1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (pmqs@cpan.org)
4 # last modified 28th October 2007
7 # Copyright (c) 1995-2009 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_no_length, $splice_end_array, $Error);
169 $VERSION = eval $VERSION; # needed for dev releases
172 local $SIG{__WARN__} = sub {$splice_end_array_no_length = join(" ",@_);};
173 my @a =(1); splice(@a, 3);
174 $splice_end_array_no_length =
175 ($splice_end_array_no_length =~ /^splice\(\) offset past end of array at /);
178 local $SIG{__WARN__} = sub {$splice_end_array = join(" ", @_);};
179 my @a =(1); splice(@a, 3, 1);
181 ($splice_end_array =~ /^splice\(\) offset past end of array at /);
184 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
185 $DB_BTREE = new DB_File::BTREEINFO ;
186 $DB_HASH = new DB_File::HASHINFO ;
187 $DB_RECNO = new DB_File::RECNOINFO ;
194 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
199 @ISA = qw(DynaLoader);
203 push @ISA, qw(Tie::Hash Exporter);
205 $DB_BTREE $DB_HASH $DB_RECNO
240 ($constname = $AUTOLOAD) =~ s/.*:://;
241 my ($error, $val) = constant($constname);
242 Carp::croak $error if $error;
244 *{$AUTOLOAD} = sub { $val };
250 # Make all Fcntl O_XXX constants available for importing
252 my @O = grep /^O_/, @Fcntl::EXPORT;
253 Fcntl->import(@O); # first we import what we want to export
258 { XSLoader::load("DB_File", $VERSION)}
260 { bootstrap DB_File $VERSION }
262 # Preloaded methods go here. Autoload methods go after __END__, and are
263 # processed by the autosplit program.
265 sub tie_hash_or_array
268 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
271 $arg[1] = File::Spec->rel2abs($arg[1])
274 $arg[4] = tied %{ $arg[4] }
275 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
277 $arg[2] = O_CREAT()|O_RDWR() if @arg >=3 && ! defined $arg[2];
278 $arg[3] = 0666 if @arg >=4 && ! defined $arg[3];
280 # make recno in Berkeley DB version 2 (or better) work like
281 # recno in version 1.
282 if ($db_version >= 4 and ! $tieHASH) {
283 $arg[2] |= O_CREAT();
286 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
287 $arg[1] and ! -e $arg[1]) {
288 open(FH, ">$arg[1]") or return undef ;
290 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
293 DoTie_($tieHASH, @arg) ;
298 tie_hash_or_array(@_) ;
303 tie_hash_or_array(@_) ;
311 my $status = $self->seq($key, $value, R_FIRST());
314 while ($status == 0) {
316 $status = $self->seq($key, $value, R_NEXT());
318 foreach $key (reverse @keys) {
319 my $s = $self->del($key);
329 my $current_length = $self->length() ;
331 if ($length < $current_length) {
333 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
336 elsif ($length > $current_length) {
337 $self->put($length-1, "") ;
346 if (not defined $offset) {
347 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
352 my $length = @_ ? shift : 0;
353 # Carping about definedness comes _after_ the OFFSET sanity check.
354 # This is so we get the same error messages as Perl's splice().
359 my $size = $self->FETCHSIZE();
361 # 'If OFFSET is negative then it start that far from the end of
365 my $new_offset = $size + $offset;
366 if ($new_offset < 0) {
367 die "Modification of non-creatable array value attempted, "
368 . "subscript $offset";
370 $offset = $new_offset;
373 if (not defined $length) {
374 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
378 if ($offset > $size) {
380 warnings::warnif('misc', 'splice() offset past end of array')
381 if $has_length ? $splice_end_array : $splice_end_array_no_length;
384 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
385 if (not defined $length) {
386 $length = $size - $offset;
389 # 'If LENGTH is negative, leave that many elements off the end of
393 $length = $size - $offset + $length;
396 # The user must have specified a length bigger than the
397 # length of the array passed in. But perl's splice()
398 # doesn't catch this, it just behaves as for length=0.
404 if ($length > $size - $offset) {
405 $length = $size - $offset;
408 # $num_elems holds the current number of elements in the database.
409 my $num_elems = $size;
411 # 'Removes the elements designated by OFFSET and LENGTH from an
415 foreach (0 .. $length - 1) {
417 my $status = $self->get($offset, $old);
419 my $msg = "error from Berkeley DB on get($offset, \$old)";
421 $msg .= ' (no such element?)';
424 $msg .= ": error status $status";
425 if (defined $! and $! ne '') {
426 $msg .= ", message $!";
433 $status = $self->del($offset);
435 my $msg = "error from Berkeley DB on del($offset)";
437 $msg .= ' (no such element?)';
440 $msg .= ": error status $status";
441 if (defined $! and $! ne '') {
442 $msg .= ", message $!";
451 # ...'and replaces them with the elements of LIST, if any.'
453 while (defined (my $elem = shift @list)) {
456 if ($pos >= $num_elems) {
457 $status = $self->put($pos, $elem);
460 $status = $self->put($pos, $elem, $self->R_IBEFORE);
464 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
466 $msg .= ' (no such element?)';
469 $msg .= ", error status $status";
470 if (defined $! and $! ne '') {
471 $msg .= ", message $!";
477 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
485 # 'In list context, returns the elements removed from the
490 elsif (defined wantarray and not wantarray) {
491 # 'In scalar context, returns the last element removed, or
492 # undef if no elements are removed.'
495 my $last = pop @removed;
502 elsif (not defined wantarray) {
507 sub ::DB_File::splice { &SPLICE }
511 croak "Usage: \$db->find_dup(key,value)\n"
515 my ($origkey, $value_wanted) = @_ ;
516 my ($key, $value) = ($origkey, 0);
519 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
521 $status = $db->seq($key, $value, R_NEXT() ) ) {
523 return 0 if $key eq $origkey and $value eq $value_wanted ;
531 croak "Usage: \$db->del_dup(key,value)\n"
535 my ($key, $value) = @_ ;
536 my ($status) = $db->find_dup($key, $value) ;
537 return $status if $status != 0 ;
539 $status = $db->del($key, R_CURSOR() ) ;
545 croak "Usage: \$db->get_dup(key [,flag])\n"
546 unless @_ == 2 or @_ == 3 ;
553 my $wantarray = wantarray ;
559 # iterate through the database until either EOF ($status == 0)
560 # or a different key is encountered ($key ne $origkey).
