new file, it will
114 be in the latest format.
115 DBI
117 This is a storage engine that stores the data in a relational database.
118 Funnily enough, this engine doesn't work with transactions (yet) as
119 InnoDB doesn't do what DBM::Deep needs it to do.
120 The parameters accepted are:
122 • dbh
124 This is a DBH that's already been opened with "connect" in DBI.
126 • dbi
128 This is a hashref containing:
130 • dsn
132 • username
134 • password
136 • connect_args
138 These correspond to the 4 parameters "connect" in DBI takes.
140 NOTE: This has only been tested with MySQL and SQLite (with
142 disappointing results). I plan on extending this to work with
143 PostgreSQL in the near future. Oracle, Sybase, and other engines will
144 come later.
145 Planned engines
147 There are plans to extend this functionality to (at least) the
148 following:
149 • BDB (and other hash engines like memcached)
151 • NoSQL engines (such as Tokyo Cabinet)
153 • DBIx::Class (and other ORMs)
155
157 Construction can be done OO-style (which is the recommended way), or
158 using Perl's tie() function. Both are examined here.
159 OO Construction
161 The recommended way to construct a DBM::Deep object is to use the new()
162 method, which gets you a blessed and tied hash (or array) reference.
163 my $db = DBM::Deep->new( "foo.db" );
165 This opens a new database handle, mapped to the file "foo.db". If this
167 file does not exist, it will automatically be created. DB files are
168 opened in "r+" (read/write) mode, and the type of object returned is a
169 hash, unless otherwise specified (see "Options" below).
170 You can pass a number of options to the constructor to specify things
172 like locking, autoflush, etc. This is done by passing an inline hash
173 (or hashref):
174 my $db = DBM::Deep->new(
176 file => "foo.db",
177 locking => 1,
178 autoflush => 1
179 );
180 Notice that the filename is now specified inside the hash with the
182 "file" parameter, as opposed to being the sole argument to the
183 constructor. This is required if any options are specified. See
184 "Options" below for the complete list.
185 You can also start with an array instead of a hash. For this, you must
187 specify the "type" parameter:
188 my $db = DBM::Deep->new(
190 file => "foo.db",
191 type => DBM::Deep->TYPE_ARRAY
192 );
193 Note: Specifying the "type" parameter only takes effect when beginning
195 a new DB file. If you create a DBM::Deep object with an existing file,
196 the "type" will be loaded from the file header, and an error will be
197 thrown if the wrong type is passed in.
198 Tie Construction
200 Alternately, you can create a DBM::Deep handle by using Perl's built-in
201 tie() function. The object returned from tie() can be used to call
202 methods, such as lock() and unlock(). (That object can be retrieved
203 from the tied variable at any time using tied() - please see perltie
204 for more info.)
205 my %hash;
207 my $db = tie %hash, "DBM::Deep", "foo.db";
208 my @array;
210 my $db = tie @array, "DBM::Deep", "bar.db";
211 As with the OO constructor, you can replace the DB filename parameter
213 with a hash containing one or more options (see "Options" just below
214 for the complete list).
215 tie %hash, "DBM::Deep", {
217 file => "foo.db",
218 locking => 1,
219 autoflush => 1
220 };
221 Options
223 There are a number of options that can be passed in when constructing
224 your DBM::Deep objects. These apply to both the OO- and tie- based
225 approaches.
226 • type
228 This parameter specifies what type of object to create, a hash or
230 array. Use one of these two constants:
231 • "DBM::Deep->TYPE_HASH"
233 • "DBM::Deep->TYPE_ARRAY"
235 This only takes effect when beginning a new file. This is an
237 optional parameter, and defaults to "DBM::Deep->TYPE_HASH".
238 • autoflush
240 Specifies whether autoflush is to be enabled on the underlying
242 filehandle. This obviously slows down write operations, but is
243 required if you may have multiple processes accessing the same DB
244 file (also consider enable locking). Pass any true value to
245 enable. This is an optional parameter, and defaults to 1 (enabled).
246 • filter_*
248 See "FILTERS" below.
250 The following parameters may be specified in the constructor the first
252 time the datafile is created. However, they will be stored in the
253 header of the file and cannot be overridden by subsequent openings of
254 the file - the values will be set from the values stored in the
255 datafile's header.
256 • num_txns
258 This is the number of transactions that can be running at one time.
260 The default is one - the HEAD. The minimum is one and the maximum
261 is 255. The more transactions, the larger and quicker the datafile
262 grows.
263 Simple access to a database, regardless of how many processes are
265 doing it, already counts as one transaction (the HEAD). So, if you
266 want, say, 5 processes to be able to call begin_work at the same
267 time, "num_txns" must be at least 6.
268 See "TRANSACTIONS" below.
270 • max_buckets
272 This is the number of entries that can be added before a
274 reindexing. The larger this number is made, the larger a file gets,
275 but the better performance you will have. The default and minimum
276 number this can be is 16. The maximum is 256, but more than 64
277 isn't recommended.
