1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (pmqs@cpan.org)
4 # last modified 11th November 2005
7 # Copyright (c) 1995-2005 Paul Marquess. All rights reserved.
8 # This program is free software; you can redistribute it and/or
9 # modify it under the same terms as Perl itself.
12 package DB_File::HASHINFO ;
20 @DB_File::HASHINFO::ISA = qw(Tie::Hash);
53 return $self->{GOT}{$key} if exists $self->{VALID}{$key} ;
56 croak "${pkg}::FETCH - Unknown element '$key'" ;
66 my $type = $self->{VALID}{$key};
70 croak "Key '$key' not associated with a code reference"
71 if $type == 2 && !ref $value && ref $value ne 'CODE';
72 $self->{GOT}{$key} = $value ;
77 croak "${pkg}::STORE - Unknown element '$key'" ;
85 if ( exists $self->{VALID}{$key} )
87 delete $self->{GOT}{$key} ;
92 croak "DB_File::HASHINFO::DELETE - Unknown element '$key'" ;
100 exists $self->{VALID}{$key} ;
108 croak ref($self) . " does not define the method ${method}" ;
111 sub FIRSTKEY { my $self = shift ; $self->NotHere("FIRSTKEY") }
112 sub NEXTKEY { my $self = shift ; $self->NotHere("NEXTKEY") }
113 sub CLEAR { my $self = shift ; $self->NotHere("CLEAR") }
115 package DB_File::RECNOINFO ;
120 @DB_File::RECNOINFO::ISA = qw(DB_File::HASHINFO) ;
126 bless { VALID => { map {$_, 1}
127 qw( bval cachesize psize flags lorder reclen bfname )
133 package DB_File::BTREEINFO ;
138 @DB_File::BTREEINFO::ISA = qw(DB_File::HASHINFO) ;
163 our ($VERSION, @ISA, @EXPORT, $AUTOLOAD, $DB_BTREE, $DB_HASH, $DB_RECNO);
164 our ($db_version, $use_XSLoader, $splice_end_array, $Error);
171 local $SIG{__WARN__} = sub {$splice_end_array = "@_";};
172 my @a =(1); splice(@a, 3);
174 ($splice_end_array =~ /^splice\(\) offset past end of array at /);
177 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
178 $DB_BTREE = new DB_File::BTREEINFO ;
179 $DB_HASH = new DB_File::HASHINFO ;
180 $DB_RECNO = new DB_File::RECNOINFO ;
187 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
192 @ISA = qw(DynaLoader);
196 push @ISA, qw(Tie::Hash Exporter);
198 $DB_BTREE $DB_HASH $DB_RECNO
233 ($constname = $AUTOLOAD) =~ s/.*:://;
234 my ($error, $val) = constant($constname);
235 Carp::croak $error if $error;
237 *{$AUTOLOAD} = sub { $val };
243 # Make all Fcntl O_XXX constants available for importing
245 my @O = grep /^O_/, @Fcntl::EXPORT;
246 Fcntl->import(@O); # first we import what we want to export
251 { XSLoader::load("DB_File", $VERSION)}
253 { bootstrap DB_File $VERSION }
255 # Preloaded methods go here. Autoload methods go after __END__, and are
256 # processed by the autosplit program.
258 sub tie_hash_or_array
261 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
263 $arg[4] = tied %{ $arg[4] }
264 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
266 $arg[2] = O_CREAT()|O_RDWR() if @arg >=3 && ! defined $arg[2];
267 $arg[3] = 0666 if @arg >=4 && ! defined $arg[3];
269 # make recno in Berkeley DB version 2 (or better) work like
270 # recno in version 1.
271 if ($db_version >= 4 and ! $tieHASH) {
272 $arg[2] |= O_CREAT();
275 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
276 $arg[1] and ! -e $arg[1]) {
277 open(FH, ">$arg[1]") or return undef ;
279 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
282 DoTie_($tieHASH, @arg) ;
287 tie_hash_or_array(@_) ;
292 tie_hash_or_array(@_) ;
300 my $status = $self->seq($key, $value, R_FIRST());
303 while ($status == 0) {
305 $status = $self->seq($key, $value, R_NEXT());
307 foreach $key (reverse @keys) {
308 my $s = $self->del($key);
318 my $current_length = $self->length() ;
320 if ($length < $current_length) {
322 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
325 elsif ($length > $current_length) {
326 $self->put($length-1, "") ;
335 if (not defined $offset) {
336 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
340 my $length = @_ ? shift : 0;
341 # Carping about definedness comes _after_ the OFFSET sanity check.
342 # This is so we get the same error messages as Perl's splice().
347 my $size = $self->FETCHSIZE();
349 # 'If OFFSET is negative then it start that far from the end of
353 my $new_offset = $size + $offset;
354 if ($new_offset < 0) {
355 die "Modification of non-creatable array value attempted, "
356 . "subscript $offset";
358 $offset = $new_offset;
361 if (not defined $length) {
362 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
366 if ($offset > $size) {
368 warnings::warnif('misc', 'splice() offset past end of array')
369 if $splice_end_array;
372 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
373 if (not defined $length) {
374 $length = $size - $offset;
377 # 'If LENGTH is negative, leave that many elements off the end of
381 $length = $size - $offset + $length;
384 # The user must have specified a length bigger than the
385 # length of the array passed in. But perl's splice()
386 # doesn't catch this, it just behaves as for length=0.
392 if ($length > $size - $offset) {
393 $length = $size - $offset;
396 # $num_elems holds the current number of elements in the database.
397 my $num_elems = $size;
399 # 'Removes the elements designated by OFFSET and LENGTH from an
403 foreach (0 .. $length - 1) {
405 my $status = $self->get($offset, $old);
407 my $msg = "error from Berkeley DB on get($offset, \$old)";
409 $msg .= ' (no such element?)';
412 $msg .= ": error status $status";
413 if (defined $! and $! ne '') {
414 $msg .= ", message $!";
421 $status = $self->del($offset);
423 my $msg = "error from Berkeley DB on del($offset)";
425 $msg .= ' (no such element?)';
428 $msg .= ": error status $status";
429 if (defined $! and $! ne '') {
430 $msg .= ", message $!";
439 # ...'and replaces them with the elements of LIST, if any.'