561 for ($status = $db->seq($key, $value, R_CURSOR()) ;
562 $status == 0 and $key eq $origkey ;
563 $status = $db->seq($key, $value, R_NEXT()) ) {
565 # save the value or count number of matches
568 { ++ $values{$value} }
570 { push (@values, $value) }
577 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
583 my $type = ref shift;
584 croak "Cannot freeze $type object\n";
589 my $type = ref shift;
590 croak "Cannot thaw $type object\n";
600 DB_File - Perl5 access to Berkeley DB version 1.x
606 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
607 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
608 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
610 $status = $X->del($key [, $flags]) ;
611 $status = $X->put($key, $value [, $flags]) ;
612 $status = $X->get($key, $value [, $flags]) ;
613 $status = $X->seq($key, $value, $flags) ;
614 $status = $X->sync([$flags]) ;
618 $count = $X->get_dup($key) ;
619 @list = $X->get_dup($key) ;
620 %list = $X->get_dup($key, 1) ;
621 $status = $X->find_dup($key, $value) ;
622 $status = $X->del_dup($key, $value) ;
630 @r = $X->splice(offset, length, elements);
633 $old_filter = $db->filter_store_key ( sub { ... } ) ;
634 $old_filter = $db->filter_store_value( sub { ... } ) ;
635 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
636 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
643 B<DB_File> is a module which allows Perl programs to make use of the
644 facilities provided by Berkeley DB version 1.x (if you have a newer
645 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
646 It is assumed that you have a copy of the Berkeley DB manual pages at
647 hand when reading this documentation. The interface defined here
648 mirrors the Berkeley DB interface closely.
650 Berkeley DB is a C library which provides a consistent interface to a
651 number of database formats. B<DB_File> provides an interface to all
652 three of the database types currently supported by Berkeley DB.
660 This database type allows arbitrary key/value pairs to be stored in data
661 files. This is equivalent to the functionality provided by other
662 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
663 the files created using DB_HASH are not compatible with any of the
664 other packages mentioned.
666 A default hashing algorithm, which will be adequate for most
667 applications, is built into Berkeley DB. If you do need to use your own
668 hashing algorithm it is possible to write your own in Perl and have
669 B<DB_File> use it instead.
673 The btree format allows arbitrary key/value pairs to be stored in a
674 sorted, balanced binary tree.
676 As with the DB_HASH format, it is possible to provide a user defined
677 Perl routine to perform the comparison of keys. By default, though, the
678 keys are stored in lexical order.
682 DB_RECNO allows both fixed-length and variable-length flat text files
683 to be manipulated using the same key/value pair interface as in DB_HASH
684 and DB_BTREE. In this case the key will consist of a record (line)
689 =head2 Using DB_File with Berkeley DB version 2 or greater
691 Although B<DB_File> is intended to be used with Berkeley DB version 1,
692 it can also be used with version 2, 3 or 4. In this case the interface is
693 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
694 version 2 or greater interface differs, B<DB_File> arranges for it to work
695 like version 1. This feature allows B<DB_File> scripts that were built
696 with version 1 to be migrated to version 2 or greater without any changes.
698 If you want to make use of the new features available in Berkeley DB
699 2.x or greater, use the Perl module B<BerkeleyDB> instead.
701 B<Note:> The database file format has changed multiple times in Berkeley
702 DB version 2, 3 and 4. If you cannot recreate your databases, you
703 must dump any existing databases with either the C<db_dump> or the
704 C<db_dump185> utility that comes with Berkeley DB.
705 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
706 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
707 documentation for further details.
709 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
712 =head2 Interface to Berkeley DB
714 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
715 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
716 allows B<DB_File> to access Berkeley DB files using either an
717 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
718 array (for the DB_RECNO file type).
720 In addition to the tie() interface, it is also possible to access most
721 of the functions provided in the Berkeley DB API directly.
722 See L<THE API INTERFACE>.
724 =head2 Opening a Berkeley DB Database File
726 Berkeley DB uses the function dbopen() to open or create a database.
727 Here is the C prototype for dbopen():
730 dbopen (const char * file, int flags, int mode,
731 DBTYPE type, const void * openinfo)
733 The parameter C<type> is an enumeration which specifies which of the 3
734 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
735 Depending on which of these is actually chosen, the final parameter,
736 I<openinfo> points to a data structure which allows tailoring of the
737 specific interface method.
739 This interface is handled slightly differently in B<DB_File>. Here is
740 an equivalent call using B<DB_File>:
742 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
744 The C<filename>, C<flags> and C<mode> parameters are the direct
745 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
746 performs the function of both the C<type> and C<openinfo> parameters in
749 In the example above $DB_HASH is actually a pre-defined reference to a
750 hash object. B<DB_File> has three of these pre-defined references.
751 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
753 The keys allowed in each of these pre-defined references is limited to
754 the names used in the equivalent C structure. So, for example, the
755 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
756 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
758 To change one of these elements, just assign to it like this:
760 $DB_HASH->{'cachesize'} = 10000 ;
762 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
763 usually adequate for most applications. If you do need to create extra
764 instances of these objects, constructors are available for each file
767 Here are examples of the constructors and the valid options available
768 for DB_HASH, DB_BTREE and DB_RECNO respectively.
