Commit | Line | Data |
a0d0e21e |
1 | # DB_File.pm -- Perl 5 interface to Berkeley DB |
2 | # |
6ca2e664 |
3 | # written by Paul Marquess (Paul.Marquess@btinternet.com) |
c5da4faf |
4 | # last modified 26th April 2001 |
5 | # version 1.77 |
36477c24 |
6 | # |
c5da4faf |
7 | # Copyright (c) 1995-2001 Paul Marquess. All rights reserved. |
36477c24 |
8 | # This program is free software; you can redistribute it and/or |
9 | # modify it under the same terms as Perl itself. |
10 | |
8e07c86e |
11 | |
12 | package DB_File::HASHINFO ; |
785da04d |
13 | |
610ab055 |
14 | require 5.003 ; |
15 | |
3245f058 |
16 | use warnings; |
785da04d |
17 | use strict; |
8e07c86e |
18 | use Carp; |
88108326 |
19 | require Tie::Hash; |
20 | @DB_File::HASHINFO::ISA = qw(Tie::Hash); |
8e07c86e |
21 | |
88108326 |
22 | sub new |
8e07c86e |
23 | { |
88108326 |
24 | my $pkg = shift ; |
25 | my %x ; |
26 | tie %x, $pkg ; |
27 | bless \%x, $pkg ; |
8e07c86e |
28 | } |
29 | |
610ab055 |
30 | |
88108326 |
31 | sub TIEHASH |
32 | { |
33 | my $pkg = shift ; |
34 | |
36477c24 |
35 | bless { VALID => { map {$_, 1} |
36 | qw( bsize ffactor nelem cachesize hash lorder) |
37 | }, |
38 | GOT => {} |
39 | }, $pkg ; |
88108326 |
40 | } |
8e07c86e |
41 | |
610ab055 |
42 | |
8e07c86e |
43 | sub FETCH |
44 | { |
88108326 |
45 | my $self = shift ; |
46 | my $key = shift ; |
8e07c86e |
47 | |
36477c24 |
48 | return $self->{GOT}{$key} if exists $self->{VALID}{$key} ; |
88108326 |
49 | |
50 | my $pkg = ref $self ; |
51 | croak "${pkg}::FETCH - Unknown element '$key'" ; |
8e07c86e |
52 | } |
53 | |
54 | |
55 | sub STORE |
56 | { |
88108326 |
57 | my $self = shift ; |
58 | my $key = shift ; |
59 | my $value = shift ; |
60 | |
36477c24 |
61 | if ( exists $self->{VALID}{$key} ) |
8e07c86e |
62 | { |
36477c24 |
63 | $self->{GOT}{$key} = $value ; |
8e07c86e |
64 | return ; |
65 | } |
66 | |
88108326 |
67 | my $pkg = ref $self ; |
68 | croak "${pkg}::STORE - Unknown element '$key'" ; |
8e07c86e |
69 | } |
70 | |
71 | sub DELETE |
72 | { |
88108326 |
73 | my $self = shift ; |
74 | my $key = shift ; |
75 | |
36477c24 |
76 | if ( exists $self->{VALID}{$key} ) |
8e07c86e |
77 | { |
36477c24 |
78 | delete $self->{GOT}{$key} ; |
8e07c86e |
79 | return ; |
80 | } |
81 | |
88108326 |
82 | my $pkg = ref $self ; |
83 | croak "DB_File::HASHINFO::DELETE - Unknown element '$key'" ; |
8e07c86e |
84 | } |
85 | |
88108326 |
86 | sub EXISTS |
8e07c86e |
87 | { |
88108326 |
88 | my $self = shift ; |
89 | my $key = shift ; |
8e07c86e |
90 | |
36477c24 |
91 | exists $self->{VALID}{$key} ; |
8e07c86e |
92 | } |
93 | |
88108326 |
94 | sub NotHere |
8e07c86e |
95 | { |
18d2dc8c |
96 | my $self = shift ; |
88108326 |
97 | my $method = shift ; |
8e07c86e |
98 | |
18d2dc8c |
99 | croak ref($self) . " does not define the method ${method}" ; |
8e07c86e |
100 | } |
101 | |
18d2dc8c |
102 | sub FIRSTKEY { my $self = shift ; $self->NotHere("FIRSTKEY") } |
103 | sub NEXTKEY { my $self = shift ; $self->NotHere("NEXTKEY") } |
104 | sub CLEAR { my $self = shift ; $self->NotHere("CLEAR") } |
8e07c86e |
105 | |
106 | package DB_File::RECNOINFO ; |
785da04d |
107 | |
3245f058 |
108 | use warnings; |
88108326 |
109 | use strict ; |
110 | |
045291aa |
111 | @DB_File::RECNOINFO::ISA = qw(DB_File::HASHINFO) ; |
8e07c86e |
112 | |
113 | sub TIEHASH |
114 | { |
88108326 |
115 | my $pkg = shift ; |
116 | |
36477c24 |
117 | bless { VALID => { map {$_, 1} |
118 | qw( bval cachesize psize flags lorder reclen bfname ) |
119 | }, |
120 | GOT => {}, |
121 | }, $pkg ; |
8e07c86e |
122 | } |
123 | |
88108326 |
124 | package DB_File::BTREEINFO ; |
8e07c86e |
125 | |
3245f058 |
126 | use warnings; |
88108326 |
127 | use strict ; |
8e07c86e |
128 | |
88108326 |
129 | @DB_File::BTREEINFO::ISA = qw(DB_File::HASHINFO) ; |
8e07c86e |
130 | |
88108326 |
131 | sub TIEHASH |
8e07c86e |
132 | { |
88108326 |
133 | my $pkg = shift ; |
134 | |
36477c24 |
135 | bless { VALID => { map {$_, 1} |
136 | qw( flags cachesize maxkeypage minkeypage psize |
137 | compare prefix lorder ) |
138 | }, |
139 | GOT => {}, |
140 | }, $pkg ; |
8e07c86e |
141 | } |
142 | |
143 | |
8e07c86e |
144 | package DB_File ; |
785da04d |
145 | |
3245f058 |
146 | use warnings; |
785da04d |
147 | use strict; |
b90e71be |
148 | use vars qw($VERSION @ISA @EXPORT $AUTOLOAD $DB_BTREE $DB_HASH $DB_RECNO |
149 | $db_version $use_XSLoader |
150 | ) ; |
8e07c86e |
151 | use Carp; |
152 | |
785da04d |
153 | |
c5da4faf |
154 | $VERSION = "1.77" ; |
8e07c86e |
155 | |
156 | #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE; |
88108326 |
157 | $DB_BTREE = new DB_File::BTREEINFO ; |
158 | $DB_HASH = new DB_File::HASHINFO ; |
159 | $DB_RECNO = new DB_File::RECNOINFO ; |
8e07c86e |
160 | |
785da04d |
161 | require Tie::Hash; |
8e07c86e |
162 | require Exporter; |
163 | use AutoLoader; |
b90e71be |
164 | BEGIN { |
165 | $use_XSLoader = 1 ; |
166 | eval { require XSLoader } ; |
167 | |
168 | if ($@) { |
169 | $use_XSLoader = 0 ; |
170 | require DynaLoader; |
171 | @ISA = qw(DynaLoader); |
172 | } |
173 | } |
174 | |
175 | push @ISA, qw(Tie::Hash Exporter); |
8e07c86e |
176 | @EXPORT = qw( |
177 | $DB_BTREE $DB_HASH $DB_RECNO |
88108326 |
178 | |
8e07c86e |
179 | BTREEMAGIC |
180 | BTREEVERSION |
181 | DB_LOCK |
182 | DB_SHMEM |
183 | DB_TXN |
184 | HASHMAGIC |
185 | HASHVERSION |
186 | MAX_PAGE_NUMBER |
187 | MAX_PAGE_OFFSET |
188 | MAX_REC_NUMBER |
189 | RET_ERROR |
190 | RET_SPECIAL |
191 | RET_SUCCESS |
192 | R_CURSOR |
193 | R_DUP |
194 | R_FIRST |
195 | R_FIXEDLEN |
196 | R_IAFTER |
197 | R_IBEFORE |
198 | R_LAST |
199 | R_NEXT |
200 | R_NOKEY |
201 | R_NOOVERWRITE |
202 | R_PREV |
203 | R_RECNOSYNC |
204 | R_SETCURSOR |
205 | R_SNAPSHOT |
206 | __R_UNUSED |
88108326 |
207 | |
045291aa |
208 | ); |
8e07c86e |
209 | |
210 | sub AUTOLOAD { |
785da04d |
211 | my($constname); |
8e07c86e |
212 | ($constname = $AUTOLOAD) =~ s/.*:://; |
785da04d |
213 | my $val = constant($constname, @_ ? $_[0] : 0); |
8e07c86e |
214 | if ($! != 0) { |
265f5c4a |
215 | if ($! =~ /Invalid/ || $!{EINVAL}) { |
8e07c86e |
216 | $AutoLoader::AUTOLOAD = $AUTOLOAD; |
217 | goto &AutoLoader::AUTOLOAD; |
218 | } |
219 | else { |
785da04d |
220 | my($pack,$file,$line) = caller; |
8e07c86e |
221 | croak "Your vendor has not defined DB macro $constname, used at $file line $line. |
222 | "; |
223 | } |
224 | } |
225 | eval "sub $AUTOLOAD { $val }"; |
226 | goto &$AUTOLOAD; |
227 | } |
228 | |
f6b705ef |
229 | |
a6ed719b |
230 | eval { |
1f70e1ea |
231 | # Make all Fcntl O_XXX constants available for importing |
232 | require Fcntl; |
233 | my @O = grep /^O_/, @Fcntl::EXPORT; |
234 | Fcntl->import(@O); # first we import what we want to export |
235 | push(@EXPORT, @O); |
a6ed719b |
236 | }; |
f6b705ef |
237 | |
b90e71be |
238 | if ($use_XSLoader) |
239 | { XSLoader::load("DB_File", $VERSION)} |
240 | else |
241 | { bootstrap DB_File $VERSION } |
8e07c86e |
242 | |
243 | # Preloaded methods go here. Autoload methods go after __END__, and are |
244 | # processed by the autosplit program. |
245 | |
05475680 |
246 | sub tie_hash_or_array |
610ab055 |
247 | { |
248 | my (@arg) = @_ ; |
05475680 |
249 | my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ; |
610ab055 |
250 | |
251 | $arg[4] = tied %{ $arg[4] } |
252 | if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ; |
253 | |
1f70e1ea |
254 | # make recno in Berkeley DB version 2 work like recno in version 1. |
255 | if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and |
256 | $arg[1] and ! -e $arg[1]) { |
257 | open(FH, ">$arg[1]") or return undef ; |
258 | close FH ; |
259 | chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ; |
260 | } |
261 | |
05475680 |
262 | DoTie_($tieHASH, @arg) ; |
610ab055 |
263 | } |
264 | |
05475680 |
265 | sub TIEHASH |
266 | { |
267 | tie_hash_or_array(@_) ; |
268 | } |
269 | |
270 | sub TIEARRAY |
271 | { |
272 | tie_hash_or_array(@_) ; |
273 | } |
88108326 |
274 | |
045291aa |
275 | sub CLEAR |
276 | { |
1f70e1ea |
277 | my $self = shift; |
3245f058 |
278 | my $key = 0 ; |
1f70e1ea |
279 | my $value = "" ; |
280 | my $status = $self->seq($key, $value, R_FIRST()); |
281 | my @keys; |
282 | |
283 | while ($status == 0) { |
284 | push @keys, $key; |
285 | $status = $self->seq($key, $value, R_NEXT()); |
286 | } |
287 | foreach $key (reverse @keys) { |
288 | my $s = $self->del($key); |
289 | } |
290 | } |
291 | |
045291aa |
292 | sub EXTEND { } |
293 | |
294 | sub STORESIZE |
295 | { |
296 | my $self = shift; |
297 | my $length = shift ; |
298 | my $current_length = $self->length() ; |
299 | |
300 | if ($length < $current_length) { |
301 | my $key ; |
302 | for ($key = $current_length - 1 ; $key >= $length ; -- $key) |
303 | { $self->del($key) } |
304 | } |
a9fd575d |
305 | elsif ($length > $current_length) { |
306 | $self->put($length-1, "") ; |
307 | } |
045291aa |
308 | } |
309 | |
c5da4faf |
310 | |
311 | sub SPLICE |
312 | { |
313 | my $self = shift; |
314 | my $offset = shift; |
315 | if (not defined $offset) { |
316 | carp 'Use of uninitialized value in splice'; |
317 | $offset = 0; |
318 | } |
319 | |
320 | my $length = @_ ? shift : 0; |
321 | # Carping about definedness comes _after_ the OFFSET sanity check. |
322 | # This is so we get the same error messages as Perl's splice(). |
323 | # |
324 | |
325 | my @list = @_; |
326 | |
327 | my $size = $self->FETCHSIZE(); |
328 | |
329 | # 'If OFFSET is negative then it start that far from the end of |
330 | # the array.' |
331 | # |
332 | if ($offset < 0) { |
333 | my $new_offset = $size + $offset; |
334 | if ($new_offset < 0) { |
335 | die "Modification of non-creatable array value attempted, " |
336 | . "subscript $offset"; |
337 | } |
338 | $offset = $new_offset; |
339 | } |
340 | |
341 | if ($offset > $size) { |
342 | $offset = $size; |
343 | } |
344 | |
345 | if (not defined $length) { |
346 | carp 'Use of uninitialized value in splice'; |
347 | $length = 0; |
348 | } |
349 | |
350 | # 'If LENGTH is omitted, removes everything from OFFSET onward.' |
351 | if (not defined $length) { |
352 | $length = $size - $offset; |
353 | } |
354 | |
355 | # 'If LENGTH is negative, leave that many elements off the end of |
356 | # the array.' |
357 | # |
358 | if ($length < 0) { |
359 | $length = $size - $offset + $length; |
360 | |
361 | if ($length < 0) { |
362 | # The user must have specified a length bigger than the |