441 while (defined (my $elem = shift @list)) {
444 if ($pos >= $num_elems) {
445 $status = $self->put($pos, $elem);
448 $status = $self->put($pos, $elem, $self->R_IBEFORE);
452 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
454 $msg .= ' (no such element?)';
457 $msg .= ", error status $status";
458 if (defined $! and $! ne '') {
459 $msg .= ", message $!";
465 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
473 # 'In list context, returns the elements removed from the
478 elsif (defined wantarray and not wantarray) {
479 # 'In scalar context, returns the last element removed, or
480 # undef if no elements are removed.'
483 my $last = pop @removed;
490 elsif (not defined wantarray) {
495 sub ::DB_File::splice { &SPLICE }
499 croak "Usage: \$db->find_dup(key,value)\n"
503 my ($origkey, $value_wanted) = @_ ;
504 my ($key, $value) = ($origkey, 0);
507 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
509 $status = $db->seq($key, $value, R_NEXT() ) ) {
511 return 0 if $key eq $origkey and $value eq $value_wanted ;
519 croak "Usage: \$db->del_dup(key,value)\n"
523 my ($key, $value) = @_ ;
524 my ($status) = $db->find_dup($key, $value) ;
525 return $status if $status != 0 ;
527 $status = $db->del($key, R_CURSOR() ) ;
533 croak "Usage: \$db->get_dup(key [,flag])\n"
534 unless @_ == 2 or @_ == 3 ;
541 my $wantarray = wantarray ;
547 # iterate through the database until either EOF ($status == 0)
548 # or a different key is encountered ($key ne $origkey).
549 for ($status = $db->seq($key, $value, R_CURSOR()) ;
550 $status == 0 and $key eq $origkey ;
551 $status = $db->seq($key, $value, R_NEXT()) ) {
553 # save the value or count number of matches
556 { ++ $values{$value} }
558 { push (@values, $value) }
565 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
574 DB_File - Perl5 access to Berkeley DB version 1.x
580 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
581 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
582 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
584 $status = $X->del($key [, $flags]) ;
585 $status = $X->put($key, $value [, $flags]) ;
586 $status = $X->get($key, $value [, $flags]) ;
587 $status = $X->seq($key, $value, $flags) ;
588 $status = $X->sync([$flags]) ;
592 $count = $X->get_dup($key) ;
593 @list = $X->get_dup($key) ;
594 %list = $X->get_dup($key, 1) ;
595 $status = $X->find_dup($key, $value) ;
596 $status = $X->del_dup($key, $value) ;
604 @r = $X->splice(offset, length, elements);
607 $old_filter = $db->filter_store_key ( sub { ... } ) ;
608 $old_filter = $db->filter_store_value( sub { ... } ) ;
609 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
610 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
617 B<DB_File> is a module which allows Perl programs to make use of the
618 facilities provided by Berkeley DB version 1.x (if you have a newer
619 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
620 It is assumed that you have a copy of the Berkeley DB manual pages at
621 hand when reading this documentation. The interface defined here
622 mirrors the Berkeley DB interface closely.
624 Berkeley DB is a C library which provides a consistent interface to a
625 number of database formats. B<DB_File> provides an interface to all
626 three of the database types currently supported by Berkeley DB.
634 This database type allows arbitrary key/value pairs to be stored in data
635 files. This is equivalent to the functionality provided by other
636 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
637 the files created using DB_HASH are not compatible with any of the
638 other packages mentioned.
640 A default hashing algorithm, which will be adequate for most
641 applications, is built into Berkeley DB. If you do need to use your own
642 hashing algorithm it is possible to write your own in Perl and have
643 B<DB_File> use it instead.
647 The btree format allows arbitrary key/value pairs to be stored in a
648 sorted, balanced binary tree.
650 As with the DB_HASH format, it is possible to provide a user defined
651 Perl routine to perform the comparison of keys. By default, though, the
652 keys are stored in lexical order.
656 DB_RECNO allows both fixed-length and variable-length flat text files
657 to be manipulated using the same key/value pair interface as in DB_HASH
658 and DB_BTREE. In this case the key will consist of a record (line)
663 =head2 Using DB_File with Berkeley DB version 2 or greater
665 Although B<DB_File> is intended to be used with Berkeley DB version 1,
666 it can also be used with version 2, 3 or 4. In this case the interface is
667 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
668 version 2 or greater interface differs, B<DB_File> arranges for it to work
669 like version 1. This feature allows B<DB_File> scripts that were built
670 with version 1 to be migrated to version 2 or greater without any changes.
672 If you want to make use of the new features available in Berkeley DB
673 2.x or greater, use the Perl module B<BerkeleyDB> instead.
675 B<Note:> The database file format has changed multiple times in Berkeley
676 DB version 2, 3 and 4. If you cannot recreate your databases, you
677 must dump any existing databases with either the C<db_dump> or the
678 C<db_dump185> utility that comes with Berkeley DB.
679 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
680 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
681 documentation for further details.
683 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
686 =head2 Interface to Berkeley DB
688 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
689 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
690 allows B<DB_File> to access Berkeley DB files using either an
691 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
692 array (for the DB_RECNO file type).
694 In addition to the tie() interface, it is also possible to access most
695 of the functions provided in the Berkeley DB API directly.
696 See L<THE API INTERFACE>.
698 =head2 Opening a Berkeley DB Database File
700 Berkeley DB uses the function dbopen() to open or create a database.
701 Here is the C prototype for dbopen():
704 dbopen (const char * file, int flags, int mode,
705 DBTYPE type, const void * openinfo)
707 The parameter C<type> is an enumeration which specifies which of the 3
708 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
709 Depending on which of these is actually chosen, the final parameter,
710 I<openinfo> points to a data structure which allows tailoring of the
711 specific interface method.
713 This interface is handled slightly differently in B<DB_File>. Here is
714 an equivalent call using B<DB_File>:
716 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
718 The C<filename>, C<flags> and C<mode> parameters are the direct
719 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
720 performs the function of both the C<type> and C<openinfo> parameters in
723 In the example above $DB_HASH is actually a pre-defined reference to a
724 hash object. B<DB_File> has three of these pre-defined references.