278 • data_sector_size
280 This is the size in bytes of a given data sector. Data sectors will
282 chain, so a value of any size can be stored. However, chaining is
283 expensive in terms of time. Setting this value to something close
284 to the expected common length of your scalars will improve your
285 performance. If it is too small, your file will have a lot of
286 chaining. If it is too large, your file will have a lot of dead
287 space in it.
288 The default for this is 64 bytes. The minimum value is 32 and the
290 maximum is 256 bytes.
291 Note: There are between 6 and 10 bytes taken up in each data sector
293 for bookkeeping. (It's 4 + the number of bytes in your
294 "pack_size".) This is included within the data_sector_size, thus
295 the effective value is 6-10 bytes less than what you specified.
296 Another note: If your strings contain any characters beyond the
298 byte range, they will be encoded as UTF-8 before being stored in
299 the file. This will make all non-ASCII characters take up more than
300 one byte each.
301 • pack_size
303 This is the size of the file pointer used throughout the file. The
305 valid values are:
306 • small
308 This uses 2-byte offsets, allowing for a maximum file size of
310 65 KB.
311 • medium (default)
313 This uses 4-byte offsets, allowing for a maximum file size of 4
315 GB.
316 • large
318 This uses 8-byte offsets, allowing for a maximum file size of
320 16 XB (exabytes). This can only be enabled if your Perl is
321 compiled for 64-bit.
322 See "LARGEFILE SUPPORT" for more information.
324 • external_refs
326 This is a boolean option. When enabled, it allows external
328 references to database entries to hold on to those entries, even
329 when they are deleted.
330 To illustrate, if you retrieve a hash (or array) reference from the
332 database,
333 $foo_hash = $db->{foo};
335 the hash reference is still tied to the database. So if you
337 delete $db->{foo};
339 $foo_hash will point to a location in the DB that is no longer
341 valid (we call this a stale reference). So if you try to retrieve
342 the data from $foo_hash,
343 for(keys %$foo_hash) {
345 you will get an error.
347 The "external_refs" option causes $foo_hash to 'hang on' to the DB
349 entry, so it will not be deleted from the database if there is
350 still a reference to it in a running program. It will be deleted,
351 instead, when the $foo_hash variable no longer exists, or is
352 overwritten.
353 This has the potential to cause database bloat if your program
355 crashes, so it is not enabled by default. (See also the "export"
356 method for an alternative workaround.)
357
359 With DBM::Deep you can access your databases using Perl's standard
360 hash/array syntax. Because all DBM::Deep objects are tied to hashes or
361 arrays, you can treat them as such (but see "external_refs", above, and
362 "Stale References", below). DBM::Deep will intercept all reads/writes
363 and direct them to the right place -- the DB file. This has nothing to
364 do with the "Tie Construction" section above. This simply tells you how
365 to use DBM::Deep using regular hashes and arrays, rather than calling
366 functions like "get()" and "put()" (although those work too). It is
367 entirely up to you how to want to access your databases.
368 Hashes
370 You can treat any DBM::Deep object like a normal Perl hash reference.
371 Add keys, or even nested hashes (or arrays) using standard Perl syntax:
372 my $db = DBM::Deep->new( "foo.db" );
374 $db->{mykey} = "myvalue";
376 $db->{myhash} = {};
377 $db->{myhash}->{subkey} = "subvalue";
378 print $db->{myhash}->{subkey} . "\n";
380 You can even step through hash keys using the normal Perl "keys()"
382 function:
383 foreach my $key (keys %$db) {
385 print "$key: " . $db->{$key} . "\n";
386 }
387 Remember that Perl's "keys()" function extracts every key from the hash
389 and pushes them onto an array, all before the loop even begins. If you
390 have an extremely large hash, this may exhaust Perl's memory. Instead,
391 consider using Perl's "each()" function, which pulls keys/values one at
392 a time, using very little memory:
393 while (my ($key, $value) = each %$db) {
395 print "$key: $value\n";
396 }
397 Please note that when using "each()", you should always pass a direct
399 hash reference, not a lookup. Meaning, you should never do this:
400 # NEVER DO THIS
402 while (my ($key, $value) = each %{$db->{foo}}) { # BAD
403 This causes an infinite loop, because for each iteration, Perl is
405 calling FETCH() on the $db handle, resulting in a "new" hash for foo
406 every time, so it effectively keeps returning the first key over and
407 over again. Instead, assign a temporary variable to "$db->{foo}", then
408 pass that to each().
409 Arrays
411 As with hashes, you can treat any DBM::Deep object like a normal Perl
412 array reference. This includes inserting, removing and manipulating
413 elements, and the "push()", "pop()", "shift()", "unshift()" and
414 "splice()" functions. The object must have first been created using
415 type "DBM::Deep->TYPE_ARRAY", or simply be a nested array reference
416 inside a hash. Example:
417 my $db = DBM::Deep->new(
419 file => "foo-array.db",
420 type => DBM::Deep->TYPE_ARRAY
421 );
422 $db->[0] = "foo";
424 push @$db, "bar", "baz";
425 unshift @$db, "bah";
426 my $last_elem = pop @$db; # baz
428 my $first_elem = shift @$db; # bah
429 my $second_elem = $db->[1]; # bar
430 my $num_elements = scalar @$db;
432
434 In addition to the tie() interface, you can also use a standard OO
435 interface to manipulate all aspects of DBM::Deep databases. Each type
436 of object (hash or array) has its own methods, but both types share the
437 following common methods: "put()", "get()", "exists()", "delete()" and
438 "clear()". "fetch()" and "store()" are aliases to "put()" and "get()",
439 respectively.