770 $a = new DB_File::HASHINFO ;
778 $b = new DB_File::BTREEINFO ;
788 $c = new DB_File::RECNOINFO ;
797 The values stored in the hashes above are mostly the direct equivalent
798 of their C counterpart. Like their C counterparts, all are set to a
799 default values - that means you don't have to set I<all> of the
800 values when you only want to change one. Here is an example:
802 $a = new DB_File::HASHINFO ;
803 $a->{'cachesize'} = 12345 ;
804 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
806 A few of the options need extra discussion here. When used, the C
807 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
808 to C functions. In B<DB_File> these keys are used to store references
809 to Perl subs. Below are templates for each of the subs:
815 # return the hash value for $data
821 my ($key, $key2) = @_ ;
823 # return 0 if $key1 eq $key2
824 # -1 if $key1 lt $key2
825 # 1 if $key1 gt $key2
826 return (-1 , 0 or 1) ;
831 my ($key, $key2) = @_ ;
833 # return number of bytes of $key2 which are
834 # necessary to determine that it is greater than $key1
838 See L<Changing the BTREE sort order> for an example of using the
841 If you are using the DB_RECNO interface and you intend making use of
842 C<bval>, you should check out L<The 'bval' Option>.
844 =head2 Default Parameters
846 It is possible to omit some or all of the final 4 parameters in the
847 call to C<tie> and let them take default values. As DB_HASH is the most
848 common file format used, the call:
850 tie %A, "DB_File", "filename" ;
854 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
856 It is also possible to omit the filename parameter as well, so the
863 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
865 See L<In Memory Databases> for a discussion on the use of C<undef>
866 in place of a filename.
868 =head2 In Memory Databases
870 Berkeley DB allows the creation of in-memory databases by using NULL
871 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
872 uses C<undef> instead of NULL to provide this functionality.
876 The DB_HASH file format is probably the most commonly used of the three
877 file formats that B<DB_File> supports. It is also very straightforward
880 =head2 A Simple Example
882 This example shows how to create a database, add key/value pairs to the
883 database, delete keys/value pairs and finally how to enumerate the
884 contents of the database.
892 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
893 or die "Cannot open file 'fruit': $!\n";
895 # Add a few key/value pairs to the file
896 $h{"apple"} = "red" ;
897 $h{"orange"} = "orange" ;
898 $h{"banana"} = "yellow" ;
899 $h{"tomato"} = "red" ;
901 # Check for existence of a key
902 print "Banana Exists\n\n" if $h{"banana"} ;
904 # Delete a key/value pair.
907 # print the contents of the file
908 while (($k, $v) = each %h)
909 { print "$k -> $v\n" }
921 Note that the like ordinary associative arrays, the order of the keys
922 retrieved is in an apparently random order.
926 The DB_BTREE format is useful when you want to store data in a given
927 order. By default the keys will be stored in lexical order, but as you
928 will see from the example shown in the next section, it is very easy to
929 define your own sorting function.
931 =head2 Changing the BTREE sort order
933 This script shows how to override the default sorting algorithm that
934 BTREE uses. Instead of using the normal lexical ordering, a case
935 insensitive compare function will be used.
945 my ($key1, $key2) = @_ ;
946 "\L$key1" cmp "\L$key2" ;
949 # specify the Perl sub that will do the comparison
950 $DB_BTREE->{'compare'} = \&Compare ;
953 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
954 or die "Cannot open file 'tree': $!\n" ;
956 # Add a key/value pair to the file
957 $h{'Wall'} = 'Larry' ;
958 $h{'Smith'} = 'John' ;
959 $h{'mouse'} = 'mickey' ;
960 $h{'duck'} = 'donald' ;
965 # Cycle through the keys printing them in order.
966 # Note it is not necessary to sort the keys as
967 # the btree will have kept them in order automatically.
973 Here is the output from the code above.
979 There are a few point to bear in mind if you want to change the
980 ordering in a BTREE database:
986 The new compare function must be specified when you create the database.
990 You cannot change the ordering once the database has been created. Thus
991 you must use the same compare function every time you access the
996 Duplicate keys are entirely defined by the comparison function.
997 In the case-insensitive example above, the keys: 'KEY' and 'key'
998 would be considered duplicates, and assigning to the second one
999 would overwrite the first. If duplicates are allowed for (with the
1000 R_DUP flag discussed below), only a single copy of duplicate keys
1001 is stored in the database --- so (again with example above) assigning
1002 three values to the keys: 'KEY', 'Key', and 'key' would leave just
1003 the first key: 'KEY' in the database with three values. For some
1004 situations this results in information loss, so care should be taken
1005 to provide fully qualified comparison functions when necessary.
1006 For example, the above comparison routine could be modified to
1007 additionally compare case-sensitively if two keys are equal in the
1008 case insensitive comparison:
1011 my($key1, $key2) = @_;
1012 lc $key1 cmp lc $key2 ||
1016 And now you will only have duplicates when the keys themselves
1017 are truly the same. (note: in versions of the db library prior to
1018 about November 1996, such duplicate keys were retained so it was
1019 possible to recover the original keys in sets of keys that
1025 =head2 Handling Duplicate Keys
1027 The BTREE file type optionally allows a single key to be associated
1028 with an arbitrary number of values. This option is enabled by setting
1029 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
1031 There are some difficulties in using the tied hash interface if you
1032 want to manipulate a BTREE database with duplicate keys. Consider this
1039 my ($filename, %h) ;
1041 $filename = "tree" ;
1044 # Enable duplicate records
1045 $DB_BTREE->{'flags'} = R_DUP ;
1047 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1048 or die "Cannot open $filename: $!\n";
1050 # Add some key/value pairs to the file
1051 $h{'Wall'} = 'Larry' ;
1052 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1053 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1054 $h{'Smith'} = 'John' ;
1055 $h{'mouse'} = 'mickey' ;
1057 # iterate through the associative array
1058 # and print each key/value pair.
1059 foreach (sort keys %h)
1060 { print "$_ -> $h{$_}\n" }
1072 As you can see 3 records have been successfully created with key C<Wall>
1073 - the only thing is, when they are retrieved from the database they
1074 I<seem> to have the same value, namely C<Larry>. The problem is caused
1075 by the way that the associative array interface works. Basically, when
1076 the associative array interface is used to fetch the value associated
1077 with a given key, it will only ever retrieve the first value.
1079 Although it may not be immediately obvious from the code above, the
1080 associative array interface can be used to write values with duplicate
1081 keys, but it cannot be used to read them back from the database.
1083 The way to get around this problem is to use the Berkeley DB API method
1084 called C<seq>. This method allows sequential access to key/value
1085 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1086 and the API in general.
