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, $Error);
169 $VERSION = eval $VERSION; # needed for dev releases
172 local $SIG{__WARN__} = sub {$splice_end_array = "@_";};
173 my @a =(1); splice(@a, 3);
175 ($splice_end_array =~ /^splice\(\) offset past end of array at /);
178 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
179 $DB_BTREE = new DB_File::BTREEINFO ;
180 $DB_HASH = new DB_File::HASHINFO ;
181 $DB_RECNO = new DB_File::RECNOINFO ;
188 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
193 @ISA = qw(DynaLoader);
197 push @ISA, qw(Tie::Hash Exporter);
199 $DB_BTREE $DB_HASH $DB_RECNO
234 ($constname = $AUTOLOAD) =~ s/.*:://;
235 my ($error, $val) = constant($constname);
236 Carp::croak $error if $error;
238 *{$AUTOLOAD} = sub { $val };
244 # Make all Fcntl O_XXX constants available for importing
246 my @O = grep /^O_/, @Fcntl::EXPORT;
247 Fcntl->import(@O); # first we import what we want to export
252 { XSLoader::load("DB_File", $VERSION)}
254 { bootstrap DB_File $VERSION }
256 # Preloaded methods go here. Autoload methods go after __END__, and are
257 # processed by the autosplit program.
259 sub tie_hash_or_array
262 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
265 $arg[1] = File::Spec->rel2abs($arg[1])
268 $arg[4] = tied %{ $arg[4] }
269 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
271 $arg[2] = O_CREAT()|O_RDWR() if @arg >=3 && ! defined $arg[2];
272 $arg[3] = 0666 if @arg >=4 && ! defined $arg[3];
274 # make recno in Berkeley DB version 2 (or better) work like
275 # recno in version 1.
276 if ($db_version >= 4 and ! $tieHASH) {
277 $arg[2] |= O_CREAT();
280 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
281 $arg[1] and ! -e $arg[1]) {
282 open(FH, ">$arg[1]") or return undef ;
284 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
287 DoTie_($tieHASH, @arg) ;
292 tie_hash_or_array(@_) ;
297 tie_hash_or_array(@_) ;
305 my $status = $self->seq($key, $value, R_FIRST());
308 while ($status == 0) {
310 $status = $self->seq($key, $value, R_NEXT());
312 foreach $key (reverse @keys) {
313 my $s = $self->del($key);
323 my $current_length = $self->length() ;
325 if ($length < $current_length) {
327 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
330 elsif ($length > $current_length) {
331 $self->put($length-1, "") ;
340 if (not defined $offset) {
341 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
345 my $length = @_ ? shift : 0;
346 # Carping about definedness comes _after_ the OFFSET sanity check.
347 # This is so we get the same error messages as Perl's splice().
352 my $size = $self->FETCHSIZE();
354 # 'If OFFSET is negative then it start that far from the end of
358 my $new_offset = $size + $offset;
359 if ($new_offset < 0) {
360 die "Modification of non-creatable array value attempted, "
361 . "subscript $offset";
363 $offset = $new_offset;
366 if (not defined $length) {
367 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
371 if ($offset > $size) {
373 warnings::warnif('misc', 'splice() offset past end of array')
374 if $splice_end_array;
377 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
378 if (not defined $length) {
379 $length = $size - $offset;
382 # 'If LENGTH is negative, leave that many elements off the end of
386 $length = $size - $offset + $length;
389 # The user must have specified a length bigger than the
390 # length of the array passed in. But perl's splice()
391 # doesn't catch this, it just behaves as for length=0.
397 if ($length > $size - $offset) {
398 $length = $size - $offset;
401 # $num_elems holds the current number of elements in the database.