440 • new() / clone()
442 These are the constructor and copy-functions.
444 • put() / store()
446 Stores a new hash key/value pair, or sets an array element value.
448 Takes two arguments, the hash key or array index, and the new
449 value. The value can be a scalar, hash ref or array ref. Returns
450 true on success, false on failure.
451 $db->put("foo", "bar"); # for hashes
453 $db->put(1, "bar"); # for arrays
454 • get() / fetch()
456 Fetches the value of a hash key or array element. Takes one
458 argument: the hash key or array index. Returns a scalar, hash ref
459 or array ref, depending on the data type stored.
460 my $value = $db->get("foo"); # for hashes
462 my $value = $db->get(1); # for arrays
463 • exists()
465 Checks if a hash key or array index exists. Takes one argument: the
467 hash key or array index. Returns true if it exists, false if not.
468 if ($db->exists("foo")) { print "yay!\n"; } # for hashes
470 if ($db->exists(1)) { print "yay!\n"; } # for arrays
471 • delete()
473 Deletes one hash key/value pair or array element. Takes one
475 argument: the hash key or array index. Returns the data that the
476 element used to contain (just like Perl's "delete" function), which
477 is "undef" if it did not exist. For arrays, the remaining elements
478 located after the deleted element are NOT moved over. The deleted
479 element is essentially just undefined, which is exactly how Perl's
480 internal arrays work.
481 $db->delete("foo"); # for hashes
483 $db->delete(1); # for arrays
484 • clear()
486 Deletes all hash keys or array elements. Takes no arguments. No
488 return value.
489 $db->clear(); # hashes or arrays
491 • lock() / unlock() / lock_exclusive() / lock_shared()
493 q.v. "LOCKING" for more info.
495 • optimize()
497 This will compress the datafile so that it takes up as little space
499 as possible. There is a freespace manager so that when space is
500 freed up, it is used before extending the size of the datafile.
501 But, that freespace just sits in the datafile unless "optimize()"
502 is called.
503 "optimize" basically copies everything into a new database, so, if
505 it is in version 1.0003 format, it will be upgraded.
506 • import()
508 Unlike simple assignment, "import()" does not tie the right-hand
510 side. Instead, a copy of your data is put into the DB. "import()"
511 takes either an arrayref (if your DB is an array) or a hashref (if
512 your DB is a hash). "import()" will die if anything else is passed
513 in.
514 • export()
516 This returns a complete copy of the data structure at the point you
518 do the export. This copy is in RAM, not on disk like the DB is.
519 • begin_work() / commit() / rollback()
521 These are the transactional functions. "TRANSACTIONS" for more
523 information.
524 • supports( $option )
526 This returns a boolean indicating whether this instance of
528 DBM::Deep supports that feature. $option can be one of:
529 • transactions
531 • unicode
533 • db_version()
535 This returns the version of the database format that the current
537 database is in. This is specified as the earliest version of
538 DBM::Deep that supports it.
539 For the File back end, this will be 1.0003 or 2.
541 For the DBI back end, it is currently always 1.0020.
543 Hashes
545 For hashes, DBM::Deep supports all the common methods described above,
546 and the following additional methods: "first_key()" and "next_key()".
547 • first_key()
549 Returns the "first" key in the hash. As with built-in Perl hashes,
551 keys are fetched in an undefined order (which appears random).
552 Takes no arguments, returns the key as a scalar value.
553 my $key = $db->first_key();
555 • next_key()
557 Returns the "next" key in the hash, given the previous one as the
559 sole argument. Returns undef if there are no more keys to be
560 fetched.
561 $key = $db->next_key($key);
563 Here are some examples of using hashes:
565 my $db = DBM::Deep->new( "foo.db" );
567 $db->put("foo", "bar");
569 print "foo: " . $db->get("foo") . "\n";
570 $db->put("baz", {}); # new child hash ref
572 $db->get("baz")->put("buz", "biz");
573 print "buz: " . $db->get("baz")->get("buz") . "\n";
574 my $key = $db->first_key();
576 while ($key) {
577 print "$key: " . $db->get($key) . "\n";
578 $key = $db->next_key($key);
579 }
580 if ($db->exists("foo")) { $db->delete("foo"); }
582 Arrays
584 For arrays, DBM::Deep supports all the common methods described above,
585 and the following additional methods: "length()", "push()", "pop()",
586 "shift()", "unshift()" and "splice()".
587 • length()
589 Returns the number of elements in the array. Takes no arguments.
591 my $len = $db->length();
593 • push()
595 Adds one or more elements onto the end of the array. Accepts
597 scalars, hash refs or array refs. No return value.