1088 Here is the script above rewritten using the C<seq> API method.
1094 my ($filename, $x, %h, $status, $key, $value) ;
1096 $filename = "tree" ;
1099 # Enable duplicate records
1100 $DB_BTREE->{'flags'} = R_DUP ;
1102 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1103 or die "Cannot open $filename: $!\n";
1105 # Add some key/value pairs to the file
1106 $h{'Wall'} = 'Larry' ;
1107 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1108 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1109 $h{'Smith'} = 'John' ;
1110 $h{'mouse'} = 'mickey' ;
1112 # iterate through the btree using seq
1113 # and print each key/value pair.
1115 for ($status = $x->seq($key, $value, R_FIRST) ;
1117 $status = $x->seq($key, $value, R_NEXT) )
1118 { print "$key -> $value\n" }
1131 This time we have got all the key/value pairs, including the multiple
1132 values associated with the key C<Wall>.
1134 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1135 a few utility methods.
1137 =head2 The get_dup() Method
1139 The C<get_dup> method assists in
1140 reading duplicate values from BTREE databases. The method can take the
1143 $count = $x->get_dup($key) ;
1144 @list = $x->get_dup($key) ;
1145 %list = $x->get_dup($key, 1) ;
1147 In a scalar context the method returns the number of values associated
1148 with the key, C<$key>.
1150 In list context, it returns all the values which match C<$key>. Note
1151 that the values will be returned in an apparently random order.
1153 In list context, if the second parameter is present and evaluates
1154 TRUE, the method returns an associative array. The keys of the
1155 associative array correspond to the values that matched in the BTREE
1156 and the values of the array are a count of the number of times that
1157 particular value occurred in the BTREE.
1159 So assuming the database created above, we can use C<get_dup> like
1166 my ($filename, $x, %h) ;
1168 $filename = "tree" ;
1170 # Enable duplicate records
1171 $DB_BTREE->{'flags'} = R_DUP ;
1173 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1174 or die "Cannot open $filename: $!\n";
1176 my $cnt = $x->get_dup("Wall") ;
1177 print "Wall occurred $cnt times\n" ;
1179 my %hash = $x->get_dup("Wall", 1) ;
1180 print "Larry is there\n" if $hash{'Larry'} ;
1181 print "There are $hash{'Brick'} Brick Walls\n" ;
1183 my @list = sort $x->get_dup("Wall") ;
1184 print "Wall => [@list]\n" ;
1186 @list = $x->get_dup("Smith") ;
1187 print "Smith => [@list]\n" ;
1189 @list = $x->get_dup("Dog") ;
1190 print "Dog => [@list]\n" ;
1195 Wall occurred 3 times
1197 There are 2 Brick Walls
1198 Wall => [Brick Brick Larry]
1202 =head2 The find_dup() Method
1204 $status = $X->find_dup($key, $value) ;
1206 This method checks for the existence of a specific key/value pair. If the
1207 pair exists, the cursor is left pointing to the pair and the method
1208 returns 0. Otherwise the method returns a non-zero value.
1210 Assuming the database from the previous example:
1216 my ($filename, $x, %h, $found) ;
1218 $filename = "tree" ;
1220 # Enable duplicate records
1221 $DB_BTREE->{'flags'} = R_DUP ;
1223 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1224 or die "Cannot open $filename: $!\n";
1226 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1227 print "Larry Wall is $found there\n" ;
1229 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1230 print "Harry Wall is $found there\n" ;
1238 Harry Wall is not there
1241 =head2 The del_dup() Method
1243 $status = $X->del_dup($key, $value) ;
1245 This method deletes a specific key/value pair. It returns
1246 0 if they exist and have been deleted successfully.
1247 Otherwise the method returns a non-zero value.
1249 Again assuming the existence of the C<tree> database
1255 my ($filename, $x, %h, $found) ;
1257 $filename = "tree" ;
1259 # Enable duplicate records
1260 $DB_BTREE->{'flags'} = R_DUP ;
1262 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1263 or die "Cannot open $filename: $!\n";
1265 $x->del_dup("Wall", "Larry") ;
1267 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1268 print "Larry Wall is $found there\n" ;
1275 Larry Wall is not there
1277 =head2 Matching Partial Keys
1279 The BTREE interface has a feature which allows partial keys to be
1280 matched. This functionality is I<only> available when the C<seq> method
1281 is used along with the R_CURSOR flag.
1283 $x->seq($key, $value, R_CURSOR) ;
1285 Here is the relevant quote from the dbopen man page where it defines
1286 the use of the R_CURSOR flag with seq:
1288 Note, for the DB_BTREE access method, the returned key is not
1289 necessarily an exact match for the specified key. The returned key
1290 is the smallest key greater than or equal to the specified key,
1291 permitting partial key matches and range searches.
1293 In the example script below, the C<match> sub uses this feature to find
1294 and print the first matching key/value pair given a partial key.
1301 my ($filename, $x, %h, $st, $key, $value) ;
1307 my $orig_key = $key ;
1308 $x->seq($key, $value, R_CURSOR) ;
1309 print "$orig_key\t-> $key\t-> $value\n" ;
1312 $filename = "tree" ;
1315 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1316 or die "Cannot open $filename: $!\n";
1318 # Add some key/value pairs to the file
1319 $h{'mouse'} = 'mickey' ;
1320 $h{'Wall'} = 'Larry' ;
1321 $h{'Walls'} = 'Brick' ;
1322 $h{'Smith'} = 'John' ;
1326 print "IN ORDER\n" ;
1327 for ($st = $x->seq($key, $value, R_FIRST) ;
1329 $st = $x->seq($key, $value, R_NEXT) )
1331 { print "$key -> $value\n" }
1333 print "\nPARTIAL MATCH\n" ;
1353 a -> mouse -> mickey
1357 DB_RECNO provides an interface to flat text files. Both variable and
1358 fixed length records are supported.
1360 In order to make RECNO more compatible with Perl, the array offset for
1361 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1363 As with normal Perl arrays, a RECNO array can be accessed using
1364 negative indexes. The index -1 refers to the last element of the array,
1365 -2 the second last, and so on. Attempting to access an element before
1366 the start of the array will raise a fatal run-time error.