363 | # length of the array passed in. But perl's splice() |
364 | # doesn't catch this, it just behaves as for length=0. |
365 | # |
366 | $length = 0; |
367 | } |
368 | } |
369 | |
370 | if ($length > $size - $offset) { |
371 | $length = $size - $offset; |
372 | } |
373 | |
374 | # $num_elems holds the current number of elements in the database. |
375 | my $num_elems = $size; |
376 | |
377 | # 'Removes the elements designated by OFFSET and LENGTH from an |
378 | # array,'... |
379 | # |
380 | my @removed = (); |
381 | foreach (0 .. $length - 1) { |
382 | my $old; |
383 | my $status = $self->get($offset, $old); |
384 | if ($status != 0) { |
385 | my $msg = "error from Berkeley DB on get($offset, \$old)"; |
386 | if ($status == 1) { |
387 | $msg .= ' (no such element?)'; |
388 | } |
389 | else { |
390 | $msg .= ": error status $status"; |
391 | if (defined $! and $! ne '') { |
392 | $msg .= ", message $!"; |
393 | } |
394 | } |
395 | die $msg; |
396 | } |
397 | push @removed, $old; |
398 | |
399 | $status = $self->del($offset); |
400 | if ($status != 0) { |
401 | my $msg = "error from Berkeley DB on del($offset)"; |
402 | if ($status == 1) { |
403 | $msg .= ' (no such element?)'; |
404 | } |
405 | else { |
406 | $msg .= ": error status $status"; |
407 | if (defined $! and $! ne '') { |
408 | $msg .= ", message $!"; |
409 | } |
410 | } |
411 | die $msg; |
412 | } |
413 | |
414 | -- $num_elems; |
415 | } |
416 | |
417 | # ...'and replaces them with the elements of LIST, if any.' |
418 | my $pos = $offset; |
419 | while (defined (my $elem = shift @list)) { |
420 | my $old_pos = $pos; |
421 | my $status; |
422 | if ($pos >= $num_elems) { |
423 | $status = $self->put($pos, $elem); |
424 | } |
425 | else { |
426 | $status = $self->put($pos, $elem, $self->R_IBEFORE); |
427 | } |
428 | |
429 | if ($status != 0) { |
430 | my $msg = "error from Berkeley DB on put($pos, $elem, ...)"; |
431 | if ($status == 1) { |
432 | $msg .= ' (no such element?)'; |
433 | } |
434 | else { |
435 | $msg .= ", error status $status"; |
436 | if (defined $! and $! ne '') { |
437 | $msg .= ", message $!"; |
438 | } |
439 | } |
440 | die $msg; |
441 | } |
442 | |
443 | die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE" |
444 | if $old_pos != $pos; |
445 | |
446 | ++ $pos; |
447 | ++ $num_elems; |
448 | } |
449 | |
450 | if (wantarray) { |
451 | # 'In list context, returns the elements removed from the |
452 | # array.' |
453 | # |
454 | return @removed; |
455 | } |
456 | elsif (defined wantarray and not wantarray) { |
457 | # 'In scalar context, returns the last element removed, or |
458 | # undef if no elements are removed.' |
459 | # |
460 | if (@removed) { |
461 | my $last = pop @removed; |
462 | return "$last"; |
463 | } |
464 | else { |
465 | return undef; |
466 | } |
467 | } |
468 | elsif (not defined wantarray) { |
469 | # Void context |
470 | } |
471 | else { die } |
472 | } |
473 | sub ::DB_File::splice { &SPLICE } |
474 | |
6ca2e664 |
475 | sub find_dup |
476 | { |
477 | croak "Usage: \$db->find_dup(key,value)\n" |
478 | unless @_ == 3 ; |
479 | |
480 | my $db = shift ; |
481 | my ($origkey, $value_wanted) = @_ ; |
482 | my ($key, $value) = ($origkey, 0); |
483 | my ($status) = 0 ; |
484 | |
485 | for ($status = $db->seq($key, $value, R_CURSOR() ) ; |
486 | $status == 0 ; |
487 | $status = $db->seq($key, $value, R_NEXT() ) ) { |
488 | |
489 | return 0 if $key eq $origkey and $value eq $value_wanted ; |
490 | } |
491 | |
492 | return $status ; |
493 | } |
494 | |
495 | sub del_dup |
496 | { |
497 | croak "Usage: \$db->del_dup(key,value)\n" |
498 | unless @_ == 3 ; |
499 | |
500 | my $db = shift ; |
501 | my ($key, $value) = @_ ; |
502 | my ($status) = $db->find_dup($key, $value) ; |
503 | return $status if $status != 0 ; |
504 | |
505 | $status = $db->del($key, R_CURSOR() ) ; |
506 | return $status ; |
507 | } |
508 | |
88108326 |
509 | sub get_dup |
510 | { |
511 | croak "Usage: \$db->get_dup(key [,flag])\n" |
512 | unless @_ == 2 or @_ == 3 ; |
513 | |
514 | my $db = shift ; |
515 | my $key = shift ; |
516 | my $flag = shift ; |
f6b705ef |
517 | my $value = 0 ; |
88108326 |
518 | my $origkey = $key ; |
519 | my $wantarray = wantarray ; |
f6b705ef |
520 | my %values = () ; |
88108326 |
521 | my @values = () ; |
522 | my $counter = 0 ; |
f6b705ef |
523 | my $status = 0 ; |
88108326 |
524 | |
f6b705ef |
525 | # iterate through the database until either EOF ($status == 0) |
526 | # or a different key is encountered ($key ne $origkey). |
527 | for ($status = $db->seq($key, $value, R_CURSOR()) ; |
528 | $status == 0 and $key eq $origkey ; |
529 | $status = $db->seq($key, $value, R_NEXT()) ) { |
88108326 |
530 | |
f6b705ef |
531 | # save the value or count number of matches |
532 | if ($wantarray) { |
533 | if ($flag) |
534 | { ++ $values{$value} } |
535 | else |
536 | { push (@values, $value) } |
537 | } |
538 | else |
539 | { ++ $counter } |
88108326 |
540 | |
88108326 |
541 | } |
542 | |
f6b705ef |
543 | return ($wantarray ? ($flag ? %values : @values) : $counter) ; |
88108326 |
544 | } |
545 | |
546 | |
8e07c86e |
547 | 1; |
548 | __END__ |
549 | |
3b35bae3 |
550 | =head1 NAME |
551 | |
1f70e1ea |
552 | DB_File - Perl5 access to Berkeley DB version 1.x |
3b35bae3 |
553 | |
554 | =head1 SYNOPSIS |
555 | |
bbc7dcd2 |
556 | use DB_File; |
557 | |
88108326 |
558 | [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ; |
559 | [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ; |
560 | [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ; |
760ac839 |
561 | |
3b35bae3 |
562 | $status = $X->del($key [, $flags]) ; |
563 | $status = $X->put($key, $value [, $flags]) ; |
564 | $status = $X->get($key, $value [, $flags]) ; |
760ac839 |
565 | $status = $X->seq($key, $value, $flags) ; |
3b35bae3 |
566 | $status = $X->sync([$flags]) ; |
567 | $status = $X->fd ; |
760ac839 |
568 | |
f6b705ef |
569 | # BTREE only |
88108326 |
570 | $count = $X->get_dup($key) ; |
571 | @list = $X->get_dup($key) ; |
572 | %list = $X->get_dup($key, 1) ; |
6ca2e664 |
573 | $status = $X->find_dup($key, $value) ; |
574 | $status = $X->del_dup($key, $value) ; |
88108326 |
575 | |
f6b705ef |
576 | # RECNO only |
577 | $a = $X->length; |
578 | $a = $X->pop ; |
579 | $X->push(list); |
580 | $a = $X->shift; |
581 | $X->unshift(list); |
c5da4faf |
582 | @r = $X->splice(offset, length, elements); |
f6b705ef |
583 | |
cad2e5aa |
584 | # DBM Filters |
585 | $old_filter = $db->filter_store_key ( sub { ... } ) ; |
586 | $old_filter = $db->filter_store_value( sub { ... } ) ; |
587 | $old_filter = $db->filter_fetch_key ( sub { ... } ) ; |
588 | $old_filter = $db->filter_fetch_value( sub { ... } ) ; |
589 | |
3b35bae3 |
590 | untie %hash ; |
591 | untie @array ; |
592 | |
593 | =head1 DESCRIPTION |
594 | |
8e07c86e |
595 | B<DB_File> is a module which allows Perl programs to make use of the |
1f70e1ea |
596 | facilities provided by Berkeley DB version 1.x (if you have a newer |
039d031f |
597 | version of DB, see L<Using DB_File with Berkeley DB version 2 or 3>). |
598 | It is assumed that you have a copy of the Berkeley DB manual pages at |
599 | hand when reading this documentation. The interface defined here |
600 | mirrors the Berkeley DB interface closely. |
68dc0745 |
601 | |
8e07c86e |
602 | Berkeley DB is a C library which provides a consistent interface to a |
603 | number of database formats. B<DB_File> provides an interface to all |
604 | three of the database types currently supported by Berkeley DB. |
3b35bae3 |
605 | |
606 | The file types are: |
607 | |
608 | =over 5 |
609 | |
88108326 |
610 | =item B<DB_HASH> |
3b35bae3 |
611 | |
88108326 |
612 | This database type allows arbitrary key/value pairs to be stored in data |
8e07c86e |
613 | files. This is equivalent to the functionality provided by other |
614 | hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though, |
615 | the files created using DB_HASH are not compatible with any of the |
616 | other packages mentioned. |
3b35bae3 |
617 | |
8e07c86e |
618 | A default hashing algorithm, which will be adequate for most |
619 | applications, is built into Berkeley DB. If you do need to use your own |
620 | hashing algorithm it is possible to write your own in Perl and have |
621 | B<DB_File> use it instead. |
3b35bae3 |
622 | |
88108326 |
623 | =item B<DB_BTREE> |
624 | |
625 | The btree format allows arbitrary key/value pairs to be stored in a |
8e07c86e |
626 | sorted, balanced binary tree. |
3b35bae3 |
627 | |
8e07c86e |
628 | As with the DB_HASH format, it is possible to provide a user defined |
629 | Perl routine to perform the comparison of keys. By default, though, the |
630 | keys are stored in lexical order. |
3b35bae3 |
631 | |
88108326 |
632 | =item B<DB_RECNO> |
3b35bae3 |
633 | |
8e07c86e |
634 | DB_RECNO allows both fixed-length and variable-length flat text files |
635 | to be manipulated using the same key/value pair interface as in DB_HASH |
636 | and DB_BTREE. In this case the key will consist of a record (line) |
637 | number. |
3b35bae3 |
638 | |
639 | =back |
640 | |
039d031f |
641 | =head2 Using DB_File with Berkeley DB version 2 or 3 |
1f70e1ea |
642 | |
643 | Although B<DB_File> is intended to be used with Berkeley DB version 1, |
c5da4faf |
644 | it can also be used with version 2 or 3. In this case the interface is |
1f70e1ea |
645 | limited to the functionality provided by Berkeley DB 1.x. Anywhere the |
039d031f |
646 | version 2 or 3 interface differs, B<DB_File> arranges for it to work |
647 | like version 1. This feature allows B<DB_File> scripts that were built |
648 | with version 1 to be migrated to version 2 or 3 without any changes. |
1f70e1ea |
649 | |
650 | If you want to make use of the new features available in Berkeley DB |
b90e71be |
651 | 2.x or greater, use the Perl module B<BerkeleyDB> instead. |
1f70e1ea |
652 | |
039d031f |
653 | B<Note:> The database file format has changed in both Berkeley DB |
654 | version 2 and 3. If you cannot recreate your databases, you must dump |
c5da4faf |
655 | any existing databases with either the C<db_dump> or the C<db_dump185> |
656 | utility that comes with Berkeley DB. |
039d031f |
657 | Once you have rebuilt DB_File to use Berkeley DB version 2 or 3, your |
1f70e1ea |
658 | databases can be recreated using C<db_load>. Refer to the Berkeley DB |