725 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
727 The keys allowed in each of these pre-defined references is limited to
728 the names used in the equivalent C structure. So, for example, the
729 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
730 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
732 To change one of these elements, just assign to it like this:
734 $DB_HASH->{'cachesize'} = 10000 ;
736 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
737 usually adequate for most applications. If you do need to create extra
738 instances of these objects, constructors are available for each file
741 Here are examples of the constructors and the valid options available
742 for DB_HASH, DB_BTREE and DB_RECNO respectively.
744 $a = new DB_File::HASHINFO ;
752 $b = new DB_File::BTREEINFO ;
762 $c = new DB_File::RECNOINFO ;
771 The values stored in the hashes above are mostly the direct equivalent
772 of their C counterpart. Like their C counterparts, all are set to a
773 default values - that means you don't have to set I<all> of the
774 values when you only want to change one. Here is an example:
776 $a = new DB_File::HASHINFO ;
777 $a->{'cachesize'} = 12345 ;
778 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
780 A few of the options need extra discussion here. When used, the C
781 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
782 to C functions. In B<DB_File> these keys are used to store references
783 to Perl subs. Below are templates for each of the subs:
789 # return the hash value for $data
795 my ($key, $key2) = @_ ;
797 # return 0 if $key1 eq $key2
798 # -1 if $key1 lt $key2
799 # 1 if $key1 gt $key2
800 return (-1 , 0 or 1) ;
805 my ($key, $key2) = @_ ;
807 # return number of bytes of $key2 which are
808 # necessary to determine that it is greater than $key1
812 See L<Changing the BTREE sort order> for an example of using the
815 If you are using the DB_RECNO interface and you intend making use of
816 C<bval>, you should check out L<The 'bval' Option>.
818 =head2 Default Parameters
820 It is possible to omit some or all of the final 4 parameters in the
821 call to C<tie> and let them take default values. As DB_HASH is the most
822 common file format used, the call:
824 tie %A, "DB_File", "filename" ;
828 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
830 It is also possible to omit the filename parameter as well, so the
837 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
839 See L<In Memory Databases> for a discussion on the use of C<undef>
840 in place of a filename.
842 =head2 In Memory Databases
844 Berkeley DB allows the creation of in-memory databases by using NULL
845 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
846 uses C<undef> instead of NULL to provide this functionality.
850 The DB_HASH file format is probably the most commonly used of the three
851 file formats that B<DB_File> supports. It is also very straightforward
854 =head2 A Simple Example
856 This example shows how to create a database, add key/value pairs to the
857 database, delete keys/value pairs and finally how to enumerate the
858 contents of the database.
866 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
867 or die "Cannot open file 'fruit': $!\n";
869 # Add a few key/value pairs to the file
870 $h{"apple"} = "red" ;
871 $h{"orange"} = "orange" ;
872 $h{"banana"} = "yellow" ;
873 $h{"tomato"} = "red" ;
875 # Check for existence of a key
876 print "Banana Exists\n\n" if $h{"banana"} ;
878 # Delete a key/value pair.
881 # print the contents of the file
882 while (($k, $v) = each %h)
883 { print "$k -> $v\n" }
895 Note that the like ordinary associative arrays, the order of the keys
896 retrieved is in an apparently random order.
900 The DB_BTREE format is useful when you want to store data in a given
901 order. By default the keys will be stored in lexical order, but as you
902 will see from the example shown in the next section, it is very easy to
903 define your own sorting function.
905 =head2 Changing the BTREE sort order
907 This script shows how to override the default sorting algorithm that
908 BTREE uses. Instead of using the normal lexical ordering, a case
909 insensitive compare function will be used.
919 my ($key1, $key2) = @_ ;
920 "\L$key1" cmp "\L$key2" ;
923 # specify the Perl sub that will do the comparison
924 $DB_BTREE->{'compare'} = \&Compare ;
927 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
928 or die "Cannot open file 'tree': $!\n" ;
930 # Add a key/value pair to the file
931 $h{'Wall'} = 'Larry' ;
932 $h{'Smith'} = 'John' ;
933 $h{'mouse'} = 'mickey' ;
934 $h{'duck'} = 'donald' ;
939 # Cycle through the keys printing them in order.
940 # Note it is not necessary to sort the keys as
941 # the btree will have kept them in order automatically.
947 Here is the output from the code above.
953 There are a few point to bear in mind if you want to change the
954 ordering in a BTREE database:
960 The new compare function must be specified when you create the database.
964 You cannot change the ordering once the database has been created. Thus
965 you must use the same compare function every time you access the
970 Duplicate keys are entirely defined by the comparison function.
971 In the case-insensitive example above, the keys: 'KEY' and 'key'
972 would be considered duplicates, and assigning to the second one
973 would overwrite the first. If duplicates are allowed for (with the
974 R_DUP flag discussed below), only a single copy of duplicate keys
975 is stored in the database --- so (again with example above) assigning
976 three values to the keys: 'KEY', 'Key', and 'key' would leave just
977 the first key: 'KEY' in the database with three values. For some
978 situations this results in information loss, so care should be taken
979 to provide fully qualified comparison functions when necessary.
980 For example, the above comparison routine could be modified to
981 additionally compare case-sensitively if two keys are equal in the
982 case insensitive comparison:
985 my($key1, $key2) = @_;
986 lc $key1 cmp lc $key2 ||
990 And now you will only have duplicates when the keys themselves
991 are truly the same. (note: in versions of the db library prior to
992 about November 1996, such duplicate keys were retained so it was
993 possible to recover the original keys in sets of keys that
999 =head2 Handling Duplicate Keys
1001 The BTREE file type optionally allows a single key to be associated
1002 with an arbitrary number of values. This option is enabled by setting
1003 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
1005 There are some difficulties in using the tied hash interface if you
1006 want to manipulate a BTREE database with duplicate keys. Consider this
1013 my ($filename, %h) ;
1015 $filename = "tree" ;
1018 # Enable duplicate records
1019 $DB_BTREE->{'flags'} = R_DUP ;
1021 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1022 or die "Cannot open $filename: $!\n";
1024 # Add some key/value pairs to the file
1025 $h{'Wall'} = 'Larry' ;
1026 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1027 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1028 $h{'Smith'} = 'John' ;
1029 $h{'mouse'} = 'mickey' ;
1031 # iterate through the associative array
1032 # and print each key/value pair.