1368 =head2 The 'bval' Option
1370 The operation of the bval option warrants some discussion. Here is the
1371 definition of bval from the Berkeley DB 1.85 recno manual page:
1373 The delimiting byte to be used to mark the end of a
1374 record for variable-length records, and the pad charac-
1375 ter for fixed-length records. If no value is speci-
1376 fied, newlines (``\n'') are used to mark the end of
1377 variable-length records and fixed-length records are
1380 The second sentence is wrong. In actual fact bval will only default to
1381 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1382 openinfo parameter is used at all, the value that happens to be in bval
1383 will be used. That means you always have to specify bval when making
1384 use of any of the options in the openinfo parameter. This documentation
1385 error will be fixed in the next release of Berkeley DB.
1387 That clarifies the situation with regards Berkeley DB itself. What
1388 about B<DB_File>? Well, the behavior defined in the quote above is
1389 quite useful, so B<DB_File> conforms to it.
1391 That means that you can specify other options (e.g. cachesize) and
1392 still have bval default to C<"\n"> for variable length records, and
1393 space for fixed length records.
1395 Also note that the bval option only allows you to specify a single byte
1398 =head2 A Simple Example
1400 Here is a simple example that uses RECNO (if you are using a version
1401 of Perl earlier than 5.004_57 this example won't work -- see
1402 L<Extra RECNO Methods> for a workaround).
1408 my $filename = "text" ;
1412 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1413 or die "Cannot open file 'text': $!\n" ;
1415 # Add a few key/value pairs to the file
1420 push @h, "green", "black" ;
1422 my $elements = scalar @h ;
1423 print "The array contains $elements entries\n" ;
1426 print "popped $last\n" ;
1428 unshift @h, "white" ;
1429 my $first = shift @h ;
1430 print "shifted $first\n" ;
1432 # Check for existence of a key
1433 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1435 # use a negative index
1436 print "The last element is $h[-1]\n" ;
1437 print "The 2nd last element is $h[-2]\n" ;
1441 Here is the output from the script:
1443 The array contains 5 entries
1446 Element 1 Exists with value blue
1447 The last element is green
1448 The 2nd last element is yellow
1450 =head2 Extra RECNO Methods
1452 If you are using a version of Perl earlier than 5.004_57, the tied
1453 array interface is quite limited. In the example script above
1454 C<push>, C<pop>, C<shift>, C<unshift>
1455 or determining the array length will not work with a tied array.
1457 To make the interface more useful for older versions of Perl, a number
1458 of methods are supplied with B<DB_File> to simulate the missing array
1459 operations. All these methods are accessed via the object returned from
1462 Here are the methods:
1466 =item B<$X-E<gt>push(list) ;>
1468 Pushes the elements of C<list> to the end of the array.
1470 =item B<$value = $X-E<gt>pop ;>
1472 Removes and returns the last element of the array.
1474 =item B<$X-E<gt>shift>
1476 Removes and returns the first element of the array.
1478 =item B<$X-E<gt>unshift(list) ;>
1480 Pushes the elements of C<list> to the start of the array.
1482 =item B<$X-E<gt>length>
1484 Returns the number of elements in the array.
1486 =item B<$X-E<gt>splice(offset, length, elements);>
1488 Returns a splice of the array.
1492 =head2 Another Example
1494 Here is a more complete example that makes use of some of the methods
1495 described above. It also makes use of the API interface directly (see
1496 L<THE API INTERFACE>).
1500 my (@h, $H, $file, $i) ;
1508 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1509 or die "Cannot open file $file: $!\n" ;
1511 # first create a text file to play with
1519 # Print the records in order.
1521 # The length method is needed here because evaluating a tied
1522 # array in a scalar context does not return the number of
1523 # elements in the array.
1525 print "\nORIGINAL\n" ;
1526 foreach $i (0 .. $H->length - 1) {
1527 print "$i: $h[$i]\n" ;
1530 # use the push & pop methods
1533 print "\nThe last record was [$a]\n" ;
1535 # and the shift & unshift methods
1537 $H->unshift("first") ;
1538 print "The first record was [$a]\n" ;
1540 # Use the API to add a new record after record 2.
1542 $H->put($i, "Newbie", R_IAFTER) ;
1544 # and a new record before record 1.
1546 $H->put($i, "New One", R_IBEFORE) ;
1551 # now print the records in reverse order
1552 print "\nREVERSE\n" ;
1553 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1554 { print "$i: $h[$i]\n" }
1556 # same again, but use the API functions instead
1557 print "\nREVERSE again\n" ;
1558 my ($s, $k, $v) = (0, 0, 0) ;
1559 for ($s = $H->seq($k, $v, R_LAST) ;
1561 $s = $H->seq($k, $v, R_PREV))
1562 { print "$k: $v\n" }
1567 and this is what it outputs:
1576 The last record was [four]
1577 The first record was [zero]
1601 Rather than iterating through the array, C<@h> like this:
1605 it is necessary to use either this:
1607 foreach $i (0 .. $H->length - 1)
1611 for ($a = $H->get($k, $v, R_FIRST) ;
1613 $a = $H->get($k, $v, R_NEXT) )
1617 Notice that both times the C<put> method was used the record index was
1618 specified using a variable, C<$i>, rather than the literal value
1619 itself. This is because C<put> will return the record number of the
1620 inserted line via that parameter.
1624 =head1 THE API INTERFACE
1626 As well as accessing Berkeley DB using a tied hash or array, it is also
1627 possible to make direct use of most of the API functions defined in the
1628 Berkeley DB documentation.
1630 To do this you need to store a copy of the object returned from the tie.
1632 $db = tie %hash, "DB_File", "filename" ;
1634 Once you have done that, you can access the Berkeley DB API functions
1635 as B<DB_File> methods directly like this:
1637 $db->put($key, $value, R_NOOVERWRITE) ;
1639 B<Important:> If you have saved a copy of the object returned from
1640 C<tie>, the underlying database file will I<not> be closed until both
1641 the tied variable is untied and all copies of the saved object are
1645 $db = tie %hash, "DB_File", "filename"
1646 or die "Cannot tie filename: $!" ;
1651 See L<The untie() Gotcha> for more details.
1653 All the functions defined in L<dbopen> are available except for
1654 close() and dbopen() itself. The B<DB_File> method interface to the
1655 supported functions have been implemented to mirror the way Berkeley DB
1656 works whenever possible. In particular note that:
1662 The methods return a status value. All return 0 on success.
1663 All return -1 to signify an error and set C<$!> to the exact
1664 error code. The return code 1 generally (but not always) means that the
1665 key specified did not exist in the database.