659 | documentation for further details. |
660 | |
039d031f |
661 | Please read L<"COPYRIGHT"> before using version 2.x or 3.x of Berkeley |
662 | DB with DB_File. |
1f70e1ea |
663 | |
68dc0745 |
664 | =head2 Interface to Berkeley DB |
3b35bae3 |
665 | |
666 | B<DB_File> allows access to Berkeley DB files using the tie() mechanism |
8e07c86e |
667 | in Perl 5 (for full details, see L<perlfunc/tie()>). This facility |
668 | allows B<DB_File> to access Berkeley DB files using either an |
669 | associative array (for DB_HASH & DB_BTREE file types) or an ordinary |
670 | array (for the DB_RECNO file type). |
3b35bae3 |
671 | |
88108326 |
672 | In addition to the tie() interface, it is also possible to access most |
673 | of the functions provided in the Berkeley DB API directly. |
f6b705ef |
674 | See L<THE API INTERFACE>. |
3b35bae3 |
675 | |
88108326 |
676 | =head2 Opening a Berkeley DB Database File |
3b35bae3 |
677 | |
8e07c86e |
678 | Berkeley DB uses the function dbopen() to open or create a database. |
f6b705ef |
679 | Here is the C prototype for dbopen(): |
3b35bae3 |
680 | |
681 | DB* |
682 | dbopen (const char * file, int flags, int mode, |
683 | DBTYPE type, const void * openinfo) |
684 | |
685 | The parameter C<type> is an enumeration which specifies which of the 3 |
686 | interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used. |
687 | Depending on which of these is actually chosen, the final parameter, |
688 | I<openinfo> points to a data structure which allows tailoring of the |
689 | specific interface method. |
690 | |
8e07c86e |
691 | This interface is handled slightly differently in B<DB_File>. Here is |
88108326 |
692 | an equivalent call using B<DB_File>: |
3b35bae3 |
693 | |
88108326 |
694 | tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ; |
3b35bae3 |
695 | |
8e07c86e |
696 | The C<filename>, C<flags> and C<mode> parameters are the direct |
697 | equivalent of their dbopen() counterparts. The final parameter $DB_HASH |
698 | performs the function of both the C<type> and C<openinfo> parameters in |
699 | dbopen(). |
3b35bae3 |
700 | |
88108326 |
701 | In the example above $DB_HASH is actually a pre-defined reference to a |
702 | hash object. B<DB_File> has three of these pre-defined references. |
703 | Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO. |
3b35bae3 |
704 | |
8e07c86e |
705 | The keys allowed in each of these pre-defined references is limited to |
706 | the names used in the equivalent C structure. So, for example, the |
707 | $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>, |
88108326 |
708 | C<ffactor>, C<hash>, C<lorder> and C<nelem>. |
709 | |
710 | To change one of these elements, just assign to it like this: |
711 | |
712 | $DB_HASH->{'cachesize'} = 10000 ; |
713 | |
714 | The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are |
715 | usually adequate for most applications. If you do need to create extra |
716 | instances of these objects, constructors are available for each file |
717 | type. |
718 | |
719 | Here are examples of the constructors and the valid options available |
720 | for DB_HASH, DB_BTREE and DB_RECNO respectively. |
721 | |
722 | $a = new DB_File::HASHINFO ; |
723 | $a->{'bsize'} ; |
724 | $a->{'cachesize'} ; |
725 | $a->{'ffactor'}; |
726 | $a->{'hash'} ; |
727 | $a->{'lorder'} ; |
728 | $a->{'nelem'} ; |
729 | |
730 | $b = new DB_File::BTREEINFO ; |
731 | $b->{'flags'} ; |
732 | $b->{'cachesize'} ; |
733 | $b->{'maxkeypage'} ; |
734 | $b->{'minkeypage'} ; |
735 | $b->{'psize'} ; |
736 | $b->{'compare'} ; |
737 | $b->{'prefix'} ; |
738 | $b->{'lorder'} ; |
739 | |
740 | $c = new DB_File::RECNOINFO ; |
741 | $c->{'bval'} ; |
742 | $c->{'cachesize'} ; |
743 | $c->{'psize'} ; |
744 | $c->{'flags'} ; |
745 | $c->{'lorder'} ; |
746 | $c->{'reclen'} ; |
747 | $c->{'bfname'} ; |
748 | |
749 | The values stored in the hashes above are mostly the direct equivalent |
750 | of their C counterpart. Like their C counterparts, all are set to a |
f6b705ef |
751 | default values - that means you don't have to set I<all> of the |
88108326 |
752 | values when you only want to change one. Here is an example: |
753 | |
754 | $a = new DB_File::HASHINFO ; |
755 | $a->{'cachesize'} = 12345 ; |
756 | tie %y, 'DB_File', "filename", $flags, 0777, $a ; |
757 | |
36477c24 |
758 | A few of the options need extra discussion here. When used, the C |
88108326 |
759 | equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers |
760 | to C functions. In B<DB_File> these keys are used to store references |
761 | to Perl subs. Below are templates for each of the subs: |
762 | |
763 | sub hash |
764 | { |
765 | my ($data) = @_ ; |
766 | ... |
767 | # return the hash value for $data |
768 | return $hash ; |
769 | } |
3b35bae3 |
770 | |
88108326 |
771 | sub compare |
772 | { |
773 | my ($key, $key2) = @_ ; |
774 | ... |
775 | # return 0 if $key1 eq $key2 |
776 | # -1 if $key1 lt $key2 |
777 | # 1 if $key1 gt $key2 |
778 | return (-1 , 0 or 1) ; |
779 | } |
3b35bae3 |
780 | |
88108326 |
781 | sub prefix |
782 | { |
783 | my ($key, $key2) = @_ ; |
784 | ... |
785 | # return number of bytes of $key2 which are |
786 | # necessary to determine that it is greater than $key1 |
787 | return $bytes ; |
788 | } |
3b35bae3 |
789 | |
f6b705ef |
790 | See L<Changing the BTREE sort order> for an example of using the |
791 | C<compare> template. |
88108326 |
792 | |
36477c24 |
793 | If you are using the DB_RECNO interface and you intend making use of |
9a2c4ce3 |
794 | C<bval>, you should check out L<The 'bval' Option>. |
36477c24 |
795 | |
88108326 |
796 | =head2 Default Parameters |
797 | |
798 | It is possible to omit some or all of the final 4 parameters in the |
799 | call to C<tie> and let them take default values. As DB_HASH is the most |
800 | common file format used, the call: |
801 | |
802 | tie %A, "DB_File", "filename" ; |
803 | |
804 | is equivalent to: |
805 | |
18d2dc8c |
806 | tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ; |
88108326 |
807 | |
808 | It is also possible to omit the filename parameter as well, so the |
809 | call: |
810 | |
811 | tie %A, "DB_File" ; |
812 | |
813 | is equivalent to: |
814 | |
18d2dc8c |
815 | tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ; |
88108326 |
816 | |
f6b705ef |
817 | See L<In Memory Databases> for a discussion on the use of C<undef> |
88108326 |
818 | in place of a filename. |
819 | |
f6b705ef |
820 | =head2 In Memory Databases |
821 | |
822 | Berkeley DB allows the creation of in-memory databases by using NULL |
823 | (that is, a C<(char *)0> in C) in place of the filename. B<DB_File> |
824 | uses C<undef> instead of NULL to provide this functionality. |
825 | |
826 | =head1 DB_HASH |
827 | |
828 | The DB_HASH file format is probably the most commonly used of the three |
829 | file formats that B<DB_File> supports. It is also very straightforward |
830 | to use. |
831 | |
68dc0745 |
832 | =head2 A Simple Example |
f6b705ef |
833 | |
834 | This example shows how to create a database, add key/value pairs to the |
835 | database, delete keys/value pairs and finally how to enumerate the |
836 | contents of the database. |
837 | |
3245f058 |
838 | use warnings ; |
610ab055 |
839 | use strict ; |
f6b705ef |
840 | use DB_File ; |
610ab055 |
841 | use vars qw( %h $k $v ) ; |
f6b705ef |
842 | |
2c2d71f5 |
843 | unlink "fruit" ; |
f6b705ef |
844 | tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0640, $DB_HASH |
845 | or die "Cannot open file 'fruit': $!\n"; |
846 | |
847 | # Add a few key/value pairs to the file |
848 | $h{"apple"} = "red" ; |
849 | $h{"orange"} = "orange" ; |
850 | $h{"banana"} = "yellow" ; |
851 | $h{"tomato"} = "red" ; |
852 | |
853 | # Check for existence of a key |
854 | print "Banana Exists\n\n" if $h{"banana"} ; |
855 | |
856 | # Delete a key/value pair. |
857 | delete $h{"apple"} ; |
858 | |
859 | # print the contents of the file |
860 | while (($k, $v) = each %h) |
861 | { print "$k -> $v\n" } |
862 | |
863 | untie %h ; |
864 | |
865 | here is the output: |
866 | |
867 | Banana Exists |
bbc7dcd2 |
868 | |
f6b705ef |
869 | orange -> orange |
870 | tomato -> red |
871 | banana -> yellow |
872 | |
873 | Note that the like ordinary associative arrays, the order of the keys |
874 | retrieved is in an apparently random order. |
875 | |
876 | =head1 DB_BTREE |
877 | |
878 | The DB_BTREE format is useful when you want to store data in a given |
879 | order. By default the keys will be stored in lexical order, but as you |
880 | will see from the example shown in the next section, it is very easy to |
881 | define your own sorting function. |
882 | |
883 | =head2 Changing the BTREE sort order |
884 | |
885 | This script shows how to override the default sorting algorithm that |
886 | BTREE uses. Instead of using the normal lexical ordering, a case |
887 | insensitive compare function will be used. |
88108326 |
888 | |
3245f058 |
889 | use warnings ; |
610ab055 |
890 | use strict ; |
f6b705ef |
891 | use DB_File ; |
610ab055 |
892 | |
893 | my %h ; |
f6b705ef |
894 | |
895 | sub Compare |
896 | { |
897 | my ($key1, $key2) = @_ ; |
898 | "\L$key1" cmp "\L$key2" ; |
899 | } |
900 | |
901 | # specify the Perl sub that will do the comparison |
902 | $DB_BTREE->{'compare'} = \&Compare ; |
903 | |
2c2d71f5 |
904 | unlink "tree" ; |
f6b705ef |
905 | tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0640, $DB_BTREE |
906 | or die "Cannot open file 'tree': $!\n" ; |
907 | |
908 | # Add a key/value pair to the file |
909 | $h{'Wall'} = 'Larry' ; |
910 | $h{'Smith'} = 'John' ; |
911 | $h{'mouse'} = 'mickey' ; |
912 | $h{'duck'} = 'donald' ; |
913 | |
914 | # Delete |
915 | delete $h{"duck"} ; |
916 | |
917 | # Cycle through the keys printing them in order. |
918 | # Note it is not necessary to sort the keys as |
919 | # the btree will have kept them in order automatically. |
920 | foreach (keys %h) |
921 | { print "$_\n" } |
922 | |
923 | untie %h ; |
924 | |
925 | Here is the output from the code above. |
926 | |
927 | mouse |
928 | Smith |
929 | Wall |
930 | |
931 | There are a few point to bear in mind if you want to change the |
932 | ordering in a BTREE database: |
933 | |
934 | =over 5 |
935 | |
936 | =item 1. |
937 | |
938 | The new compare function must be specified when you create the database. |
939 | |