1033 foreach (sort keys %h)
1034 { print "$_ -> $h{$_}\n" }
1046 As you can see 3 records have been successfully created with key C<Wall>
1047 - the only thing is, when they are retrieved from the database they
1048 I<seem> to have the same value, namely C<Larry>. The problem is caused
1049 by the way that the associative array interface works. Basically, when
1050 the associative array interface is used to fetch the value associated
1051 with a given key, it will only ever retrieve the first value.
1053 Although it may not be immediately obvious from the code above, the
1054 associative array interface can be used to write values with duplicate
1055 keys, but it cannot be used to read them back from the database.
1057 The way to get around this problem is to use the Berkeley DB API method
1058 called C<seq>. This method allows sequential access to key/value
1059 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1060 and the API in general.
1062 Here is the script above rewritten using the C<seq> API method.
1068 my ($filename, $x, %h, $status, $key, $value) ;
1070 $filename = "tree" ;
1073 # Enable duplicate records
1074 $DB_BTREE->{'flags'} = R_DUP ;
1076 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1077 or die "Cannot open $filename: $!\n";
1079 # Add some key/value pairs to the file
1080 $h{'Wall'} = 'Larry' ;
1081 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1082 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1083 $h{'Smith'} = 'John' ;
1084 $h{'mouse'} = 'mickey' ;
1086 # iterate through the btree using seq
1087 # and print each key/value pair.
1089 for ($status = $x->seq($key, $value, R_FIRST) ;
1091 $status = $x->seq($key, $value, R_NEXT) )
1092 { print "$key -> $value\n" }
1105 This time we have got all the key/value pairs, including the multiple
1106 values associated with the key C<Wall>.
1108 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1109 a few utility methods.
1111 =head2 The get_dup() Method
1113 The C<get_dup> method assists in
1114 reading duplicate values from BTREE databases. The method can take the
1117 $count = $x->get_dup($key) ;
1118 @list = $x->get_dup($key) ;
1119 %list = $x->get_dup($key, 1) ;
1121 In a scalar context the method returns the number of values associated
1122 with the key, C<$key>.
1124 In list context, it returns all the values which match C<$key>. Note
1125 that the values will be returned in an apparently random order.
1127 In list context, if the second parameter is present and evaluates
1128 TRUE, the method returns an associative array. The keys of the
1129 associative array correspond to the values that matched in the BTREE
1130 and the values of the array are a count of the number of times that
1131 particular value occurred in the BTREE.
1133 So assuming the database created above, we can use C<get_dup> like
1140 my ($filename, $x, %h) ;
1142 $filename = "tree" ;
1144 # Enable duplicate records
1145 $DB_BTREE->{'flags'} = R_DUP ;
1147 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1148 or die "Cannot open $filename: $!\n";
1150 my $cnt = $x->get_dup("Wall") ;
1151 print "Wall occurred $cnt times\n" ;
1153 my %hash = $x->get_dup("Wall", 1) ;
1154 print "Larry is there\n" if $hash{'Larry'} ;
1155 print "There are $hash{'Brick'} Brick Walls\n" ;
1157 my @list = sort $x->get_dup("Wall") ;
1158 print "Wall => [@list]\n" ;
1160 @list = $x->get_dup("Smith") ;
1161 print "Smith => [@list]\n" ;
1163 @list = $x->get_dup("Dog") ;
1164 print "Dog => [@list]\n" ;
1169 Wall occurred 3 times
1171 There are 2 Brick Walls
1172 Wall => [Brick Brick Larry]
1176 =head2 The find_dup() Method
1178 $status = $X->find_dup($key, $value) ;
1180 This method checks for the existence of a specific key/value pair. If the
1181 pair exists, the cursor is left pointing to the pair and the method
1182 returns 0. Otherwise the method returns a non-zero value.
1184 Assuming the database from the previous example:
1190 my ($filename, $x, %h, $found) ;
1192 $filename = "tree" ;
1194 # Enable duplicate records
1195 $DB_BTREE->{'flags'} = R_DUP ;
1197 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1198 or die "Cannot open $filename: $!\n";
1200 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1201 print "Larry Wall is $found there\n" ;
1203 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1204 print "Harry Wall is $found there\n" ;
1212 Harry Wall is not there
1215 =head2 The del_dup() Method
1217 $status = $X->del_dup($key, $value) ;
1219 This method deletes a specific key/value pair. It returns
1220 0 if they exist and have been deleted successfully.
1221 Otherwise the method returns a non-zero value.
1223 Again assuming the existence of the C<tree> database
1229 my ($filename, $x, %h, $found) ;
1231 $filename = "tree" ;
1233 # Enable duplicate records
1234 $DB_BTREE->{'flags'} = R_DUP ;
1236 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1237 or die "Cannot open $filename: $!\n";
1239 $x->del_dup("Wall", "Larry") ;
1241 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1242 print "Larry Wall is $found there\n" ;
1249 Larry Wall is not there
1251 =head2 Matching Partial Keys
1253 The BTREE interface has a feature which allows partial keys to be
1254 matched. This functionality is I<only> available when the C<seq> method
1255 is used along with the R_CURSOR flag.
1257 $x->seq($key, $value, R_CURSOR) ;
1259 Here is the relevant quote from the dbopen man page where it defines
1260 the use of the R_CURSOR flag with seq:
1262 Note, for the DB_BTREE access method, the returned key is not
1263 necessarily an exact match for the specified key. The returned key
1264 is the smallest key greater than or equal to the specified key,
1265 permitting partial key matches and range searches.
1267 In the example script below, the C<match> sub uses this feature to find
1268 and print the first matching key/value pair given a partial key.
1275 my ($filename, $x, %h, $st, $key, $value) ;
1281 my $orig_key = $key ;
1282 $x->seq($key, $value, R_CURSOR) ;
1283 print "$orig_key\t-> $key\t-> $value\n" ;
1286 $filename = "tree" ;
1289 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1290 or die "Cannot open $filename: $!\n";
1292 # Add some key/value pairs to the file
1293 $h{'mouse'} = 'mickey' ;
1294 $h{'Wall'} = 'Larry' ;
1295 $h{'Walls'} = 'Brick' ;
1296 $h{'Smith'} = 'John' ;
1300 print "IN ORDER\n" ;
1301 for ($st = $x->seq($key, $value, R_FIRST) ;
1303 $st = $x->seq($key, $value, R_NEXT) )
1305 { print "$key -> $value\n" }
1307 print "\nPARTIAL MATCH\n" ;
1327 a -> mouse -> mickey
1331 DB_RECNO provides an interface to flat text files. Both variable and
1332 fixed length records are supported.