1667 Other return codes are defined. See below and in the Berkeley DB
1668 documentation for details. The Berkeley DB documentation should be used
1669 as the definitive source.
1673 Whenever a Berkeley DB function returns data via one of its parameters,
1674 the equivalent B<DB_File> method does exactly the same.
1678 If you are careful, it is possible to mix API calls with the tied
1679 hash/array interface in the same piece of code. Although only a few of
1680 the methods used to implement the tied interface currently make use of
1681 the cursor, you should always assume that the cursor has been changed
1682 any time the tied hash/array interface is used. As an example, this
1683 code will probably not do what you expect:
1685 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1686 or die "Cannot tie $filename: $!" ;
1688 # Get the first key/value pair and set the cursor
1689 $X->seq($key, $value, R_FIRST) ;
1691 # this line will modify the cursor
1692 $count = scalar keys %x ;
1694 # Get the second key/value pair.
1695 # oops, it didn't, it got the last key/value pair!
1696 $X->seq($key, $value, R_NEXT) ;
1698 The code above can be rearranged to get around the problem, like this:
1700 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1701 or die "Cannot tie $filename: $!" ;
1703 # this line will modify the cursor
1704 $count = scalar keys %x ;
1706 # Get the first key/value pair and set the cursor
1707 $X->seq($key, $value, R_FIRST) ;
1709 # Get the second key/value pair.
1711 $X->seq($key, $value, R_NEXT) ;
1715 All the constants defined in L<dbopen> for use in the flags parameters
1716 in the methods defined below are also available. Refer to the Berkeley
1717 DB documentation for the precise meaning of the flags values.
1719 Below is a list of the methods available.
1723 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1725 Given a key (C<$key>) this method reads the value associated with it
1726 from the database. The value read from the database is returned in the
1727 C<$value> parameter.
1729 If the key does not exist the method returns 1.
1731 No flags are currently defined for this method.
1733 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1735 Stores the key/value pair in the database.
1737 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1738 will have the record number of the inserted key/value pair set.
1740 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1743 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1745 Removes all key/value pairs with key C<$key> from the database.
1747 A return code of 1 means that the requested key was not in the
1750 R_CURSOR is the only valid flag at present.
1752 =item B<$status = $X-E<gt>fd ;>
1754 Returns the file descriptor for the underlying database.
1756 See L<Locking: The Trouble with fd> for an explanation for why you should
1757 not use C<fd> to lock your database.
1759 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1761 This interface allows sequential retrieval from the database. See
1762 L<dbopen> for full details.
1764 Both the C<$key> and C<$value> parameters will be set to the key/value
1765 pair read from the database.
1767 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1768 R_FIRST, R_LAST, R_NEXT and R_PREV.
1770 =item B<$status = $X-E<gt>sync([$flags]) ;>
1772 Flushes any cached buffers to disk.
1774 R_RECNOSYNC is the only valid flag at present.
1780 A DBM Filter is a piece of code that is be used when you I<always>
1781 want to make the same transformation to all keys and/or values in a
1784 There are four methods associated with DBM Filters. All work identically,
1785 and each is used to install (or uninstall) a single DBM Filter. Each
1786 expects a single parameter, namely a reference to a sub. The only
1787 difference between them is the place that the filter is installed.
1793 =item B<filter_store_key>
1795 If a filter has been installed with this method, it will be invoked
1796 every time you write a key to a DBM database.
1798 =item B<filter_store_value>
1800 If a filter has been installed with this method, it will be invoked
1801 every time you write a value to a DBM database.
1804 =item B<filter_fetch_key>
1806 If a filter has been installed with this method, it will be invoked
1807 every time you read a key from a DBM database.
1809 =item B<filter_fetch_value>
1811 If a filter has been installed with this method, it will be invoked
1812 every time you read a value from a DBM database.
1816 You can use any combination of the methods, from none, to all four.
1818 All filter methods return the existing filter, if present, or C<undef>
1821 To delete a filter pass C<undef> to it.
1825 When each filter is called by Perl, a local copy of C<$_> will contain
1826 the key or value to be filtered. Filtering is achieved by modifying
1827 the contents of C<$_>. The return code from the filter is ignored.
1829 =head2 An Example -- the NULL termination problem.
1831 Consider the following scenario. You have a DBM database
1832 that you need to share with a third-party C application. The C application
1833 assumes that I<all> keys and values are NULL terminated. Unfortunately
1834 when Perl writes to DBM databases it doesn't use NULL termination, so
1835 your Perl application will have to manage NULL termination itself. When
1836 you write to the database you will have to use something like this:
1838 $hash{"$key\0"} = "$value\0" ;
1840 Similarly the NULL needs to be taken into account when you are considering
1841 the length of existing keys/values.
1843 It would be much better if you could ignore the NULL terminations issue
1844 in the main application code and have a mechanism that automatically
1845 added the terminating NULL to all keys and values whenever you write to
1846 the database and have them removed when you read from the database. As I'm
1847 sure you have already guessed, this is a problem that DBM Filters can
1855 my $filename = "filt" ;
1858 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1859 or die "Cannot open $filename: $!\n" ;
1861 # Install DBM Filters
1862 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1863 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1864 $db->filter_fetch_value( sub { s/\0$// } ) ;
1865 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1867 $hash{"abc"} = "def" ;
1868 my $a = $hash{"ABC"} ;
1873 Hopefully the contents of each of the filters should be
1874 self-explanatory. Both "fetch" filters remove the terminating NULL,
1875 and both "store" filters add a terminating NULL.