940 | =item 2. |
941 | |
942 | You cannot change the ordering once the database has been created. Thus |
943 | you must use the same compare function every time you access the |
88108326 |
944 | database. |
945 | |
f6b705ef |
946 | =back |
947 | |
68dc0745 |
948 | =head2 Handling Duplicate Keys |
f6b705ef |
949 | |
950 | The BTREE file type optionally allows a single key to be associated |
951 | with an arbitrary number of values. This option is enabled by setting |
952 | the flags element of C<$DB_BTREE> to R_DUP when creating the database. |
953 | |
88108326 |
954 | There are some difficulties in using the tied hash interface if you |
955 | want to manipulate a BTREE database with duplicate keys. Consider this |
956 | code: |
957 | |
3245f058 |
958 | use warnings ; |
610ab055 |
959 | use strict ; |
88108326 |
960 | use DB_File ; |
610ab055 |
961 | |
962 | use vars qw($filename %h ) ; |
963 | |
88108326 |
964 | $filename = "tree" ; |
965 | unlink $filename ; |
bbc7dcd2 |
966 | |
88108326 |
967 | # Enable duplicate records |
968 | $DB_BTREE->{'flags'} = R_DUP ; |
bbc7dcd2 |
969 | |
88108326 |
970 | tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
971 | or die "Cannot open $filename: $!\n"; |
bbc7dcd2 |
972 | |
88108326 |
973 | # Add some key/value pairs to the file |
974 | $h{'Wall'} = 'Larry' ; |
975 | $h{'Wall'} = 'Brick' ; # Note the duplicate key |
f6b705ef |
976 | $h{'Wall'} = 'Brick' ; # Note the duplicate key and value |
88108326 |
977 | $h{'Smith'} = 'John' ; |
978 | $h{'mouse'} = 'mickey' ; |
979 | |
980 | # iterate through the associative array |
981 | # and print each key/value pair. |
2c2d71f5 |
982 | foreach (sort keys %h) |
88108326 |
983 | { print "$_ -> $h{$_}\n" } |
984 | |
f6b705ef |
985 | untie %h ; |
986 | |
88108326 |
987 | Here is the output: |
988 | |
989 | Smith -> John |
990 | Wall -> Larry |
991 | Wall -> Larry |
f6b705ef |
992 | Wall -> Larry |
88108326 |
993 | mouse -> mickey |
994 | |
f6b705ef |
995 | As you can see 3 records have been successfully created with key C<Wall> |
88108326 |
996 | - the only thing is, when they are retrieved from the database they |
f6b705ef |
997 | I<seem> to have the same value, namely C<Larry>. The problem is caused |
998 | by the way that the associative array interface works. Basically, when |
999 | the associative array interface is used to fetch the value associated |
1000 | with a given key, it will only ever retrieve the first value. |
88108326 |
1001 | |
1002 | Although it may not be immediately obvious from the code above, the |
1003 | associative array interface can be used to write values with duplicate |
1004 | keys, but it cannot be used to read them back from the database. |
1005 | |
1006 | The way to get around this problem is to use the Berkeley DB API method |
1007 | called C<seq>. This method allows sequential access to key/value |
f6b705ef |
1008 | pairs. See L<THE API INTERFACE> for details of both the C<seq> method |
1009 | and the API in general. |
88108326 |
1010 | |
1011 | Here is the script above rewritten using the C<seq> API method. |
1012 | |
3245f058 |
1013 | use warnings ; |
610ab055 |
1014 | use strict ; |
88108326 |
1015 | use DB_File ; |
bbc7dcd2 |
1016 | |
610ab055 |
1017 | use vars qw($filename $x %h $status $key $value) ; |
1018 | |
88108326 |
1019 | $filename = "tree" ; |
1020 | unlink $filename ; |
bbc7dcd2 |
1021 | |
88108326 |
1022 | # Enable duplicate records |
1023 | $DB_BTREE->{'flags'} = R_DUP ; |
bbc7dcd2 |
1024 | |
88108326 |
1025 | $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
1026 | or die "Cannot open $filename: $!\n"; |
bbc7dcd2 |
1027 | |
88108326 |
1028 | # Add some key/value pairs to the file |
1029 | $h{'Wall'} = 'Larry' ; |
1030 | $h{'Wall'} = 'Brick' ; # Note the duplicate key |
f6b705ef |
1031 | $h{'Wall'} = 'Brick' ; # Note the duplicate key and value |
88108326 |
1032 | $h{'Smith'} = 'John' ; |
1033 | $h{'mouse'} = 'mickey' ; |
bbc7dcd2 |
1034 | |
f6b705ef |
1035 | # iterate through the btree using seq |
88108326 |
1036 | # and print each key/value pair. |
610ab055 |
1037 | $key = $value = 0 ; |
f6b705ef |
1038 | for ($status = $x->seq($key, $value, R_FIRST) ; |
1039 | $status == 0 ; |
1040 | $status = $x->seq($key, $value, R_NEXT) ) |
88108326 |
1041 | { print "$key -> $value\n" } |
bbc7dcd2 |
1042 | |
88108326 |
1043 | undef $x ; |
1044 | untie %h ; |
1045 | |
1046 | that prints: |
1047 | |
1048 | Smith -> John |
1049 | Wall -> Brick |
f6b705ef |
1050 | Wall -> Brick |
88108326 |
1051 | Wall -> Larry |
1052 | mouse -> mickey |
1053 | |
f6b705ef |
1054 | This time we have got all the key/value pairs, including the multiple |
88108326 |
1055 | values associated with the key C<Wall>. |
1056 | |
6ca2e664 |
1057 | To make life easier when dealing with duplicate keys, B<DB_File> comes with |
1058 | a few utility methods. |
1059 | |
68dc0745 |
1060 | =head2 The get_dup() Method |
f6b705ef |
1061 | |
6ca2e664 |
1062 | The C<get_dup> method assists in |
88108326 |
1063 | reading duplicate values from BTREE databases. The method can take the |
1064 | following forms: |
1065 | |
1066 | $count = $x->get_dup($key) ; |
1067 | @list = $x->get_dup($key) ; |
1068 | %list = $x->get_dup($key, 1) ; |
1069 | |
1070 | In a scalar context the method returns the number of values associated |
1071 | with the key, C<$key>. |
1072 | |
1073 | In list context, it returns all the values which match C<$key>. Note |
f6b705ef |
1074 | that the values will be returned in an apparently random order. |
88108326 |
1075 | |
7a2e2cd6 |
1076 | In list context, if the second parameter is present and evaluates |
1077 | TRUE, the method returns an associative array. The keys of the |
1078 | associative array correspond to the values that matched in the BTREE |
1079 | and the values of the array are a count of the number of times that |
1080 | particular value occurred in the BTREE. |
88108326 |
1081 | |
f6b705ef |
1082 | So assuming the database created above, we can use C<get_dup> like |
88108326 |
1083 | this: |
1084 | |
3245f058 |
1085 | use warnings ; |
2c2d71f5 |
1086 | use strict ; |
1087 | use DB_File ; |
bbc7dcd2 |
1088 | |
2c2d71f5 |
1089 | use vars qw($filename $x %h ) ; |
1090 | |
1091 | $filename = "tree" ; |
bbc7dcd2 |
1092 | |
2c2d71f5 |
1093 | # Enable duplicate records |
1094 | $DB_BTREE->{'flags'} = R_DUP ; |
bbc7dcd2 |
1095 | |
2c2d71f5 |
1096 | $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
1097 | or die "Cannot open $filename: $!\n"; |
1098 | |
610ab055 |
1099 | my $cnt = $x->get_dup("Wall") ; |
88108326 |
1100 | print "Wall occurred $cnt times\n" ; |
1101 | |
610ab055 |
1102 | my %hash = $x->get_dup("Wall", 1) ; |
88108326 |
1103 | print "Larry is there\n" if $hash{'Larry'} ; |
f6b705ef |
1104 | print "There are $hash{'Brick'} Brick Walls\n" ; |
88108326 |
1105 | |
2c2d71f5 |
1106 | my @list = sort $x->get_dup("Wall") ; |
88108326 |
1107 | print "Wall => [@list]\n" ; |
1108 | |
f6b705ef |
1109 | @list = $x->get_dup("Smith") ; |
88108326 |
1110 | print "Smith => [@list]\n" ; |
bbc7dcd2 |
1111 | |
f6b705ef |
1112 | @list = $x->get_dup("Dog") ; |
88108326 |
1113 | print "Dog => [@list]\n" ; |
1114 | |
1115 | |
1116 | and it will print: |
1117 | |
f6b705ef |
1118 | Wall occurred 3 times |
88108326 |
1119 | Larry is there |
f6b705ef |
1120 | There are 2 Brick Walls |
1121 | Wall => [Brick Brick Larry] |
88108326 |
1122 | Smith => [John] |
1123 | Dog => [] |
3b35bae3 |
1124 | |
6ca2e664 |
1125 | =head2 The find_dup() Method |
1126 | |
1127 | $status = $X->find_dup($key, $value) ; |
1128 | |
b90e71be |
1129 | This method checks for the existence of a specific key/value pair. If the |
6ca2e664 |
1130 | pair exists, the cursor is left pointing to the pair and the method |
1131 | returns 0. Otherwise the method returns a non-zero value. |
1132 | |
1133 | Assuming the database from the previous example: |
1134 | |
3245f058 |
1135 | use warnings ; |
6ca2e664 |
1136 | use strict ; |
1137 | use DB_File ; |
bbc7dcd2 |
1138 | |
6ca2e664 |
1139 | use vars qw($filename $x %h $found) ; |
1140 | |
1141 | my $filename = "tree" ; |
bbc7dcd2 |
1142 | |
6ca2e664 |
1143 | # Enable duplicate records |
1144 | $DB_BTREE->{'flags'} = R_DUP ; |
bbc7dcd2 |
1145 | |
6ca2e664 |
1146 | $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
1147 | or die "Cannot open $filename: $!\n"; |
1148 | |
1149 | $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ; |
1150 | print "Larry Wall is $found there\n" ; |
bbc7dcd2 |
1151 | |
6ca2e664 |
1152 | $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ; |
1153 | print "Harry Wall is $found there\n" ; |
bbc7dcd2 |
1154 | |
6ca2e664 |
1155 | undef $x ; |
1156 | untie %h ; |
1157 | |
1158 | prints this |
1159 | |
2c2d71f5 |
1160 | Larry Wall is there |
6ca2e664 |
1161 | Harry Wall is not there |
1162 | |
1163 | |
1164 | =head2 The del_dup() Method |
1165 | |
1166 | $status = $X->del_dup($key, $value) ; |
1167 | |
1168 | This method deletes a specific key/value pair. It returns |
1169 | 0 if they exist and have been deleted successfully. |
1170 | Otherwise the method returns a non-zero value. |
1171 | |
b90e71be |
1172 | Again assuming the existence of the C<tree> database |
6ca2e664 |
1173 | |
3245f058 |
1174 | use warnings ; |
6ca2e664 |
1175 | use strict ; |
1176 | use DB_File ; |
bbc7dcd2 |
1177 | |
6ca2e664 |
1178 | use vars qw($filename $x %h $found) ; |
1179 | |
1180 | my $filename = "tree" ; |
bbc7dcd2 |
1181 | |
6ca2e664 |
1182 | # Enable duplicate records |
1183 | $DB_BTREE->{'flags'} = R_DUP ; |
bbc7dcd2 |
1184 | |
6ca2e664 |
1185 | $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
1186 | or die "Cannot open $filename: $!\n"; |
1187 | |
1188 | $x->del_dup("Wall", "Larry") ; |
1189 | |
1190 | $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ; |
1191 | print "Larry Wall is $found there\n" ; |
bbc7dcd2 |
1192 | |
6ca2e664 |
1193 | undef $x ; |
1194 | untie %h ; |
1195 | |
1196 | prints this |
1197 | |
1198 | Larry Wall is not there |
1199 | |
f6b705ef |
1200 | =head2 Matching Partial Keys |
1201 | |
1202 | The BTREE interface has a feature which allows partial keys to be |
1203 | matched. This functionality is I<only> available when the C<seq> method |
1204 | is used along with the R_CURSOR flag. |
1205 | |
1206 | $x->seq($key, $value, R_CURSOR) ; |
1207 | |