598 $db->push("foo", "bar", {});
600 • pop()
602 Fetches the last element in the array, and deletes it. Takes no
604 arguments. Returns undef if array is empty. Returns the element
605 value.
606 my $elem = $db->pop();
608 • shift()
610 Fetches the first element in the array, deletes it, then shifts all
612 the remaining elements over to take up the space. Returns the
613 element value. This method is not recommended with large arrays --
614 see "Large Arrays" below for details.
615 my $elem = $db->shift();
617 • unshift()
619 Inserts one or more elements onto the beginning of the array,
621 shifting all existing elements over to make room. Accepts scalars,
622 hash refs or array refs. No return value. This method is not
623 recommended with large arrays -- see <Large Arrays> below for
624 details.
625 $db->unshift("foo", "bar", {});
627 • splice()
629 Performs exactly like Perl's built-in function of the same name.
631 See "splice" in perlfunc for usage -- it is too complicated to
632 document here. This method is not recommended with large arrays --
633 see "Large Arrays" below for details.
634 Here are some examples of using arrays:
636 my $db = DBM::Deep->new(
638 file => "foo.db",
639 type => DBM::Deep->TYPE_ARRAY
640 );
641 $db->push("bar", "baz");
643 $db->unshift("foo");
644 $db->put(3, "buz");
645 my $len = $db->length();
647 print "length: $len\n"; # 4
648 for (my $k=0; $k<$len; $k++) {
650 print "$k: " . $db->get($k) . "\n";
651 }
652 $db->splice(1, 2, "biz", "baf");
654 while (my $elem = shift @$db) {
656 print "shifted: $elem\n";
657 }
658
660 Enable or disable automatic file locking by passing a boolean value to
661 the "locking" parameter when constructing your DBM::Deep object (see
662 "SETUP" above).
663 my $db = DBM::Deep->new(
665 file => "foo.db",
666 locking => 1
667 );
668 This causes DBM::Deep to "flock()" the underlying filehandle with
670 exclusive mode for writes, and shared mode for reads. This is required
671 if you have multiple processes accessing the same database file, to
672 avoid file corruption. Please note that "flock()" does NOT work for
673 files over NFS. See "DB over NFS" below for more.
674 Explicit Locking
676 You can explicitly lock a database, so it remains locked for multiple
677 actions. This is done by calling the "lock_exclusive()" method (for
678 when you want to write) or the "lock_shared()" method (for when you
679 want to read). This is particularly useful for things like counters,
680 where the current value needs to be fetched, then incremented, then
681 stored again.
682 $db->lock_exclusive();
684 my $counter = $db->get("counter");
685 $counter++;
686 $db->put("counter", $counter);
687 $db->unlock();
688 # or...
690 $db->lock_exclusive();
692 $db->{counter}++;
693 $db->unlock();
694 Win32/Cygwin
696 Due to Win32 actually enforcing the read-only status of a shared lock,
697 all locks on Win32 and cygwin are exclusive. This is because of how
698 autovivification currently works. Hopefully, this will go away in a
699 future release.
700
702 You can import existing complex structures by calling the "import()"
703 method, and export an entire database into an in-memory structure using
704 the "export()" method. Both are examined here.
705 Importing
707 Say you have an existing hash with nested hashes/arrays inside it.
708 Instead of walking the structure and adding keys/elements to the
709 database as you go, simply pass a reference to the "import()" method.
710 This recursively adds everything to an existing DBM::Deep object for
711 you. Here is an example:
712 my $struct = {
714 key1 => "value1",
715 key2 => "value2",
716 array1 => [ "elem0", "elem1", "elem2" ],
717 hash1 => {
718 subkey1 => "subvalue1",
719 subkey2 => "subvalue2"
720 }
721 };
722 my $db = DBM::Deep->new( "foo.db" );
724 $db->import( $struct );
725 print $db->{key1} . "\n"; # prints "value1"
727 This recursively imports the entire $struct object into $db, including
729 all nested hashes and arrays. If the DBM::Deep object contains existing
730 data, keys are merged with the existing ones, replacing if they already
731 exist. The "import()" method can be called on any database level (not
732 just the base level), and works with both hash and array DB types.
733 Note: Make sure your existing structure has no circular references in
735 it. These will cause an infinite loop when importing. There are plans
736 to fix this in a later release.
737 Exporting
739 Calling the "export()" method on an existing DBM::Deep object will
740 return a reference to a new in-memory copy of the database. The export
741 is done recursively, so all nested hashes/arrays are all exported to
742 standard Perl objects. Here is an example:
743 my $db = DBM::Deep->new( "foo.db" );
745 $db->{key1} = "value1";
747 $db->{key2} = "value2";
748 $db->{hash1} = {};
749 $db->{hash1}->{subkey1} = "subvalue1";
750 $db->{hash1}->{subkey2} = "subvalue2";
751 my $struct = $db->export();
753 print $struct->{key1} . "\n"; # prints "value1"
755 This makes a complete copy of the database in memory, and returns a
757 reference to it. The "export()" method can be called on any database
758 level (not just the base level), and works with both hash and array DB
759 types. Be careful of large databases -- you can store a lot more data
760 in a DBM::Deep object than an in-memory Perl structure.