1878 =head2 Another Example -- Key is a C int.
1880 Here is another real-life example. By default, whenever Perl writes to
1881 a DBM database it always writes the key and value as strings. So when
1884 $hash{12345} = "something" ;
1886 the key 12345 will get stored in the DBM database as the 5 byte string
1887 "12345". If you actually want the key to be stored in the DBM database
1888 as a C int, you will have to use C<pack> when writing, and C<unpack>
1891 Here is a DBM Filter that does it:
1897 my $filename = "filt" ;
1901 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1902 or die "Cannot open $filename: $!\n" ;
1904 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1905 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1906 $hash{123} = "def" ;
1911 This time only two filters have been used -- we only need to manipulate
1912 the contents of the key, so it wasn't necessary to install any value
1915 =head1 HINTS AND TIPS
1918 =head2 Locking: The Trouble with fd
1920 Until version 1.72 of this module, the recommended technique for locking
1921 B<DB_File> databases was to flock the filehandle returned from the "fd"
1922 function. Unfortunately this technique has been shown to be fundamentally
1923 flawed (Kudos to David Harris for tracking this down). Use it at your own
1926 The locking technique went like this.
1928 $db = tie(%db, 'DB_File', 'foo.db', O_CREAT|O_RDWR, 0644)
1929 || die "dbcreat foo.db $!";
1931 open(DB_FH, "+<&=$fd") || die "dup $!";
1932 flock (DB_FH, LOCK_EX) || die "flock: $!";
1934 $db{"Tom"} = "Jerry" ;
1936 flock(DB_FH, LOCK_UN);
1941 In simple terms, this is what happens:
1947 Use "tie" to open the database.
1951 Lock the database with fd & flock.
1955 Read & Write to the database.
1959 Unlock and close the database.
1963 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1964 database in step 2 is that an initial block from the database will get
1965 read from disk and cached in memory.
1967 To see why this is a problem, consider what can happen when two processes,
1968 say "A" and "B", both want to update the same B<DB_File> database
1969 using the locking steps outlined above. Assume process "A" has already
1970 opened the database and has a write lock, but it hasn't actually updated
1971 the database yet (it has finished step 2, but not started step 3 yet). Now
1972 process "B" tries to open the same database - step 1 will succeed,
1973 but it will block on step 2 until process "A" releases the lock. The
1974 important thing to notice here is that at this point in time both
1975 processes will have cached identical initial blocks from the database.
1977 Now process "A" updates the database and happens to change some of the
1978 data held in the initial buffer. Process "A" terminates, flushing
1979 all cached data to disk and releasing the database lock. At this point
1980 the database on disk will correctly reflect the changes made by process
1983 With the lock released, process "B" can now continue. It also updates the
1984 database and unfortunately it too modifies the data that was in its
1985 initial buffer. Once that data gets flushed to disk it will overwrite
1986 some/all of the changes process "A" made to the database.
1988 The result of this scenario is at best a database that doesn't contain
1989 what you expect. At worst the database will corrupt.
1991 The above won't happen every time competing process update the same
1992 B<DB_File> database, but it does illustrate why the technique should
1995 =head2 Safe ways to lock a database
1997 Starting with version 2.x, Berkeley DB has internal support for locking.
1998 The companion module to this one, B<BerkeleyDB>, provides an interface
1999 to this locking functionality. If you are serious about locking
2000 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
2002 If using B<BerkeleyDB> isn't an option, there are a number of modules
2003 available on CPAN that can be used to implement locking. Each one
2004 implements locking differently and has different goals in mind. It is
2005 therefore worth knowing the difference, so that you can pick the right
2006 one for your application. Here are the three locking wrappers:
2010 =item B<Tie::DB_Lock>
2012 A B<DB_File> wrapper which creates copies of the database file for
2013 read access, so that you have a kind of a multiversioning concurrent read
2014 system. However, updates are still serial. Use for databases where reads
2015 may be lengthy and consistency problems may occur.
2017 =item B<Tie::DB_LockFile>
2019 A B<DB_File> wrapper that has the ability to lock and unlock the database
2020 while it is being used. Avoids the tie-before-flock problem by simply
2021 re-tie-ing the database when you get or drop a lock. Because of the
2022 flexibility in dropping and re-acquiring the lock in the middle of a
2023 session, this can be massaged into a system that will work with long
2024 updates and/or reads if the application follows the hints in the POD
2027 =item B<DB_File::Lock>
2029 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
2030 before tie-ing the database and drops the lock after the untie. Allows
2031 one to use the same lockfile for multiple databases to avoid deadlock
2032 problems, if desired. Use for databases where updates are reads are
2033 quick and simple flock locking semantics are enough.
2037 =head2 Sharing Databases With C Applications
2039 There is no technical reason why a Berkeley DB database cannot be
2040 shared by both a Perl and a C application.
2042 The vast majority of problems that are reported in this area boil down
2043 to the fact that C strings are NULL terminated, whilst Perl strings are
2044 not. See L<DBM FILTERS> for a generic way to work around this problem.
2046 Here is a real example. Netscape 2.0 keeps a record of the locations you
2047 visit along with the time you last visited them in a DB_HASH database.
2048 This is usually stored in the file F<~/.netscape/history.db>. The key
2049 field in the database is the location string and the value field is the
2050 time the location was last visited stored as a 4 byte binary value.
2052 If you haven't already guessed, the location string is stored with a
2053 terminating NULL. This means you need to be careful when accessing the
2056 Here is a snippet of code that is loosely based on Tom Christiansen's
2057 I<ggh> script (available from your nearest CPAN archive in
2058 F<authors/id/TOMC/scripts/nshist.gz>).
2065 my ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2066 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2068 $HISTORY = "$dotdir/.netscape/history.db";
2070 tie %hist_db, 'DB_File', $HISTORY
2071 or die "Cannot open $HISTORY: $!\n" ;;
2073 # Dump the complete database
2074 while ( ($href, $binary_time) = each %hist_db ) {
2076 # remove the terminating NULL
2077 $href =~ s/\x00$// ;
2079 # convert the binary time into a user friendly string
2080 $date = localtime unpack("V", $binary_time);
2081 print "$date $href\n" ;
2084 # check for the existence of a specific key
2085 # remember to add the NULL
2086 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2087 $date = localtime unpack("V", $binary_time) ;
2088 print "Last visited mox.perl.com on $date\n" ;
2091 print "Never visited mox.perl.com\n"
2096 =head2 The untie() Gotcha
2098 If you make use of the Berkeley DB API, it is I<very> strongly
2099 recommended that you read L<perltie/The untie Gotcha>.