1208 | Here is the relevant quote from the dbopen man page where it defines |
1209 | the use of the R_CURSOR flag with seq: |
1210 | |
f6b705ef |
1211 | Note, for the DB_BTREE access method, the returned key is not |
1212 | necessarily an exact match for the specified key. The returned key |
1213 | is the smallest key greater than or equal to the specified key, |
1214 | permitting partial key matches and range searches. |
1215 | |
f6b705ef |
1216 | In the example script below, the C<match> sub uses this feature to find |
1217 | and print the first matching key/value pair given a partial key. |
1218 | |
3245f058 |
1219 | use warnings ; |
610ab055 |
1220 | use strict ; |
f6b705ef |
1221 | use DB_File ; |
1222 | use Fcntl ; |
610ab055 |
1223 | |
1224 | use vars qw($filename $x %h $st $key $value) ; |
f6b705ef |
1225 | |
1226 | sub match |
1227 | { |
1228 | my $key = shift ; |
610ab055 |
1229 | my $value = 0; |
f6b705ef |
1230 | my $orig_key = $key ; |
1231 | $x->seq($key, $value, R_CURSOR) ; |
1232 | print "$orig_key\t-> $key\t-> $value\n" ; |
1233 | } |
1234 | |
1235 | $filename = "tree" ; |
1236 | unlink $filename ; |
1237 | |
1238 | $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_BTREE |
1239 | or die "Cannot open $filename: $!\n"; |
bbc7dcd2 |
1240 | |
f6b705ef |
1241 | # Add some key/value pairs to the file |
1242 | $h{'mouse'} = 'mickey' ; |
1243 | $h{'Wall'} = 'Larry' ; |
1244 | $h{'Walls'} = 'Brick' ; |
1245 | $h{'Smith'} = 'John' ; |
bbc7dcd2 |
1246 | |
f6b705ef |
1247 | |
610ab055 |
1248 | $key = $value = 0 ; |
f6b705ef |
1249 | print "IN ORDER\n" ; |
1250 | for ($st = $x->seq($key, $value, R_FIRST) ; |
1251 | $st == 0 ; |
1252 | $st = $x->seq($key, $value, R_NEXT) ) |
bbc7dcd2 |
1253 | |
2c2d71f5 |
1254 | { print "$key -> $value\n" } |
bbc7dcd2 |
1255 | |
f6b705ef |
1256 | print "\nPARTIAL MATCH\n" ; |
1257 | |
1258 | match "Wa" ; |
1259 | match "A" ; |
1260 | match "a" ; |
1261 | |
1262 | undef $x ; |
1263 | untie %h ; |
1264 | |
1265 | Here is the output: |
1266 | |
1267 | IN ORDER |
1268 | Smith -> John |
1269 | Wall -> Larry |
1270 | Walls -> Brick |
1271 | mouse -> mickey |
1272 | |
1273 | PARTIAL MATCH |
1274 | Wa -> Wall -> Larry |
1275 | A -> Smith -> John |
1276 | a -> mouse -> mickey |
1277 | |
1278 | =head1 DB_RECNO |
1279 | |
1280 | DB_RECNO provides an interface to flat text files. Both variable and |
1281 | fixed length records are supported. |
3b35bae3 |
1282 | |
6ca2e664 |
1283 | In order to make RECNO more compatible with Perl, the array offset for |
88108326 |
1284 | all RECNO arrays begins at 0 rather than 1 as in Berkeley DB. |
3b35bae3 |
1285 | |
88108326 |
1286 | As with normal Perl arrays, a RECNO array can be accessed using |
1287 | negative indexes. The index -1 refers to the last element of the array, |
1288 | -2 the second last, and so on. Attempting to access an element before |
1289 | the start of the array will raise a fatal run-time error. |
3b35bae3 |
1290 | |
68dc0745 |
1291 | =head2 The 'bval' Option |
36477c24 |
1292 | |
1293 | The operation of the bval option warrants some discussion. Here is the |
1294 | definition of bval from the Berkeley DB 1.85 recno manual page: |
1295 | |
1296 | The delimiting byte to be used to mark the end of a |
1297 | record for variable-length records, and the pad charac- |
1298 | ter for fixed-length records. If no value is speci- |
1299 | fied, newlines (``\n'') are used to mark the end of |
1300 | variable-length records and fixed-length records are |
1301 | padded with spaces. |
1302 | |
1303 | The second sentence is wrong. In actual fact bval will only default to |
1304 | C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL |
1305 | openinfo parameter is used at all, the value that happens to be in bval |
1306 | will be used. That means you always have to specify bval when making |
1307 | use of any of the options in the openinfo parameter. This documentation |
1308 | error will be fixed in the next release of Berkeley DB. |
1309 | |
1310 | That clarifies the situation with regards Berkeley DB itself. What |
1311 | about B<DB_File>? Well, the behavior defined in the quote above is |
6ca2e664 |
1312 | quite useful, so B<DB_File> conforms to it. |
36477c24 |
1313 | |
1314 | That means that you can specify other options (e.g. cachesize) and |
1315 | still have bval default to C<"\n"> for variable length records, and |
1316 | space for fixed length records. |
1317 | |
c5da4faf |
1318 | Also note that the bval option only allows you to specify a single byte |
1319 | as a delimeter. |
1320 | |
f6b705ef |
1321 | =head2 A Simple Example |
3b35bae3 |
1322 | |
6ca2e664 |
1323 | Here is a simple example that uses RECNO (if you are using a version |
1324 | of Perl earlier than 5.004_57 this example won't work -- see |
1325 | L<Extra RECNO Methods> for a workaround). |
f6b705ef |
1326 | |
3245f058 |
1327 | use warnings ; |
610ab055 |
1328 | use strict ; |
f6b705ef |
1329 | use DB_File ; |
f6b705ef |
1330 | |
2c2d71f5 |
1331 | my $filename = "text" ; |
1332 | unlink $filename ; |
1333 | |
610ab055 |
1334 | my @h ; |
2c2d71f5 |
1335 | tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0640, $DB_RECNO |
f6b705ef |
1336 | or die "Cannot open file 'text': $!\n" ; |
1337 | |
1338 | # Add a few key/value pairs to the file |
1339 | $h[0] = "orange" ; |
1340 | $h[1] = "blue" ; |
1341 | $h[2] = "yellow" ; |
1342 | |
6ca2e664 |
1343 | push @h, "green", "black" ; |
1344 | |
1345 | my $elements = scalar @h ; |
1346 | print "The array contains $elements entries\n" ; |
1347 | |
1348 | my $last = pop @h ; |
1349 | print "popped $last\n" ; |
1350 | |
1351 | unshift @h, "white" ; |
1352 | my $first = shift @h ; |
1353 | print "shifted $first\n" ; |
1354 | |
f6b705ef |
1355 | # Check for existence of a key |
1356 | print "Element 1 Exists with value $h[1]\n" if $h[1] ; |
1357 | |
1358 | # use a negative index |
1359 | print "The last element is $h[-1]\n" ; |
1360 | print "The 2nd last element is $h[-2]\n" ; |
1361 | |
1362 | untie @h ; |
3b35bae3 |
1363 | |
f6b705ef |
1364 | Here is the output from the script: |
1365 | |
6ca2e664 |
1366 | The array contains 5 entries |
1367 | popped black |
2c2d71f5 |
1368 | shifted white |
f6b705ef |
1369 | Element 1 Exists with value blue |
6ca2e664 |
1370 | The last element is green |
1371 | The 2nd last element is yellow |
f6b705ef |
1372 | |
6ca2e664 |
1373 | =head2 Extra RECNO Methods |
f6b705ef |
1374 | |
045291aa |
1375 | If you are using a version of Perl earlier than 5.004_57, the tied |
6ca2e664 |
1376 | array interface is quite limited. In the example script above |
1377 | C<push>, C<pop>, C<shift>, C<unshift> |
1378 | or determining the array length will not work with a tied array. |
045291aa |
1379 | |
1380 | To make the interface more useful for older versions of Perl, a number |
1381 | of methods are supplied with B<DB_File> to simulate the missing array |
1382 | operations. All these methods are accessed via the object returned from |
1383 | the tie call. |
f6b705ef |
1384 | |
1385 | Here are the methods: |
1386 | |
1387 | =over 5 |
3b35bae3 |
1388 | |
f6b705ef |
1389 | =item B<$X-E<gt>push(list) ;> |
1390 | |
1391 | Pushes the elements of C<list> to the end of the array. |
1392 | |
1393 | =item B<$value = $X-E<gt>pop ;> |
1394 | |
1395 | Removes and returns the last element of the array. |
1396 | |
1397 | =item B<$X-E<gt>shift> |
1398 | |
1399 | Removes and returns the first element of the array. |
1400 | |
1401 | =item B<$X-E<gt>unshift(list) ;> |
1402 | |
1403 | Pushes the elements of C<list> to the start of the array. |
1404 | |
1405 | =item B<$X-E<gt>length> |
1406 | |
1407 | Returns the number of elements in the array. |
1408 | |
c5da4faf |
1409 | =item B<$X-E<gt>splice(offset, length, elements);> |
1410 | |
1411 | Returns a splice of the the array. |
1412 | |
f6b705ef |
1413 | =back |
1414 | |
1415 | =head2 Another Example |
1416 | |
1417 | Here is a more complete example that makes use of some of the methods |
1418 | described above. It also makes use of the API interface directly (see |
1419 | L<THE API INTERFACE>). |
1420 | |
3245f058 |
1421 | use warnings ; |
f6b705ef |
1422 | use strict ; |
1423 | use vars qw(@h $H $file $i) ; |
1424 | use DB_File ; |
1425 | use Fcntl ; |
bbc7dcd2 |
1426 | |
f6b705ef |
1427 | $file = "text" ; |
1428 | |
1429 | unlink $file ; |
1430 | |
1431 | $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0640, $DB_RECNO |
1432 | or die "Cannot open file $file: $!\n" ; |
bbc7dcd2 |
1433 | |
f6b705ef |
1434 | # first create a text file to play with |
1435 | $h[0] = "zero" ; |
1436 | $h[1] = "one" ; |
1437 | $h[2] = "two" ; |
1438 | $h[3] = "three" ; |
1439 | $h[4] = "four" ; |
1440 | |
bbc7dcd2 |
1441 | |
f6b705ef |
1442 | # Print the records in order. |
1443 | # |
1444 | # The length method is needed here because evaluating a tied |
1445 | # array in a scalar context does not return the number of |
1446 | # elements in the array. |
1447 | |
1448 | print "\nORIGINAL\n" ; |
1449 | foreach $i (0 .. $H->length - 1) { |
1450 | print "$i: $h[$i]\n" ; |
1451 | } |
1452 | |
1453 | # use the push & pop methods |
1454 | $a = $H->pop ; |
1455 | $H->push("last") ; |
1456 | print "\nThe last record was [$a]\n" ; |
1457 | |
1458 | # and the shift & unshift methods |
1459 | $a = $H->shift ; |
1460 | $H->unshift("first") ; |
1461 | print "The first record was [$a]\n" ; |
1462 | |
1463 | # Use the API to add a new record after record 2. |
1464 | $i = 2 ; |
1465 | $H->put($i, "Newbie", R_IAFTER) ; |
1466 | |
1467 | # and a new record before record 1. |
1468 | $i = 1 ; |
1469 | $H->put($i, "New One", R_IBEFORE) ; |
1470 | |
1471 | # delete record 3 |
1472 | $H->del(3) ; |
1473 | |
1474 | # now print the records in reverse order |
1475 | print "\nREVERSE\n" ; |
1476 | for ($i = $H->length - 1 ; $i >= 0 ; -- $i) |
1477 | { print "$i: $h[$i]\n" } |
1478 | |
1479 | # same again, but use the API functions instead |
1480 | print "\nREVERSE again\n" ; |
610ab055 |
1481 | my ($s, $k, $v) = (0, 0, 0) ; |
f6b705ef |
1482 | for ($s = $H->seq($k, $v, R_LAST) ; |
1483 | $s == 0 ; |
1484 | $s = $H->seq($k, $v, R_PREV)) |
1485 | { print "$k: $v\n" } |
1486 | |
1487 | undef $H ; |
1488 | untie @h ; |
1489 | |
1490 | and this is what it outputs: |
1491 | |
1492 | ORIGINAL |
1493 | 0: zero |
1494 | 1: one |
1495 | 2: two |
1496 | 3: three |
1497 | 4: four |
1498 | |