761 Note: Make sure your database has no circular references in it. These
763 will cause an infinite loop when exporting. There are plans to fix this
764 in a later release.
765
767 DBM::Deep has a number of hooks where you can specify your own Perl
768 function to perform filtering on incoming or outgoing data. This is a
769 perfect way to extend the engine, and implement things like real-time
770 compression or encryption. Filtering applies to the base DB level, and
771 all child hashes / arrays. Filter hooks can be specified when your
772 DBM::Deep object is first constructed, or by calling the "set_filter()"
773 method at any time. There are four available filter hooks.
774 set_filter()
776 This method takes two parameters - the filter type and the filter
777 subreference. The four types are:
778 • filter_store_key
780 This filter is called whenever a hash key is stored. It is passed
782 the incoming key, and expected to return a transformed key.
783 • filter_store_value
785 This filter is called whenever a hash key or array element is
787 stored. It is passed the incoming value, and expected to return a
788 transformed value.
789 • filter_fetch_key
791 This filter is called whenever a hash key is fetched (i.e. via
793 "first_key()" or "next_key()"). It is passed the transformed key,
794 and expected to return the plain key.
795 • filter_fetch_value
797 This filter is called whenever a hash key or array element is
799 fetched. It is passed the transformed value, and expected to
800 return the plain value.
801 Here are the two ways to setup a filter hook:
803 my $db = DBM::Deep->new(
805 file => "foo.db",
806 filter_store_value => \&my_filter_store,
807 filter_fetch_value => \&my_filter_fetch
808 );
809 # or...
811 $db->set_filter( "store_value", \&my_filter_store );
813 $db->set_filter( "fetch_value", \&my_filter_fetch );
814 Your filter function will be called only when dealing with SCALAR keys
816 or values. When nested hashes and arrays are being stored/fetched,
817 filtering is bypassed. Filters are called as static functions, passed a
818 single SCALAR argument, and expected to return a single SCALAR value.
819 If you want to remove a filter, set the function reference to "undef":
820 $db->set_filter( "store_value", undef );
822 Examples
824 Please read DBM::Deep::Cookbook for examples of filters.
825
827 Most DBM::Deep methods return a true value for success, and call die()
828 on failure. You can wrap calls in an eval block to catch the die.
829 my $db = DBM::Deep->new( "foo.db" ); # create hash
831 eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call
832 print $@; # prints error message
834
836 If you have a 64-bit system, and your Perl is compiled with both
837 LARGEFILE and 64-bit support, you may be able to create databases
838 larger than 4 GB. DBM::Deep by default uses 32-bit file offset tags,
839 but these can be changed by specifying the 'pack_size' parameter when
840 constructing the file.
841 DBM::Deep->new(
843 file => $filename,
844 pack_size => 'large',
845 );
846 This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad
848 words instead of 32-bit longs. After setting these values your DB files
849 have a theoretical maximum size of 16 XB (exabytes).
850 You can also use "pack_size => 'small'" in order to use 16-bit file
852 offsets.
853 Note: Changing these values will NOT work for existing database files.
855 Only change this for new files. Once the value has been set, it is
856 stored in the file's header and cannot be changed for the life of the
857 file. These parameters are per-file, meaning you can access 32-bit and
858 64-bit files, as you choose.
859 Note: We have not personally tested files larger than 4 GB -- all our
861 systems have only a 32-bit Perl. However, we have received user reports
862 that this does indeed work.
863
865 If you require low-level access to the underlying filehandle that
866 DBM::Deep uses, you can call the "_fh()" method, which returns the
867 handle:
868 my $fh = $db->_fh();
870 This method can be called on the root level of the database, or any
872 child hashes or arrays. All levels share a root structure, which
873 contains things like the filehandle, a reference counter, and all the
874 options specified when you created the object. You can get access to
875 this file object by calling the "_storage()" method.
876 my $file_obj = $db->_storage();
878 This is useful for changing options after the object has already been
880 created, such as enabling/disabling locking. You can also store your
881 own temporary user data in this structure (be wary of name collision),
882 which is then accessible from any child hash or array.
883
885 DBM::Deep has full support for circular references. Meaning you can
886 have a nested hash key or array element that points to a parent object.
887 This relationship is stored in the DB file, and is preserved between
888 sessions. Here is an example:
889 my $db = DBM::Deep->new( "foo.db" );
891 $db->{foo} = "bar";
893 $db->{circle} = $db; # ref to self
894 print $db->{foo} . "\n"; # prints "bar"
896 print $db->{circle}->{foo} . "\n"; # prints "bar" again
897 This also works as expected with array and hash references. So, the
899 following works as expected:
900 $db->{foo} = [ 1 .. 3 ];
902 $db->{bar} = $db->{foo};
903 push @{$db->{foo}}, 42;
905 is( $db->{bar}[-1], 42 ); # Passes
906 This, however, does not extend to assignments from one DB file to
908 another. So, the following will throw an error:
909 my $db1 = DBM::Deep->new( "foo.db" );
911 my $db2 = DBM::Deep->new( "bar.db" );
912 $db1->{foo} = [];
914 $db2->{foo} = $db1->{foo}; # dies
915 Note: Passing the object to a function that recursively walks the
917 object tree (such as Data::Dumper or even the built-in "optimize()" or
918 "export()" methods) will result in an infinite loop. This will be fixed
919 in a future release by adding singleton support.