2101 Even if you don't currently make use of the API interface, it is still
2104 Here is an example which illustrates the problem from a B<DB_File>
2113 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2114 or die "Cannot tie first time: $!" ;
2120 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2121 or die "Cannot tie second time: $!" ;
2125 When run, the script will produce this error message:
2127 Cannot tie second time: Invalid argument at bad.file line 14.
2129 Although the error message above refers to the second tie() statement
2130 in the script, the source of the problem is really with the untie()
2131 statement that precedes it.
2133 Having read L<perltie> you will probably have already guessed that the
2134 error is caused by the extra copy of the tied object stored in C<$X>.
2135 If you haven't, then the problem boils down to the fact that the
2136 B<DB_File> destructor, DESTROY, will not be called until I<all>
2137 references to the tied object are destroyed. Both the tied variable,
2138 C<%x>, and C<$X> above hold a reference to the object. The call to
2139 untie() will destroy the first, but C<$X> still holds a valid
2140 reference, so the destructor will not get called and the database file
2141 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2142 attempt to open a database that is already open via the catch-all
2143 "Invalid argument" doesn't help.
2145 If you run the script with the C<-w> flag the error message becomes:
2147 untie attempted while 1 inner references still exist at bad.file line 12.
2148 Cannot tie second time: Invalid argument at bad.file line 14.
2150 which pinpoints the real problem. Finally the script can now be
2151 modified to fix the original problem by destroying the API object
2160 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2164 =head1 COMMON QUESTIONS
2166 =head2 Why is there Perl source in my database?
2168 If you look at the contents of a database file created by DB_File,
2169 there can sometimes be part of a Perl script included in it.
2171 This happens because Berkeley DB uses dynamic memory to allocate
2172 buffers which will subsequently be written to the database file. Being
2173 dynamic, the memory could have been used for anything before DB
2174 malloced it. As Berkeley DB doesn't clear the memory once it has been
2175 allocated, the unused portions will contain random junk. In the case
2176 where a Perl script gets written to the database, the random junk will
2177 correspond to an area of dynamic memory that happened to be used during
2178 the compilation of the script.
2180 Unless you don't like the possibility of there being part of your Perl
2181 scripts embedded in a database file, this is nothing to worry about.
2183 =head2 How do I store complex data structures with DB_File?
2185 Although B<DB_File> cannot do this directly, there is a module which
2186 can layer transparently over B<DB_File> to accomplish this feat.
2188 Check out the MLDBM module, available on CPAN in the directory
2189 F<modules/by-module/MLDBM>.
2191 =head2 What does "Invalid Argument" mean?
2193 You will get this error message when one of the parameters in the
2194 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2195 get wrong, so it can be difficult to figure out which one it is.
2197 Here are a couple of possibilities:
2203 Attempting to reopen a database without closing it.
2207 Using the O_WRONLY flag.
2211 =head2 What does "Bareword 'DB_File' not allowed" mean?
2213 You will encounter this particular error message when you have the
2214 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2215 Consider this script:
2221 tie %x, DB_File, "filename" ;
2223 Running it produces the error in question:
2225 Bareword "DB_File" not allowed while "strict subs" in use
2227 To get around the error, place the word C<DB_File> in either single or
2228 double quotes, like this:
2230 tie %x, "DB_File", "filename" ;
2232 Although it might seem like a real pain, it is really worth the effort
2233 of having a C<use strict> in all your scripts.
2237 Articles that are either about B<DB_File> or make use of it.
2243 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2244 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2250 Moved to the Changes file.
2254 Some older versions of Berkeley DB had problems with fixed length
2255 records using the RECNO file format. This problem has been fixed since
2256 version 1.85 of Berkeley DB.
2258 I am sure there are bugs in the code. If you do find any, or can
2259 suggest any enhancements, I would welcome your comments.
2263 B<DB_File> comes with the standard Perl source distribution. Look in
2264 the directory F<ext/DB_File>. Given the amount of time between releases
2265 of Perl the version that ships with Perl is quite likely to be out of
2266 date, so the most recent version can always be found on CPAN (see
2267 L<perlmodlib/CPAN> for details), in the directory
2268 F<modules/by-module/DB_File>.
2270 This version of B<DB_File> will work with either version 1.x, 2.x or
2271 3.x of Berkeley DB, but is limited to the functionality provided by
2274 The official web site for Berkeley DB is F<http://www.oracle.com/technology/products/berkeley-db/db/index.html>.
2275 All versions of Berkeley DB are available there.
2277 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2278 archive in F<src/misc/db.1.85.tar.gz>.
2282 Copyright (c) 1995-2012 Paul Marquess. All rights reserved. This program
2283 is free software; you can redistribute it and/or modify it under the
2284 same terms as Perl itself.
2286 Although B<DB_File> is covered by the Perl license, the library it
2287 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2288 copyright and its own license. Please take the time to read it.
2290 Here are are few words taken from the Berkeley DB FAQ (at
2291 F<http://www.oracle.com/technology/products/berkeley-db/db/index.html>) regarding the license:
2293 Do I have to license DB to use it in Perl scripts?
2295 No. The Berkeley DB license requires that software that uses
2296 Berkeley DB be freely redistributable. In the case of Perl, that
2297 software is Perl, and not your scripts. Any Perl scripts that you
2298 write are your property, including scripts that make use of
2299 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2300 place any restriction on what you may do with them.
2302 If you are in any doubt about the license situation, contact either the
2303 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2308 L<perl>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2313 The DB_File interface was written by Paul Marquess
2314 E<lt>pmqs@cpan.orgE<gt>.