1334 In order to make RECNO more compatible with Perl, the array offset for
1335 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1337 As with normal Perl arrays, a RECNO array can be accessed using
1338 negative indexes. The index -1 refers to the last element of the array,
1339 -2 the second last, and so on. Attempting to access an element before
1340 the start of the array will raise a fatal run-time error.
1342 =head2 The 'bval' Option
1344 The operation of the bval option warrants some discussion. Here is the
1345 definition of bval from the Berkeley DB 1.85 recno manual page:
1347 The delimiting byte to be used to mark the end of a
1348 record for variable-length records, and the pad charac-
1349 ter for fixed-length records. If no value is speci-
1350 fied, newlines (``\n'') are used to mark the end of
1351 variable-length records and fixed-length records are
1354 The second sentence is wrong. In actual fact bval will only default to
1355 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1356 openinfo parameter is used at all, the value that happens to be in bval
1357 will be used. That means you always have to specify bval when making
1358 use of any of the options in the openinfo parameter. This documentation
1359 error will be fixed in the next release of Berkeley DB.
1361 That clarifies the situation with regards Berkeley DB itself. What
1362 about B<DB_File>? Well, the behavior defined in the quote above is
1363 quite useful, so B<DB_File> conforms to it.
1365 That means that you can specify other options (e.g. cachesize) and
1366 still have bval default to C<"\n"> for variable length records, and
1367 space for fixed length records.
1369 Also note that the bval option only allows you to specify a single byte
1372 =head2 A Simple Example
1374 Here is a simple example that uses RECNO (if you are using a version
1375 of Perl earlier than 5.004_57 this example won't work -- see
1376 L<Extra RECNO Methods> for a workaround).
1382 my $filename = "text" ;
1386 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1387 or die "Cannot open file 'text': $!\n" ;
1389 # Add a few key/value pairs to the file
1394 push @h, "green", "black" ;
1396 my $elements = scalar @h ;
1397 print "The array contains $elements entries\n" ;
1400 print "popped $last\n" ;
1402 unshift @h, "white" ;
1403 my $first = shift @h ;
1404 print "shifted $first\n" ;
1406 # Check for existence of a key
1407 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1409 # use a negative index
1410 print "The last element is $h[-1]\n" ;
1411 print "The 2nd last element is $h[-2]\n" ;
1415 Here is the output from the script:
1417 The array contains 5 entries
1420 Element 1 Exists with value blue
1421 The last element is green
1422 The 2nd last element is yellow
1424 =head2 Extra RECNO Methods
1426 If you are using a version of Perl earlier than 5.004_57, the tied
1427 array interface is quite limited. In the example script above
1428 C<push>, C<pop>, C<shift>, C<unshift>
1429 or determining the array length will not work with a tied array.
1431 To make the interface more useful for older versions of Perl, a number
1432 of methods are supplied with B<DB_File> to simulate the missing array
1433 operations. All these methods are accessed via the object returned from
1436 Here are the methods:
1440 =item B<$X-E<gt>push(list) ;>
1442 Pushes the elements of C<list> to the end of the array.
1444 =item B<$value = $X-E<gt>pop ;>
1446 Removes and returns the last element of the array.
1448 =item B<$X-E<gt>shift>
1450 Removes and returns the first element of the array.
1452 =item B<$X-E<gt>unshift(list) ;>
1454 Pushes the elements of C<list> to the start of the array.
1456 =item B<$X-E<gt>length>
1458 Returns the number of elements in the array.
1460 =item B<$X-E<gt>splice(offset, length, elements);>
1462 Returns a splice of the array.
1466 =head2 Another Example
1468 Here is a more complete example that makes use of some of the methods
1469 described above. It also makes use of the API interface directly (see
1470 L<THE API INTERFACE>).
1474 my (@h, $H, $file, $i) ;
1482 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1483 or die "Cannot open file $file: $!\n" ;
1485 # first create a text file to play with
1493 # Print the records in order.
1495 # The length method is needed here because evaluating a tied
1496 # array in a scalar context does not return the number of
1497 # elements in the array.
1499 print "\nORIGINAL\n" ;
1500 foreach $i (0 .. $H->length - 1) {
1501 print "$i: $h[$i]\n" ;
1504 # use the push & pop methods
1507 print "\nThe last record was [$a]\n" ;
1509 # and the shift & unshift methods
1511 $H->unshift("first") ;
1512 print "The first record was [$a]\n" ;
1514 # Use the API to add a new record after record 2.
1516 $H->put($i, "Newbie", R_IAFTER) ;
1518 # and a new record before record 1.
1520 $H->put($i, "New One", R_IBEFORE) ;
1525 # now print the records in reverse order
1526 print "\nREVERSE\n" ;
1527 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1528 { print "$i: $h[$i]\n" }
1530 # same again, but use the API functions instead
1531 print "\nREVERSE again\n" ;
1532 my ($s, $k, $v) = (0, 0, 0) ;
1533 for ($s = $H->seq($k, $v, R_LAST) ;
1535 $s = $H->seq($k, $v, R_PREV))
1536 { print "$k: $v\n" }
1541 and this is what it outputs:
1550 The last record was [four]
1551 The first record was [zero]
1575 Rather than iterating through the array, C<@h> like this:
1579 it is necessary to use either this:
1581 foreach $i (0 .. $H->length - 1)
1585 for ($a = $H->get($k, $v, R_FIRST) ;
1587 $a = $H->get($k, $v, R_NEXT) )
1591 Notice that both times the C<put> method was used the record index was
1592 specified using a variable, C<$i>, rather than the literal value
1593 itself. This is because C<put> will return the record number of the
1594 inserted line via that parameter.