920
922 As of 1.0000, DBM::Deep has ACID transactions. Every DBM::Deep object
923 is completely transaction-ready - it is not an option you have to turn
924 on. You do have to specify how many transactions may run simultaneously
925 (q.v. "num_txns").
926 Three new methods have been added to support them. They are:
928 • begin_work()
930 This starts a transaction.
932 • commit()
934 This applies the changes done within the transaction to the
936 mainline and ends the transaction.
937 • rollback()
939 This discards the changes done within the transaction to the
941 mainline and ends the transaction.
942 Transactions in DBM::Deep are done using a variant of the MVCC method,
944 the same method used by the InnoDB MySQL engine.
945
947 As of 1.0000, the file format has changed. To aid in upgrades, a
948 migration script is provided within the CPAN distribution, called
949 utils/upgrade_db.pl.
950 NOTE: This script is not installed onto your system because it carries
952 a copy of every version prior to the current version.
953 As of version 2.0000, databases created by old versions back to 1.0003
955 can be read, but new features may not be available unless the database
956 is upgraded first.
957
959 The following are items that are planned to be added in future
960 releases. These are separate from the "CAVEATS, ISSUES & BUGS" below.
961 Sub-Transactions
963 Right now, you cannot run a transaction within a transaction. Removing
964 this restriction is technically straightforward, but the combinatorial
965 explosion of possible usecases hurts my head. If this is something you
966 want to see immediately, please submit many testcases.
967 Caching
969 If a client is willing to assert upon opening the file that this
970 process will be the only consumer of that datafile, then there are a
971 number of caching possibilities that can be taken advantage of. This
972 does, however, mean that DBM::Deep is more vulnerable to losing data
973 due to unflushed changes. It also means a much larger in-memory
974 footprint. As such, it's not clear exactly how this should be done.
975 Suggestions are welcome.
976 Ram-only
978 The techniques used in DBM::Deep simply require a seekable contiguous
979 datastore. This could just as easily be a large string as a file. By
980 using substr, the STM capabilities of DBM::Deep could be used within a
981 single-process. I have no idea how I'd specify this, though.
982 Suggestions are welcome.
983 Different contention resolution mechanisms
985 Currently, the only contention resolution mechanism is last-write-wins.
986 This is the mechanism used by most RDBMSes and should be good enough
987 for most uses. For advanced uses of STM, other contention mechanisms
988 will be needed. If you have an idea of how you'd like to see contention
989 resolution in DBM::Deep, please let me know.
990
992 This section describes all the known issues with DBM::Deep. These are
993 issues that are either intractable or depend on some feature within
994 Perl working exactly right. It you have found something that is not
995 listed below, please send an e-mail to bug-DBM-Deep@rt.cpan.org
996 <mailto:bug-DBM-Deep@rt.cpan.org>. Likewise, if you think you know of
997 a way around one of these issues, please let me know.
998 References
1000 (The following assumes a high level of Perl understanding, specifically
1001 of references. Most users can safely skip this section.)
1002 Currently, the only references supported are HASH and ARRAY. The other
1004 reference types (SCALAR, CODE, GLOB, and REF) cannot be supported for
1005 various reasons.
1006 • GLOB
1008 These are things like filehandles and other sockets. They can't be
1010 supported because it's completely unclear how DBM::Deep should
1011 serialize them.
1012 • SCALAR / REF
1014 The discussion here refers to the following type of example:
1016 my $x = 25;
1018 $db->{key1} = \$x;
1019 $x = 50;
1021 # In some other process ...
1023 my $val = ${ $db->{key1} };
1025 is( $val, 50, "What actually gets stored in the DB file?" );
1027 The problem is one of synchronization. When the variable being
1029 referred to changes value, the reference isn't notified, which is
1030 kind of the point of references. This means that the new value
1031 won't be stored in the datafile for other processes to read. There
1032 is no TIEREF.
1033 It is theoretically possible to store references to values already
1035 within a DBM::Deep object because everything already is
1036 synchronized, but the change to the internals would be quite large.
1037 Specifically, DBM::Deep would have to tie every single value that
1038 is stored. This would bloat the RAM footprint of DBM::Deep at least
1039 twofold (if not more) and be a significant performance drain, all
1040 to support a feature that has never been requested.
1041 • CODE
1043 Data::Dump::Streamer provides a mechanism for serializing coderefs,
1045 including saving off all closure state. This would allow for
1046 DBM::Deep to store the code for a subroutine. Then, whenever the
1047 subroutine is read, the code could be "eval()"'ed into being.