1499 | The last record was [four] |
1500 | The first record was [zero] |
1501 | |
1502 | REVERSE |
1503 | 5: last |
1504 | 4: three |
1505 | 3: Newbie |
1506 | 2: one |
1507 | 1: New One |
1508 | 0: first |
1509 | |
1510 | REVERSE again |
1511 | 5: last |
1512 | 4: three |
1513 | 3: Newbie |
1514 | 2: one |
1515 | 1: New One |
1516 | 0: first |
1517 | |
1518 | Notes: |
1519 | |
1520 | =over 5 |
1521 | |
1522 | =item 1. |
1523 | |
1524 | Rather than iterating through the array, C<@h> like this: |
1525 | |
1526 | foreach $i (@h) |
1527 | |
1528 | it is necessary to use either this: |
1529 | |
1530 | foreach $i (0 .. $H->length - 1) |
1531 | |
1532 | or this: |
1533 | |
1534 | for ($a = $H->get($k, $v, R_FIRST) ; |
1535 | $a == 0 ; |
1536 | $a = $H->get($k, $v, R_NEXT) ) |
1537 | |
1538 | =item 2. |
1539 | |
1540 | Notice that both times the C<put> method was used the record index was |
1541 | specified using a variable, C<$i>, rather than the literal value |
1542 | itself. This is because C<put> will return the record number of the |
1543 | inserted line via that parameter. |
1544 | |
1545 | =back |
1546 | |
1547 | =head1 THE API INTERFACE |
3b35bae3 |
1548 | |
1549 | As well as accessing Berkeley DB using a tied hash or array, it is also |
88108326 |
1550 | possible to make direct use of most of the API functions defined in the |
8e07c86e |
1551 | Berkeley DB documentation. |
3b35bae3 |
1552 | |
88108326 |
1553 | To do this you need to store a copy of the object returned from the tie. |
3b35bae3 |
1554 | |
88108326 |
1555 | $db = tie %hash, "DB_File", "filename" ; |
3b35bae3 |
1556 | |
8e07c86e |
1557 | Once you have done that, you can access the Berkeley DB API functions |
88108326 |
1558 | as B<DB_File> methods directly like this: |
3b35bae3 |
1559 | |
1560 | $db->put($key, $value, R_NOOVERWRITE) ; |
1561 | |
88108326 |
1562 | B<Important:> If you have saved a copy of the object returned from |
1563 | C<tie>, the underlying database file will I<not> be closed until both |
1564 | the tied variable is untied and all copies of the saved object are |
610ab055 |
1565 | destroyed. |
88108326 |
1566 | |
1567 | use DB_File ; |
1568 | $db = tie %hash, "DB_File", "filename" |
1569 | or die "Cannot tie filename: $!" ; |
1570 | ... |
1571 | undef $db ; |
1572 | untie %hash ; |
1573 | |
9a2c4ce3 |
1574 | See L<The untie() Gotcha> for more details. |
778183f3 |
1575 | |
88108326 |
1576 | All the functions defined in L<dbopen> are available except for |
1577 | close() and dbopen() itself. The B<DB_File> method interface to the |
1578 | supported functions have been implemented to mirror the way Berkeley DB |
1579 | works whenever possible. In particular note that: |
1580 | |
1581 | =over 5 |
1582 | |
1583 | =item * |
1584 | |
1585 | The methods return a status value. All return 0 on success. |
1586 | All return -1 to signify an error and set C<$!> to the exact |
1587 | error code. The return code 1 generally (but not always) means that the |
1588 | key specified did not exist in the database. |
1589 | |
1590 | Other return codes are defined. See below and in the Berkeley DB |
1591 | documentation for details. The Berkeley DB documentation should be used |
1592 | as the definitive source. |
1593 | |
1594 | =item * |
3b35bae3 |
1595 | |
88108326 |
1596 | Whenever a Berkeley DB function returns data via one of its parameters, |
1597 | the equivalent B<DB_File> method does exactly the same. |
3b35bae3 |
1598 | |
88108326 |
1599 | =item * |
1600 | |
1601 | If you are careful, it is possible to mix API calls with the tied |
1602 | hash/array interface in the same piece of code. Although only a few of |
1603 | the methods used to implement the tied interface currently make use of |
1604 | the cursor, you should always assume that the cursor has been changed |
1605 | any time the tied hash/array interface is used. As an example, this |
1606 | code will probably not do what you expect: |
1607 | |
1608 | $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE |
1609 | or die "Cannot tie $filename: $!" ; |
1610 | |
1611 | # Get the first key/value pair and set the cursor |
1612 | $X->seq($key, $value, R_FIRST) ; |
1613 | |
1614 | # this line will modify the cursor |
1615 | $count = scalar keys %x ; |
1616 | |
1617 | # Get the second key/value pair. |
1618 | # oops, it didn't, it got the last key/value pair! |
1619 | $X->seq($key, $value, R_NEXT) ; |
1620 | |
1621 | The code above can be rearranged to get around the problem, like this: |
1622 | |
1623 | $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE |
1624 | or die "Cannot tie $filename: $!" ; |
1625 | |
1626 | # this line will modify the cursor |
1627 | $count = scalar keys %x ; |
1628 | |
1629 | # Get the first key/value pair and set the cursor |
1630 | $X->seq($key, $value, R_FIRST) ; |
1631 | |
1632 | # Get the second key/value pair. |
1633 | # worked this time. |
1634 | $X->seq($key, $value, R_NEXT) ; |
1635 | |
1636 | =back |
1637 | |
1638 | All the constants defined in L<dbopen> for use in the flags parameters |
1639 | in the methods defined below are also available. Refer to the Berkeley |
1640 | DB documentation for the precise meaning of the flags values. |
1641 | |
1642 | Below is a list of the methods available. |
3b35bae3 |
1643 | |
1644 | =over 5 |
1645 | |
f6b705ef |
1646 | =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;> |
88108326 |
1647 | |
1648 | Given a key (C<$key>) this method reads the value associated with it |
1649 | from the database. The value read from the database is returned in the |
1650 | C<$value> parameter. |
3b35bae3 |
1651 | |
88108326 |
1652 | If the key does not exist the method returns 1. |
3b35bae3 |
1653 | |
88108326 |
1654 | No flags are currently defined for this method. |
3b35bae3 |
1655 | |
f6b705ef |
1656 | =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;> |
3b35bae3 |
1657 | |
88108326 |
1658 | Stores the key/value pair in the database. |
1659 | |
1660 | If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter |
8e07c86e |
1661 | will have the record number of the inserted key/value pair set. |
3b35bae3 |
1662 | |
88108326 |
1663 | Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and |
1664 | R_SETCURSOR. |
1665 | |
f6b705ef |
1666 | =item B<$status = $X-E<gt>del($key [, $flags]) ;> |
3b35bae3 |
1667 | |
88108326 |
1668 | Removes all key/value pairs with key C<$key> from the database. |
3b35bae3 |
1669 | |
88108326 |
1670 | A return code of 1 means that the requested key was not in the |
1671 | database. |
3b35bae3 |
1672 | |
88108326 |
1673 | R_CURSOR is the only valid flag at present. |
3b35bae3 |
1674 | |
f6b705ef |
1675 | =item B<$status = $X-E<gt>fd ;> |
3b35bae3 |
1676 | |
88108326 |
1677 | Returns the file descriptor for the underlying database. |
3b35bae3 |
1678 | |
b90e71be |
1679 | See L<Locking: The Trouble with fd> for an explanation for why you should |
1680 | not use C<fd> to lock your database. |
3b35bae3 |
1681 | |
f6b705ef |
1682 | =item B<$status = $X-E<gt>seq($key, $value, $flags) ;> |
3b35bae3 |
1683 | |
88108326 |
1684 | This interface allows sequential retrieval from the database. See |
1685 | L<dbopen> for full details. |
1686 | |
1687 | Both the C<$key> and C<$value> parameters will be set to the key/value |
1688 | pair read from the database. |
1689 | |
1690 | The flags parameter is mandatory. The valid flag values are R_CURSOR, |
1691 | R_FIRST, R_LAST, R_NEXT and R_PREV. |
1692 | |
f6b705ef |
1693 | =item B<$status = $X-E<gt>sync([$flags]) ;> |
88108326 |
1694 | |
1695 | Flushes any cached buffers to disk. |
1696 | |
1697 | R_RECNOSYNC is the only valid flag at present. |
3b35bae3 |
1698 | |
1699 | =back |
1700 | |
cad2e5aa |
1701 | =head1 DBM FILTERS |
1702 | |
1703 | A DBM Filter is a piece of code that is be used when you I<always> |
1704 | want to make the same transformation to all keys and/or values in a |
1705 | DBM database. |
1706 | |
1707 | There are four methods associated with DBM Filters. All work identically, |
1708 | and each is used to install (or uninstall) a single DBM Filter. Each |
1709 | expects a single parameter, namely a reference to a sub. The only |
1710 | difference between them is the place that the filter is installed. |
1711 | |
1712 | To summarise: |
1713 | |
1714 | =over 5 |
1715 | |
1716 | =item B<filter_store_key> |
1717 | |
1718 | If a filter has been installed with this method, it will be invoked |
1719 | every time you write a key to a DBM database. |
1720 | |
1721 | =item B<filter_store_value> |
1722 | |
1723 | If a filter has been installed with this method, it will be invoked |
1724 | every time you write a value to a DBM database. |
1725 | |
1726 | |
1727 | =item B<filter_fetch_key> |
1728 | |
1729 | If a filter has been installed with this method, it will be invoked |
1730 | every time you read a key from a DBM database. |
1731 | |
1732 | =item B<filter_fetch_value> |
1733 | |
1734 | If a filter has been installed with this method, it will be invoked |
1735 | every time you read a value from a DBM database. |
1736 | |
1737 | =back |
1738 | |
1739 | You can use any combination of the methods, from none, to all four. |
1740 | |
1741 | All filter methods return the existing filter, if present, or C<undef> |
1742 | in not. |
1743 | |
1744 | To delete a filter pass C<undef> to it. |
1745 | |
1746 | =head2 The Filter |
1747 | |
1748 | When each filter is called by Perl, a local copy of C<$_> will contain |
1749 | the key or value to be filtered. Filtering is achieved by modifying |
1750 | the contents of C<$_>. The return code from the filter is ignored. |
1751 | |
1752 | =head2 An Example -- the NULL termination problem. |
1753 | |
1754 | Consider the following scenario. You have a DBM database |
1755 | that you need to share with a third-party C application. The C application |
1756 | assumes that I<all> keys and values are NULL terminated. Unfortunately |
1757 | when Perl writes to DBM databases it doesn't use NULL termination, so |
1758 | your Perl application will have to manage NULL termination itself. When |
1759 | you write to the database you will have to use something like this: |
1760 | |
1761 | $hash{"$key\0"} = "$value\0" ; |
1762 | |
1763 | Similarly the NULL needs to be taken into account when you are considering |
1764 | the length of existing keys/values. |
1765 | |