1598 =head1 THE API INTERFACE
1600 As well as accessing Berkeley DB using a tied hash or array, it is also
1601 possible to make direct use of most of the API functions defined in the
1602 Berkeley DB documentation.
1604 To do this you need to store a copy of the object returned from the tie.
1606 $db = tie %hash, "DB_File", "filename" ;
1608 Once you have done that, you can access the Berkeley DB API functions
1609 as B<DB_File> methods directly like this:
1611 $db->put($key, $value, R_NOOVERWRITE) ;
1613 B<Important:> If you have saved a copy of the object returned from
1614 C<tie>, the underlying database file will I<not> be closed until both
1615 the tied variable is untied and all copies of the saved object are
1619 $db = tie %hash, "DB_File", "filename"
1620 or die "Cannot tie filename: $!" ;
1625 See L<The untie() Gotcha> for more details.
1627 All the functions defined in L<dbopen> are available except for
1628 close() and dbopen() itself. The B<DB_File> method interface to the
1629 supported functions have been implemented to mirror the way Berkeley DB
1630 works whenever possible. In particular note that:
1636 The methods return a status value. All return 0 on success.
1637 All return -1 to signify an error and set C<$!> to the exact
1638 error code. The return code 1 generally (but not always) means that the
1639 key specified did not exist in the database.
1641 Other return codes are defined. See below and in the Berkeley DB
1642 documentation for details. The Berkeley DB documentation should be used
1643 as the definitive source.
1647 Whenever a Berkeley DB function returns data via one of its parameters,
1648 the equivalent B<DB_File> method does exactly the same.
1652 If you are careful, it is possible to mix API calls with the tied
1653 hash/array interface in the same piece of code. Although only a few of
1654 the methods used to implement the tied interface currently make use of
1655 the cursor, you should always assume that the cursor has been changed
1656 any time the tied hash/array interface is used. As an example, this
1657 code will probably not do what you expect:
1659 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1660 or die "Cannot tie $filename: $!" ;
1662 # Get the first key/value pair and set the cursor
1663 $X->seq($key, $value, R_FIRST) ;
1665 # this line will modify the cursor
1666 $count = scalar keys %x ;
1668 # Get the second key/value pair.
1669 # oops, it didn't, it got the last key/value pair!
1670 $X->seq($key, $value, R_NEXT) ;
1672 The code above can be rearranged to get around the problem, like this:
1674 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1675 or die "Cannot tie $filename: $!" ;
1677 # this line will modify the cursor
1678 $count = scalar keys %x ;
1680 # Get the first key/value pair and set the cursor
1681 $X->seq($key, $value, R_FIRST) ;
1683 # Get the second key/value pair.
1685 $X->seq($key, $value, R_NEXT) ;
1689 All the constants defined in L<dbopen> for use in the flags parameters
1690 in the methods defined below are also available. Refer to the Berkeley
1691 DB documentation for the precise meaning of the flags values.
1693 Below is a list of the methods available.
1697 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1699 Given a key (C<$key>) this method reads the value associated with it
1700 from the database. The value read from the database is returned in the
1701 C<$value> parameter.
1703 If the key does not exist the method returns 1.
1705 No flags are currently defined for this method.
1707 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1709 Stores the key/value pair in the database.
1711 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1712 will have the record number of the inserted key/value pair set.
1714 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1717 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1719 Removes all key/value pairs with key C<$key> from the database.
1721 A return code of 1 means that the requested key was not in the
1724 R_CURSOR is the only valid flag at present.
1726 =item B<$status = $X-E<gt>fd ;>
1728 Returns the file descriptor for the underlying database.
1730 See L<Locking: The Trouble with fd> for an explanation for why you should
1731 not use C<fd> to lock your database.
1733 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1735 This interface allows sequential retrieval from the database. See
1736 L<dbopen> for full details.
1738 Both the C<$key> and C<$value> parameters will be set to the key/value
1739 pair read from the database.
1741 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1742 R_FIRST, R_LAST, R_NEXT and R_PREV.
1744 =item B<$status = $X-E<gt>sync([$flags]) ;>
1746 Flushes any cached buffers to disk.
1748 R_RECNOSYNC is the only valid flag at present.
1754 A DBM Filter is a piece of code that is be used when you I<always>
1755 want to make the same transformation to all keys and/or values in a
1758 There are four methods associated with DBM Filters. All work identically,
1759 and each is used to install (or uninstall) a single DBM Filter. Each
1760 expects a single parameter, namely a reference to a sub. The only
1761 difference between them is the place that the filter is installed.
1767 =item B<filter_store_key>
1769 If a filter has been installed with this method, it will be invoked
1770 every time you write a key to a DBM database.
1772 =item B<filter_store_value>
1774 If a filter has been installed with this method, it will be invoked
1775 every time you write a value to a DBM database.
1778 =item B<filter_fetch_key>
1780 If a filter has been installed with this method, it will be invoked
1781 every time you read a key from a DBM database.
1783 =item B<filter_fetch_value>
1785 If a filter has been installed with this method, it will be invoked
1786 every time you read a value from a DBM database.
1790 You can use any combination of the methods, from none, to all four.
1792 All filter methods return the existing filter, if present, or C<undef>
1795 To delete a filter pass C<undef> to it.
1799 When each filter is called by Perl, a local copy of C<$_> will contain
1800 the key or value to be filtered. Filtering is achieved by modifying
1801 the contents of C<$_>. The return code from the filter is ignored.
1803 =head2 An Example -- the NULL termination problem.
1805 Consider the following scenario. You have a DBM database
1806 that you need to share with a third-party C application. The C application
1807 assumes that I<all> keys and values are NULL terminated. Unfortunately
1808 when Perl writes to DBM databases it doesn't use NULL termination, so
1809 your Perl application will have to manage NULL termination itself. When
1810 you write to the database you will have to use something like this:
1812 $hash{"$key\0"} = "$value\0" ;
1814 Similarly the NULL needs to be taken into account when you are considering
1815 the length of existing keys/values.
1817 It would be much better if you could ignore the NULL terminations issue
1818 in the main application code and have a mechanism that automatically
1819 added the terminating NULL to all keys and values whenever you write to
1820 the database and have them removed when you read from the database. As I'm
1821 sure you have already guessed, this is a problem that DBM Filters can
1829 my $filename = "filt" ;
1832 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1833 or die "Cannot open $filename: $!\n" ;
1835 # Install DBM Filters
1836 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1837 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1838 $db->filter_fetch_value( sub { s/\0$// } ) ;
1839 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1841 $hash{"abc"} = "def" ;
1842 my $a = $hash{"ABC"} ;
1847 Hopefully the contents of each of the filters should be
1848 self-explanatory. Both "fetch" filters remove the terminating NULL,
1849 and both "store" filters add a terminating NULL.