1048 However, just as for SCALAR and REF, that closure state may change
1049 without notifying the DBM::Deep object storing the reference.
1050 Again, this would generally be considered a feature.
1051 External references and transactions
1053 If you do "my $x = $db->{foo};", then start a transaction, $x will be
1054 referencing the database from outside the transaction. A fix for this
1055 (and other issues with how external references into the database) is
1056 being looked into. This is the skipped set of tests in
1057 t/39_singletons.t and a related issue is the focus of
1058 t/37_delete_edge_cases.t
1059 File corruption
1061 The current level of error handling in DBM::Deep is minimal. Files are
1062 checked for a 32-bit signature when opened, but any other form of
1063 corruption in the datafile can cause segmentation faults. DBM::Deep may
1064 try to "seek()" past the end of a file, or get stuck in an infinite
1065 loop depending on the level and type of corruption. File write
1066 operations are not checked for failure (for speed), so if you happen to
1067 run out of disk space, DBM::Deep will probably fail in a bad way. These
1068 things will be addressed in a later version of DBM::Deep.
1069 DB over NFS
1071 Beware of using DBM::Deep files over NFS. DBM::Deep uses flock(), which
1072 works well on local filesystems, but will NOT protect you from file
1073 corruption over NFS. I've heard about setting up your NFS server with a
1074 locking daemon, then using "lockf()" to lock your files, but your
1075 mileage may vary there as well. From what I understand, there is no
1076 real way to do it. However, if you need access to the underlying
1077 filehandle in DBM::Deep for using some other kind of locking scheme
1078 like "lockf()", see the "LOW-LEVEL ACCESS" section above.
1079 Copying Objects
1081 Beware of copying tied objects in Perl. Very strange things can happen.
1082 Instead, use DBM::Deep's "clone()" method which safely copies the
1083 object and returns a new, blessed and tied hash or array to the same
1084 level in the DB.
1085 my $copy = $db->clone();
1087 Note: Since clone() here is cloning the object, not the database
1089 location, any modifications to either $db or $copy will be visible to
1090 both.
1091 Stale References
1093 If you take a reference to an array or hash from the database, it is
1094 tied to the database itself. This means that if the datum in question
1095 is subsequently deleted from the database, the reference to it will
1096 point to an invalid location and unpredictable things will happen if
1097 you try to use it.
1098 So a seemingly innocuous piece of code like this:
1100 my %hash = %{ $db->{some_hash} };
1102 can fail if another process deletes or clobbers "$db->{some_hash}"
1104 while the data are being extracted, since "%{ ... }" is not atomic.
1105 (This actually happened.) The solution is to lock the database before
1106 reading the data:
1107 $db->lock_exclusive;
1109 my %hash = %{ $db->{some_hash} };
1110 $db->unlock;
1111 As of version 1.0024, if you assign a stale reference to a location in
1113 the database, DBM::Deep will warn, if you have uninitialized warnings
1114 enabled, and treat the stale reference as "undef". An attempt to use a
1115 stale reference as an array or hash reference will cause an error.
1116 Large Arrays
1118 Beware of using "shift()", "unshift()" or "splice()" with large arrays.
1119 These functions cause every element in the array to move, which can be
1120 murder on DBM::Deep, as every element has to be fetched from disk, then
1121 stored again in a different location. This will be addressed in a
1122 future version.
1123 This has been somewhat addressed so that the cost is constant,
1125 regardless of what is stored at those locations. So, small arrays with
1126 huge data structures in them are faster. But, large arrays are still
1127 large.
1128 Writeonly Files
1130 If you pass in a filehandle to new(), you may have opened it in either
1131 a readonly or writeonly mode. STORE will verify that the filehandle is
1132 writable. However, there doesn't seem to be a good way to determine if
1133 a filehandle is readable. And, if the filehandle isn't readable, it's
1134 not clear what will happen. So, don't do that.
1135 Assignments Within Transactions
1137 The following will not work as one might expect:
1138 my $x = { a => 1 };
1140 $db->begin_work;
1142 $db->{foo} = $x;
1143 $db->rollback;
1144 is( $x->{a}, 1 ); # This will fail!
1146 The problem is that the moment a reference used as the rvalue to a
1148 DBM::Deep object's lvalue, it becomes tied itself. This is so that
1149 future changes to $x can be tracked within the DBM::Deep file and is
1150 considered to be a feature. By the time the rollback occurs, there is
1151 no knowledge that there had been an $x or what memory location to
1152 assign an "export()" to.
1153 NOTE: This does not affect importing because imports do a walk over the
1155 reference to be imported in order to explicitly leave it untied.
1156
1158 Devel::Cover is used to test the code coverage of the tests. Below is
1159 the Devel::Cover report on this distribution's test suite.