1766 | It would be much better if you could ignore the NULL terminations issue |
1767 | in the main application code and have a mechanism that automatically |
1768 | added the terminating NULL to all keys and values whenever you write to |
1769 | the database and have them removed when you read from the database. As I'm |
1770 | sure you have already guessed, this is a problem that DBM Filters can |
1771 | fix very easily. |
1772 | |
3245f058 |
1773 | use warnings ; |
cad2e5aa |
1774 | use strict ; |
1775 | use DB_File ; |
1776 | |
1777 | my %hash ; |
1778 | my $filename = "/tmp/filt" ; |
1779 | unlink $filename ; |
1780 | |
1781 | my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH |
1782 | or die "Cannot open $filename: $!\n" ; |
1783 | |
1784 | # Install DBM Filters |
1785 | $db->filter_fetch_key ( sub { s/\0$// } ) ; |
1786 | $db->filter_store_key ( sub { $_ .= "\0" } ) ; |
1787 | $db->filter_fetch_value( sub { s/\0$// } ) ; |
1788 | $db->filter_store_value( sub { $_ .= "\0" } ) ; |
1789 | |
1790 | $hash{"abc"} = "def" ; |
1791 | my $a = $hash{"ABC"} ; |
1792 | # ... |
1793 | undef $db ; |
1794 | untie %hash ; |
1795 | |
1796 | Hopefully the contents of each of the filters should be |
1797 | self-explanatory. Both "fetch" filters remove the terminating NULL, |
1798 | and both "store" filters add a terminating NULL. |
1799 | |
1800 | |
1801 | =head2 Another Example -- Key is a C int. |
1802 | |
1803 | Here is another real-life example. By default, whenever Perl writes to |
1804 | a DBM database it always writes the key and value as strings. So when |
1805 | you use this: |
1806 | |
1807 | $hash{12345} = "soemthing" ; |
1808 | |
1809 | the key 12345 will get stored in the DBM database as the 5 byte string |
1810 | "12345". If you actually want the key to be stored in the DBM database |
1811 | as a C int, you will have to use C<pack> when writing, and C<unpack> |
1812 | when reading. |
1813 | |
1814 | Here is a DBM Filter that does it: |
1815 | |
3245f058 |
1816 | use warnings ; |
cad2e5aa |
1817 | use strict ; |
1818 | use DB_File ; |
1819 | my %hash ; |
1820 | my $filename = "/tmp/filt" ; |
1821 | unlink $filename ; |
1822 | |
1823 | |
1824 | my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH |
1825 | or die "Cannot open $filename: $!\n" ; |
1826 | |
1827 | $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ; |
1828 | $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ; |
1829 | $hash{123} = "def" ; |
1830 | # ... |
1831 | undef $db ; |
1832 | untie %hash ; |
1833 | |
1834 | This time only two filters have been used -- we only need to manipulate |
1835 | the contents of the key, so it wasn't necessary to install any value |
1836 | filters. |
1837 | |
f6b705ef |
1838 | =head1 HINTS AND TIPS |
3b35bae3 |
1839 | |
3b35bae3 |
1840 | |
b90e71be |
1841 | =head2 Locking: The Trouble with fd |
3b35bae3 |
1842 | |
b90e71be |
1843 | Until version 1.72 of this module, the recommended technique for locking |
1844 | B<DB_File> databases was to flock the filehandle returned from the "fd" |
1845 | function. Unfortunately this technique has been shown to be fundamentally |
1846 | flawed (Kudos to David Harris for tracking this down). Use it at your own |
1847 | peril! |
3b35bae3 |
1848 | |
b90e71be |
1849 | The locking technique went like this. |
cb1a09d0 |
1850 | |
b90e71be |
1851 | $db = tie(%db, 'DB_File', '/tmp/foo.db', O_CREAT|O_RDWR, 0644) |
1852 | || die "dbcreat /tmp/foo.db $!"; |
1853 | $fd = $db->fd; |
1854 | open(DB_FH, "+<&=$fd") || die "dup $!"; |
1855 | flock (DB_FH, LOCK_EX) || die "flock: $!"; |
1856 | ... |
1857 | $db{"Tom"} = "Jerry" ; |
1858 | ... |
1859 | flock(DB_FH, LOCK_UN); |
1860 | undef $db; |
1861 | untie %db; |
1862 | close(DB_FH); |
cb1a09d0 |
1863 | |
b90e71be |
1864 | In simple terms, this is what happens: |
cb1a09d0 |
1865 | |
b90e71be |
1866 | =over 5 |
cb1a09d0 |
1867 | |
b90e71be |
1868 | =item 1. |
cb1a09d0 |
1869 | |
b90e71be |
1870 | Use "tie" to open the database. |
cb1a09d0 |
1871 | |
b90e71be |
1872 | =item 2. |
cb1a09d0 |
1873 | |
b90e71be |
1874 | Lock the database with fd & flock. |
cb1a09d0 |
1875 | |
b90e71be |
1876 | =item 3. |
cb1a09d0 |
1877 | |
b90e71be |
1878 | Read & Write to the database. |
cb1a09d0 |
1879 | |
b90e71be |
1880 | =item 4. |
cb1a09d0 |
1881 | |
b90e71be |
1882 | Unlock and close the database. |
cb1a09d0 |
1883 | |
b90e71be |
1884 | =back |
1885 | |
1886 | Here is the crux of the problem. A side-effect of opening the B<DB_File> |
1887 | database in step 2 is that an initial block from the database will get |
1888 | read from disk and cached in memory. |
1889 | |
1890 | To see why this is a problem, consider what can happen when two processes, |
1891 | say "A" and "B", both want to update the same B<DB_File> database |
1892 | using the locking steps outlined above. Assume process "A" has already |
1893 | opened the database and has a write lock, but it hasn't actually updated |
1894 | the database yet (it has finished step 2, but not started step 3 yet). Now |
1895 | process "B" tries to open the same database - step 1 will succeed, |
1896 | but it will block on step 2 until process "A" releases the lock. The |
1897 | important thing to notice here is that at this point in time both |
1898 | processes will have cached identical initial blocks from the database. |
1899 | |
1900 | Now process "A" updates the database and happens to change some of the |
1901 | data held in the initial buffer. Process "A" terminates, flushing |
1902 | all cached data to disk and releasing the database lock. At this point |
1903 | the database on disk will correctly reflect the changes made by process |
1904 | "A". |
1905 | |
1906 | With the lock released, process "B" can now continue. It also updates the |
1907 | database and unfortunately it too modifies the data that was in its |
1908 | initial buffer. Once that data gets flushed to disk it will overwrite |
1909 | some/all of the changes process "A" made to the database. |
1910 | |
1911 | The result of this scenario is at best a database that doesn't contain |
1912 | what you expect. At worst the database will corrupt. |
1913 | |
1914 | The above won't happen every time competing process update the same |
1915 | B<DB_File> database, but it does illustrate why the technique should |
1916 | not be used. |
1917 | |
1918 | =head2 Safe ways to lock a database |
1919 | |
1920 | Starting with version 2.x, Berkeley DB has internal support for locking. |
1921 | The companion module to this one, B<BerkeleyDB>, provides an interface |
1922 | to this locking functionality. If you are serious about locking |
1923 | Berkeley DB databases, I strongly recommend using B<BerkeleyDB>. |
1924 | |
1925 | If using B<BerkeleyDB> isn't an option, there are a number of modules |
1926 | available on CPAN that can be used to implement locking. Each one |
1927 | implements locking differently and has different goals in mind. It is |
1928 | therefore worth knowing the difference, so that you can pick the right |
1929 | one for your application. Here are the three locking wrappers: |
1930 | |
1931 | =over 5 |
1932 | |
1933 | =item B<Tie::DB_Lock> |
1934 | |
1935 | A B<DB_File> wrapper which creates copies of the database file for |
1936 | read access, so that you have a kind of a multiversioning concurrent read |
1937 | system. However, updates are still serial. Use for databases where reads |
1938 | may be lengthy and consistency problems may occur. |
1939 | |
1940 | =item B<Tie::DB_LockFile> |
1941 | |
1942 | A B<DB_File> wrapper that has the ability to lock and unlock the database |
1943 | while it is being used. Avoids the tie-before-flock problem by simply |
1944 | re-tie-ing the database when you get or drop a lock. Because of the |
1945 | flexibility in dropping and re-acquiring the lock in the middle of a |
1946 | session, this can be massaged into a system that will work with long |
1947 | updates and/or reads if the application follows the hints in the POD |
1948 | documentation. |
1949 | |
1950 | =item B<DB_File::Lock> |
1951 | |
1952 | An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile |
1953 | before tie-ing the database and drops the lock after the untie. Allows |
1954 | one to use the same lockfile for multiple databases to avoid deadlock |
1955 | problems, if desired. Use for databases where updates are reads are |
1956 | quick and simple flock locking semantics are enough. |
1957 | |
1958 | =back |
cb1a09d0 |
1959 | |
68dc0745 |
1960 | =head2 Sharing Databases With C Applications |
f6b705ef |
1961 | |
1962 | There is no technical reason why a Berkeley DB database cannot be |
1963 | shared by both a Perl and a C application. |
1964 | |
1965 | The vast majority of problems that are reported in this area boil down |
1966 | to the fact that C strings are NULL terminated, whilst Perl strings are |
cad2e5aa |
1967 | not. See L<DBM FILTERS> for a generic way to work around this problem. |
f6b705ef |
1968 | |
1969 | Here is a real example. Netscape 2.0 keeps a record of the locations you |
1970 | visit along with the time you last visited them in a DB_HASH database. |
1971 | This is usually stored in the file F<~/.netscape/history.db>. The key |
1972 | field in the database is the location string and the value field is the |
1973 | time the location was last visited stored as a 4 byte binary value. |
1974 | |
1975 | If you haven't already guessed, the location string is stored with a |
1976 | terminating NULL. This means you need to be careful when accessing the |
1977 | database. |
1978 | |
1979 | Here is a snippet of code that is loosely based on Tom Christiansen's |
1980 | I<ggh> script (available from your nearest CPAN archive in |
1981 | F<authors/id/TOMC/scripts/nshist.gz>). |
1982 | |
3245f058 |
1983 | use warnings ; |
610ab055 |
1984 | use strict ; |
f6b705ef |
1985 | use DB_File ; |
1986 | use Fcntl ; |
f6b705ef |
1987 | |
610ab055 |
1988 | use vars qw( $dotdir $HISTORY %hist_db $href $binary_time $date ) ; |
f6b705ef |
1989 | $dotdir = $ENV{HOME} || $ENV{LOGNAME}; |
1990 | |
1991 | $HISTORY = "$dotdir/.netscape/history.db"; |
1992 | |
1993 | tie %hist_db, 'DB_File', $HISTORY |
1994 | or die "Cannot open $HISTORY: $!\n" ;; |
1995 | |
1996 | # Dump the complete database |
1997 | while ( ($href, $binary_time) = each %hist_db ) { |
1998 | |