1852 =head2 Another Example -- Key is a C int.
1854 Here is another real-life example. By default, whenever Perl writes to
1855 a DBM database it always writes the key and value as strings. So when
1858 $hash{12345} = "something" ;
1860 the key 12345 will get stored in the DBM database as the 5 byte string
1861 "12345". If you actually want the key to be stored in the DBM database
1862 as a C int, you will have to use C<pack> when writing, and C<unpack>
1865 Here is a DBM Filter that does it:
1871 my $filename = "filt" ;
1875 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1876 or die "Cannot open $filename: $!\n" ;
1878 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1879 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1880 $hash{123} = "def" ;
1885 This time only two filters have been used -- we only need to manipulate
1886 the contents of the key, so it wasn't necessary to install any value
1889 =head1 HINTS AND TIPS
1892 =head2 Locking: The Trouble with fd
1894 Until version 1.72 of this module, the recommended technique for locking
1895 B<DB_File> databases was to flock the filehandle returned from the "fd"
1896 function. Unfortunately this technique has been shown to be fundamentally
1897 flawed (Kudos to David Harris for tracking this down). Use it at your own
1900 The locking technique went like this.
1902 $db = tie(%db, 'DB_File', 'foo.db', O_CREAT|O_RDWR, 0644)
1903 || die "dbcreat foo.db $!";
1905 open(DB_FH, "+<&=$fd") || die "dup $!";
1906 flock (DB_FH, LOCK_EX) || die "flock: $!";
1908 $db{"Tom"} = "Jerry" ;
1910 flock(DB_FH, LOCK_UN);
1915 In simple terms, this is what happens:
1921 Use "tie" to open the database.
1925 Lock the database with fd & flock.
1929 Read & Write to the database.
1933 Unlock and close the database.
1937 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1938 database in step 2 is that an initial block from the database will get
1939 read from disk and cached in memory.
1941 To see why this is a problem, consider what can happen when two processes,
1942 say "A" and "B", both want to update the same B<DB_File> database
1943 using the locking steps outlined above. Assume process "A" has already
1944 opened the database and has a write lock, but it hasn't actually updated
1945 the database yet (it has finished step 2, but not started step 3 yet). Now
1946 process "B" tries to open the same database - step 1 will succeed,
1947 but it will block on step 2 until process "A" releases the lock. The
1948 important thing to notice here is that at this point in time both
1949 processes will have cached identical initial blocks from the database.
1951 Now process "A" updates the database and happens to change some of the
1952 data held in the initial buffer. Process "A" terminates, flushing
1953 all cached data to disk and releasing the database lock. At this point
1954 the database on disk will correctly reflect the changes made by process
1957 With the lock released, process "B" can now continue. It also updates the
1958 database and unfortunately it too modifies the data that was in its
1959 initial buffer. Once that data gets flushed to disk it will overwrite
1960 some/all of the changes process "A" made to the database.
1962 The result of this scenario is at best a database that doesn't contain
1963 what you expect. At worst the database will corrupt.
1965 The above won't happen every time competing process update the same
1966 B<DB_File> database, but it does illustrate why the technique should
1969 =head2 Safe ways to lock a database
1971 Starting with version 2.x, Berkeley DB has internal support for locking.
1972 The companion module to this one, B<BerkeleyDB>, provides an interface
1973 to this locking functionality. If you are serious about locking
1974 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1976 If using B<BerkeleyDB> isn't an option, there are a number of modules
1977 available on CPAN that can be used to implement locking. Each one
1978 implements locking differently and has different goals in mind. It is
1979 therefore worth knowing the difference, so that you can pick the right
1980 one for your application. Here are the three locking wrappers:
1984 =item B<Tie::DB_Lock>
1986 A B<DB_File> wrapper which creates copies of the database file for
1987 read access, so that you have a kind of a multiversioning concurrent read
1988 system. However, updates are still serial. Use for databases where reads
1989 may be lengthy and consistency problems may occur.
1991 =item B<Tie::DB_LockFile>
1993 A B<DB_File> wrapper that has the ability to lock and unlock the database
1994 while it is being used. Avoids the tie-before-flock problem by simply
1995 re-tie-ing the database when you get or drop a lock. Because of the
1996 flexibility in dropping and re-acquiring the lock in the middle of a
1997 session, this can be massaged into a system that will work with long
1998 updates and/or reads if the application follows the hints in the POD
2001 =item B<DB_File::Lock>
2003 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
2004 before tie-ing the database and drops the lock after the untie. Allows
2005 one to use the same lockfile for multiple databases to avoid deadlock
2006 problems, if desired. Use for databases where updates are reads are
2007 quick and simple flock locking semantics are enough.
2011 =head2 Sharing Databases With C Applications
2013 There is no technical reason why a Berkeley DB database cannot be
2014 shared by both a Perl and a C application.
2016 The vast majority of problems that are reported in this area boil down
2017 to the fact that C strings are NULL terminated, whilst Perl strings are
2018 not. See L<DBM FILTERS> for a generic way to work around this problem.
2020 Here is a real example. Netscape 2.0 keeps a record of the locations you
2021 visit along with the time you last visited them in a DB_HASH database.
2022 This is usually stored in the file F<~/.netscape/history.db>. The key
2023 field in the database is the location string and the value field is the
2024 time the location was last visited stored as a 4 byte binary value.
2026 If you haven't already guessed, the location string is stored with a
2027 terminating NULL. This means you need to be careful when accessing the
2030 Here is a snippet of code that is loosely based on Tom Christiansen's
2031 I<ggh> script (available from your nearest CPAN archive in
2032 F<authors/id/TOMC/scripts/nshist.gz>).
2039 my ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2040 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2042 $HISTORY = "$dotdir/.netscape/history.db";
2044 tie %hist_db, 'DB_File', $HISTORY
2045 or die "Cannot open $HISTORY: $!\n" ;;
2047 # Dump the complete database
2048 while ( ($href, $binary_time) = each %hist_db ) {
2050 # remove the terminating NULL
2051 $href =~ s/\x00$// ;
2053 # convert the binary time into a user friendly string
2054 $date = localtime unpack("V", $binary_time);
2055 print "$date $href\n" ;
2058 # check for the existence of a specific key
2059 # remember to add the NULL
2060 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2061 $date = localtime unpack("V", $binary_time) ;
2062 print "Last visited mox.perl.com on $date\n" ;
2065 print "Never visited mox.perl.com\n"
2070 =head2 The untie() Gotcha
2072 If you make use of the Berkeley DB API, it is I<very> strongly
2073 recommended that you read L<perltie/The untie Gotcha>.