1160 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1162 File stmt bran cond sub pod time total
1163 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1164 blib/lib/DBM/Deep.pm 100.0 89.1 82.9 100.0 100.0 32.5 98.1
1165 blib/lib/DBM/Deep/Array.pm 100.0 94.4 100.0 100.0 100.0 5.2 98.8
1166 blib/lib/DBM/Deep/Engine.pm 100.0 92.9 100.0 100.0 100.0 7.4 100.0
1167 ...ib/DBM/Deep/Engine/DBI.pm 95.0 73.1 100.0 100.0 100.0 1.5 90.4
1168 ...b/DBM/Deep/Engine/File.pm 92.3 78.5 88.9 100.0 100.0 4.9 90.3
1169 blib/lib/DBM/Deep/Hash.pm 100.0 100.0 100.0 100.0 100.0 3.8 100.0
1170 .../lib/DBM/Deep/Iterator.pm 100.0 n/a n/a 100.0 100.0 0.0 100.0
1171 .../DBM/Deep/Iterator/DBI.pm 100.0 100.0 n/a 100.0 100.0 1.2 100.0
1172 ...DBM/Deep/Iterator/File.pm 92.5 84.6 n/a 100.0 66.7 0.6 90.0
1173 ...erator/File/BucketList.pm 100.0 75.0 n/a 100.0 66.7 0.4 93.8
1174 ...ep/Iterator/File/Index.pm 100.0 100.0 n/a 100.0 100.0 0.2 100.0
1175 blib/lib/DBM/Deep/Null.pm 87.5 n/a n/a 75.0 n/a 0.0 83.3
1176 blib/lib/DBM/Deep/Sector.pm 91.7 n/a n/a 83.3 0.0 6.7 74.4
1177 ...ib/DBM/Deep/Sector/DBI.pm 96.8 83.3 n/a 100.0 0.0 1.0 89.8
1178 ...p/Sector/DBI/Reference.pm 100.0 95.5 100.0 100.0 0.0 2.2 91.2
1179 ...Deep/Sector/DBI/Scalar.pm 100.0 100.0 n/a 100.0 0.0 1.1 92.9
1180 ...b/DBM/Deep/Sector/File.pm 96.0 87.5 100.0 92.3 25.0 2.2 91.0
1181 ...Sector/File/BucketList.pm 98.2 85.7 83.3 100.0 0.0 3.3 89.4
1182 .../Deep/Sector/File/Data.pm 100.0 n/a n/a 100.0 0.0 0.1 90.9
1183 ...Deep/Sector/File/Index.pm 100.0 80.0 33.3 100.0 0.0 0.8 83.1
1184 .../Deep/Sector/File/Null.pm 100.0 100.0 n/a 100.0 0.0 0.0 91.7
1185 .../Sector/File/Reference.pm 100.0 90.0 80.0 100.0 0.0 1.4 91.5
1186 ...eep/Sector/File/Scalar.pm 98.4 87.5 n/a 100.0 0.0 0.8 91.9
1187 blib/lib/DBM/Deep/Storage.pm 100.0 n/a n/a 100.0 100.0 0.0 100.0
1188 ...b/DBM/Deep/Storage/DBI.pm 97.3 70.8 n/a 100.0 38.5 6.7 87.0
1189 .../DBM/Deep/Storage/File.pm 96.6 77.1 80.0 95.7 100.0 16.0 91.8
1190 Total 99.3 85.2 84.9 99.8 63.3 100.0 97.6
1191 ---------------------------- ------ ------ ------ ------ ------ ------ ------
1192
1194 Check out the DBM::Deep Google Group at
1195 <http://groups.google.com/group/DBM-Deep> or send email to
1196 DBM-Deep@googlegroups.com <mailto:DBM-Deep@googlegroups.com>. You can
1197 also visit #dbm-deep on irc.perl.org
1198 The source code repository is at <http://github.com/robkinyon/dbm-deep>
1200
1202 Rob Kinyon, rkinyon@cpan.org <mailto:rkinyon@cpan.org>
1203 Originally written by Joseph Huckaby, jhuckaby@cpan.org
1205 <mailto:jhuckaby@cpan.org>
1206
1208 Stonehenge Consulting (<http://www.stonehenge.com/>) sponsored the
1209 development of transactions and freespace management, leading to the
1210 1.0000 release. A great debt of gratitude goes out to them for their
1211 continuing leadership in and support of the Perl community.
1212
1214 The following have contributed greatly to make DBM::Deep what it is
1215 today:
1216 • Adam Sah and Rich Gaushell for innumerable contributions early on.
1218 • Dan Golden and others at YAPC::NA 2006 for helping me design
1220 through transactions.
1221 • James Stanley for bug fix
1223 • David Steinbrunner for fixing typos and adding repository cpan
1225 metadata
1226 • H. Merijn Brandt for fixing the POD escapes.
1228 • Breno G. de Oliveira for minor packaging tweaks
1230
1232 DBM::Deep::Cookbook(3)
1233 perltie(1), Tie::Hash(3), Fcntl(3), flock(2), lockf(3), nfs(5)
1235
1237 Copyright (c) 2007-14 Rob Kinyon. All Rights Reserved. This is free
1238 software, you may use it and distribute it under the same terms as Perl
1239 itself.
1240perl v5.32.1 2021-01-27 DBM::Deep(3)