1999 | # remove the terminating NULL |
2000 | $href =~ s/\x00$// ; |
2001 | |
2002 | # convert the binary time into a user friendly string |
2003 | $date = localtime unpack("V", $binary_time); |
2004 | print "$date $href\n" ; |
2005 | } |
2006 | |
2007 | # check for the existence of a specific key |
2008 | # remember to add the NULL |
2009 | if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) { |
2010 | $date = localtime unpack("V", $binary_time) ; |
2011 | print "Last visited mox.perl.com on $date\n" ; |
2012 | } |
2013 | else { |
2014 | print "Never visited mox.perl.com\n" |
2015 | } |
2016 | |
2017 | untie %hist_db ; |
2018 | |
68dc0745 |
2019 | =head2 The untie() Gotcha |
778183f3 |
2020 | |
7a2e2cd6 |
2021 | If you make use of the Berkeley DB API, it is I<very> strongly |
68dc0745 |
2022 | recommended that you read L<perltie/The untie Gotcha>. |
778183f3 |
2023 | |
2024 | Even if you don't currently make use of the API interface, it is still |
2025 | worth reading it. |
2026 | |
2027 | Here is an example which illustrates the problem from a B<DB_File> |
2028 | perspective: |
2029 | |
2030 | use DB_File ; |
2031 | use Fcntl ; |
2032 | |
2033 | my %x ; |
2034 | my $X ; |
2035 | |
2036 | $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC |
2037 | or die "Cannot tie first time: $!" ; |
2038 | |
2039 | $x{123} = 456 ; |
2040 | |
2041 | untie %x ; |
2042 | |
2043 | tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT |
2044 | or die "Cannot tie second time: $!" ; |
2045 | |
2046 | untie %x ; |
2047 | |
2048 | When run, the script will produce this error message: |
2049 | |
2050 | Cannot tie second time: Invalid argument at bad.file line 14. |
2051 | |
2052 | Although the error message above refers to the second tie() statement |
2053 | in the script, the source of the problem is really with the untie() |
2054 | statement that precedes it. |
2055 | |
2056 | Having read L<perltie> you will probably have already guessed that the |
2057 | error is caused by the extra copy of the tied object stored in C<$X>. |
2058 | If you haven't, then the problem boils down to the fact that the |
2059 | B<DB_File> destructor, DESTROY, will not be called until I<all> |
2060 | references to the tied object are destroyed. Both the tied variable, |
2061 | C<%x>, and C<$X> above hold a reference to the object. The call to |
2062 | untie() will destroy the first, but C<$X> still holds a valid |
2063 | reference, so the destructor will not get called and the database file |
2064 | F<tst.fil> will remain open. The fact that Berkeley DB then reports the |
b90e71be |
2065 | attempt to open a database that is already open via the catch-all |
778183f3 |
2066 | "Invalid argument" doesn't help. |
2067 | |
2068 | If you run the script with the C<-w> flag the error message becomes: |
2069 | |
2070 | untie attempted while 1 inner references still exist at bad.file line 12. |
2071 | Cannot tie second time: Invalid argument at bad.file line 14. |
2072 | |
2073 | which pinpoints the real problem. Finally the script can now be |
2074 | modified to fix the original problem by destroying the API object |
2075 | before the untie: |
2076 | |
2077 | ... |
2078 | $x{123} = 456 ; |
2079 | |
2080 | undef $X ; |
2081 | untie %x ; |
2082 | |
2083 | $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT |
2084 | ... |
2085 | |
f6b705ef |
2086 | |
2087 | =head1 COMMON QUESTIONS |
2088 | |
2089 | =head2 Why is there Perl source in my database? |
2090 | |
2091 | If you look at the contents of a database file created by DB_File, |
2092 | there can sometimes be part of a Perl script included in it. |
2093 | |
2094 | This happens because Berkeley DB uses dynamic memory to allocate |
2095 | buffers which will subsequently be written to the database file. Being |
2096 | dynamic, the memory could have been used for anything before DB |
2097 | malloced it. As Berkeley DB doesn't clear the memory once it has been |
2098 | allocated, the unused portions will contain random junk. In the case |
2099 | where a Perl script gets written to the database, the random junk will |
2100 | correspond to an area of dynamic memory that happened to be used during |
2101 | the compilation of the script. |
2102 | |
2103 | Unless you don't like the possibility of there being part of your Perl |
2104 | scripts embedded in a database file, this is nothing to worry about. |
2105 | |
2106 | =head2 How do I store complex data structures with DB_File? |
2107 | |
2108 | Although B<DB_File> cannot do this directly, there is a module which |
2109 | can layer transparently over B<DB_File> to accomplish this feat. |
2110 | |
2111 | Check out the MLDBM module, available on CPAN in the directory |
2112 | F<modules/by-module/MLDBM>. |
2113 | |
2114 | =head2 What does "Invalid Argument" mean? |
2115 | |
2116 | You will get this error message when one of the parameters in the |
2117 | C<tie> call is wrong. Unfortunately there are quite a few parameters to |
2118 | get wrong, so it can be difficult to figure out which one it is. |
2119 | |
2120 | Here are a couple of possibilities: |
2121 | |
2122 | =over 5 |
2123 | |
2124 | =item 1. |
2125 | |
610ab055 |
2126 | Attempting to reopen a database without closing it. |
f6b705ef |
2127 | |
2128 | =item 2. |
2129 | |
2130 | Using the O_WRONLY flag. |
2131 | |
2132 | =back |
2133 | |
2134 | =head2 What does "Bareword 'DB_File' not allowed" mean? |
2135 | |
2136 | You will encounter this particular error message when you have the |
2137 | C<strict 'subs'> pragma (or the full strict pragma) in your script. |
2138 | Consider this script: |
2139 | |
3245f058 |
2140 | use warnings ; |
f6b705ef |
2141 | use strict ; |
2142 | use DB_File ; |
2143 | use vars qw(%x) ; |
2144 | tie %x, DB_File, "filename" ; |
2145 | |
2146 | Running it produces the error in question: |
2147 | |
2148 | Bareword "DB_File" not allowed while "strict subs" in use |
2149 | |
2150 | To get around the error, place the word C<DB_File> in either single or |
2151 | double quotes, like this: |
2152 | |
2153 | tie %x, "DB_File", "filename" ; |
2154 | |
2155 | Although it might seem like a real pain, it is really worth the effort |
2156 | of having a C<use strict> in all your scripts. |
2157 | |
cad2e5aa |
2158 | =head1 REFERENCES |
2159 | |
2160 | Articles that are either about B<DB_File> or make use of it. |
2161 | |
2162 | =over 5 |
2163 | |
2164 | =item 1. |
2165 | |
2166 | I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com), |
2167 | Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41 |
2168 | |
2169 | =back |
2170 | |
cb1a09d0 |
2171 | =head1 HISTORY |
2172 | |
1f70e1ea |
2173 | Moved to the Changes file. |
610ab055 |
2174 | |
1f70e1ea |
2175 | =head1 BUGS |
05475680 |
2176 | |
1f70e1ea |
2177 | Some older versions of Berkeley DB had problems with fixed length |
2178 | records using the RECNO file format. This problem has been fixed since |
2179 | version 1.85 of Berkeley DB. |
e858de61 |
2180 | |
1f70e1ea |
2181 | I am sure there are bugs in the code. If you do find any, or can |
2182 | suggest any enhancements, I would welcome your comments. |
a6ed719b |
2183 | |
1f70e1ea |
2184 | =head1 AVAILABILITY |
a6ed719b |
2185 | |
1f70e1ea |
2186 | B<DB_File> comes with the standard Perl source distribution. Look in |
2187 | the directory F<ext/DB_File>. Given the amount of time between releases |
2188 | of Perl the version that ships with Perl is quite likely to be out of |
2189 | date, so the most recent version can always be found on CPAN (see |
2190 | L<perlmod/CPAN> for details), in the directory |
2191 | F<modules/by-module/DB_File>. |
a6ed719b |
2192 | |
039d031f |
2193 | This version of B<DB_File> will work with either version 1.x, 2.x or |
2194 | 3.x of Berkeley DB, but is limited to the functionality provided by |
2195 | version 1. |
a6ed719b |
2196 | |
cad2e5aa |
2197 | The official web site for Berkeley DB is F<http://www.sleepycat.com>. |
039d031f |
2198 | All versions of Berkeley DB are available there. |
93af7a87 |
2199 | |
1f70e1ea |
2200 | Alternatively, Berkeley DB version 1 is available at your nearest CPAN |
2201 | archive in F<src/misc/db.1.85.tar.gz>. |
e858de61 |
2202 | |
1f70e1ea |
2203 | If you are running IRIX, then get Berkeley DB version 1 from |
2204 | F<http://reality.sgi.com/ariel>. It has the patches necessary to |
2205 | compile properly on IRIX 5.3. |
610ab055 |
2206 | |
1f70e1ea |
2207 | =head1 COPYRIGHT |
3b35bae3 |
2208 | |
c5da4faf |
2209 | Copyright (c) 1995-2001 Paul Marquess. All rights reserved. This program |
a9fd575d |
2210 | is free software; you can redistribute it and/or modify it under the |
2211 | same terms as Perl itself. |
3b35bae3 |
2212 | |
1f70e1ea |
2213 | Although B<DB_File> is covered by the Perl license, the library it |
2214 | makes use of, namely Berkeley DB, is not. Berkeley DB has its own |
2215 | copyright and its own license. Please take the time to read it. |
3b35bae3 |
2216 | |
a9fd575d |
2217 | Here are are few words taken from the Berkeley DB FAQ (at |
b90e71be |
2218 | F<http://www.sleepycat.com>) regarding the license: |
68dc0745 |
2219 | |
a9fd575d |
2220 | Do I have to license DB to use it in Perl scripts? |
3b35bae3 |
2221 | |
a9fd575d |
2222 | No. The Berkeley DB license requires that software that uses |
2223 | Berkeley DB be freely redistributable. In the case of Perl, that |
2224 | software is Perl, and not your scripts. Any Perl scripts that you |
2225 | write are your property, including scripts that make use of |
2226 | Berkeley DB. Neither the Perl license nor the Berkeley DB license |
2227 | place any restriction on what you may do with them. |
88108326 |
2228 | |
1f70e1ea |
2229 | If you are in any doubt about the license situation, contact either the |
2230 | Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details. |
a0b8c8c1 |
2231 | |
2232 | |
3b35bae3 |
2233 | =head1 SEE ALSO |
2234 | |
9fe6733a |
2235 | L<perl(1)>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>, |
2236 | L<dbmfilter> |
3b35bae3 |
2237 | |
3b35bae3 |
2238 | =head1 AUTHOR |
2239 | |
8e07c86e |
2240 | The DB_File interface was written by Paul Marquess |
6ca2e664 |
2241 | E<lt>Paul.Marquess@btinternet.comE<gt>. |
d3ef3b8a |
2242 | Questions about the DB system itself may be addressed to |
2243 | E<lt>db@sleepycat.com<gt>. |
3b35bae3 |
2244 | |
2245 | =cut |