2075 Even if you don't currently make use of the API interface, it is still
2078 Here is an example which illustrates the problem from a B<DB_File>
2087 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2088 or die "Cannot tie first time: $!" ;
2094 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2095 or die "Cannot tie second time: $!" ;
2099 When run, the script will produce this error message:
2101 Cannot tie second time: Invalid argument at bad.file line 14.
2103 Although the error message above refers to the second tie() statement
2104 in the script, the source of the problem is really with the untie()
2105 statement that precedes it.
2107 Having read L<perltie> you will probably have already guessed that the
2108 error is caused by the extra copy of the tied object stored in C<$X>.
2109 If you haven't, then the problem boils down to the fact that the
2110 B<DB_File> destructor, DESTROY, will not be called until I<all>
2111 references to the tied object are destroyed. Both the tied variable,
2112 C<%x>, and C<$X> above hold a reference to the object. The call to
2113 untie() will destroy the first, but C<$X> still holds a valid
2114 reference, so the destructor will not get called and the database file
2115 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2116 attempt to open a database that is already open via the catch-all
2117 "Invalid argument" doesn't help.
2119 If you run the script with the C<-w> flag the error message becomes:
2121 untie attempted while 1 inner references still exist at bad.file line 12.
2122 Cannot tie second time: Invalid argument at bad.file line 14.
2124 which pinpoints the real problem. Finally the script can now be
2125 modified to fix the original problem by destroying the API object
2134 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2138 =head1 COMMON QUESTIONS
2140 =head2 Why is there Perl source in my database?
2142 If you look at the contents of a database file created by DB_File,
2143 there can sometimes be part of a Perl script included in it.
2145 This happens because Berkeley DB uses dynamic memory to allocate
2146 buffers which will subsequently be written to the database file. Being
2147 dynamic, the memory could have been used for anything before DB
2148 malloced it. As Berkeley DB doesn't clear the memory once it has been
2149 allocated, the unused portions will contain random junk. In the case
2150 where a Perl script gets written to the database, the random junk will
2151 correspond to an area of dynamic memory that happened to be used during
2152 the compilation of the script.
2154 Unless you don't like the possibility of there being part of your Perl
2155 scripts embedded in a database file, this is nothing to worry about.
2157 =head2 How do I store complex data structures with DB_File?
2159 Although B<DB_File> cannot do this directly, there is a module which
2160 can layer transparently over B<DB_File> to accomplish this feat.
2162 Check out the MLDBM module, available on CPAN in the directory
2163 F<modules/by-module/MLDBM>.
2165 =head2 What does "Invalid Argument" mean?
2167 You will get this error message when one of the parameters in the
2168 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2169 get wrong, so it can be difficult to figure out which one it is.
2171 Here are a couple of possibilities:
2177 Attempting to reopen a database without closing it.
2181 Using the O_WRONLY flag.
2185 =head2 What does "Bareword 'DB_File' not allowed" mean?
2187 You will encounter this particular error message when you have the
2188 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2189 Consider this script:
2195 tie %x, DB_File, "filename" ;
2197 Running it produces the error in question:
2199 Bareword "DB_File" not allowed while "strict subs" in use
2201 To get around the error, place the word C<DB_File> in either single or
2202 double quotes, like this:
2204 tie %x, "DB_File", "filename" ;
2206 Although it might seem like a real pain, it is really worth the effort
2207 of having a C<use strict> in all your scripts.
2211 Articles that are either about B<DB_File> or make use of it.
2217 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2218 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2224 Moved to the Changes file.
2228 Some older versions of Berkeley DB had problems with fixed length
2229 records using the RECNO file format. This problem has been fixed since
2230 version 1.85 of Berkeley DB.
2232 I am sure there are bugs in the code. If you do find any, or can
2233 suggest any enhancements, I would welcome your comments.
2237 B<DB_File> comes with the standard Perl source distribution. Look in
2238 the directory F<ext/DB_File>. Given the amount of time between releases
2239 of Perl the version that ships with Perl is quite likely to be out of
2240 date, so the most recent version can always be found on CPAN (see
2241 L<perlmodlib/CPAN> for details), in the directory
2242 F<modules/by-module/DB_File>.
2244 This version of B<DB_File> will work with either version 1.x, 2.x or
2245 3.x of Berkeley DB, but is limited to the functionality provided by
2248 The official web site for Berkeley DB is F<http://www.sleepycat.com>.
2249 All versions of Berkeley DB are available there.
2251 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2252 archive in F<src/misc/db.1.85.tar.gz>.
2254 If you are running IRIX, then get Berkeley DB version 1 from
2255 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2256 compile properly on IRIX 5.3.
2260 Copyright (c) 1995-2005 Paul Marquess. All rights reserved. This program
2261 is free software; you can redistribute it and/or modify it under the
2262 same terms as Perl itself.
2264 Although B<DB_File> is covered by the Perl license, the library it
2265 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2266 copyright and its own license. Please take the time to read it.
2268 Here are are few words taken from the Berkeley DB FAQ (at
2269 F<http://www.sleepycat.com>) regarding the license:
2271 Do I have to license DB to use it in Perl scripts?
2273 No. The Berkeley DB license requires that software that uses
2274 Berkeley DB be freely redistributable. In the case of Perl, that
2275 software is Perl, and not your scripts. Any Perl scripts that you
2276 write are your property, including scripts that make use of
2277 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2278 place any restriction on what you may do with them.
2280 If you are in any doubt about the license situation, contact either the
2281 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2286 L<perl>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2291 The DB_File interface was written by Paul Marquess
2292 E<lt>pmqs@cpan.orgE<gt>.
2293 Questions about the DB system itself may be addressed to
2294 E<lt>db@sleepycat.comE<gt>.