402 my $num_elems = $size;
404 # 'Removes the elements designated by OFFSET and LENGTH from an
408 foreach (0 .. $length - 1) {
410 my $status = $self->get($offset, $old);
412 my $msg = "error from Berkeley DB on get($offset, \$old)";
414 $msg .= ' (no such element?)';
417 $msg .= ": error status $status";
418 if (defined $! and $! ne '') {
419 $msg .= ", message $!";
426 $status = $self->del($offset);
428 my $msg = "error from Berkeley DB on del($offset)";
430 $msg .= ' (no such element?)';
433 $msg .= ": error status $status";
434 if (defined $! and $! ne '') {
435 $msg .= ", message $!";
444 # ...'and replaces them with the elements of LIST, if any.'
446 while (defined (my $elem = shift @list)) {
449 if ($pos >= $num_elems) {
450 $status = $self->put($pos, $elem);
453 $status = $self->put($pos, $elem, $self->R_IBEFORE);
457 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
459 $msg .= ' (no such element?)';
462 $msg .= ", error status $status";
463 if (defined $! and $! ne '') {
464 $msg .= ", message $!";
470 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
478 # 'In list context, returns the elements removed from the
483 elsif (defined wantarray and not wantarray) {
484 # 'In scalar context, returns the last element removed, or
485 # undef if no elements are removed.'
488 my $last = pop @removed;
495 elsif (not defined wantarray) {
500 sub ::DB_File::splice { &SPLICE }
504 croak "Usage: \$db->find_dup(key,value)\n"
508 my ($origkey, $value_wanted) = @_ ;
509 my ($key, $value) = ($origkey, 0);
512 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
514 $status = $db->seq($key, $value, R_NEXT() ) ) {
516 return 0 if $key eq $origkey and $value eq $value_wanted ;
524 croak "Usage: \$db->del_dup(key,value)\n"
528 my ($key, $value) = @_ ;
529 my ($status) = $db->find_dup($key, $value) ;
530 return $status if $status != 0 ;
532 $status = $db->del($key, R_CURSOR() ) ;
538 croak "Usage: \$db->get_dup(key [,flag])\n"
539 unless @_ == 2 or @_ == 3 ;
546 my $wantarray = wantarray ;
552 # iterate through the database until either EOF ($status == 0)
553 # or a different key is encountered ($key ne $origkey).
554 for ($status = $db->seq($key, $value, R_CURSOR()) ;
555 $status == 0 and $key eq $origkey ;
556 $status = $db->seq($key, $value, R_NEXT()) ) {
558 # save the value or count number of matches
561 { ++ $values{$value} }
563 { push (@values, $value) }
570 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
579 DB_File - Perl5 access to Berkeley DB version 1.x
585 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
586 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
587 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
589 $status = $X->del($key [, $flags]) ;
590 $status = $X->put($key, $value [, $flags]) ;
591 $status = $X->get($key, $value [, $flags]) ;
592 $status = $X->seq($key, $value, $flags) ;
593 $status = $X->sync([$flags]) ;
597 $count = $X->get_dup($key) ;
598 @list = $X->get_dup($key) ;
599 %list = $X->get_dup($key, 1) ;
600 $status = $X->find_dup($key, $value) ;
601 $status = $X->del_dup($key, $value) ;
609 @r = $X->splice(offset, length, elements);
612 $old_filter = $db->filter_store_key ( sub { ... } ) ;
613 $old_filter = $db->filter_store_value( sub { ... } ) ;
614 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
615 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
622 B<DB_File> is a module which allows Perl programs to make use of the
623 facilities provided by Berkeley DB version 1.x (if you have a newer
624 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
625 It is assumed that you have a copy of the Berkeley DB manual pages at
626 hand when reading this documentation. The interface defined here
627 mirrors the Berkeley DB interface closely.
629 Berkeley DB is a C library which provides a consistent interface to a
630 number of database formats. B<DB_File> provides an interface to all
631 three of the database types currently supported by Berkeley DB.
639 This database type allows arbitrary key/value pairs to be stored in data
640 files. This is equivalent to the functionality provided by other
641 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
642 the files created using DB_HASH are not compatible with any of the
643 other packages mentioned.
645 A default hashing algorithm, which will be adequate for most
646 applications, is built into Berkeley DB. If you do need to use your own
647 hashing algorithm it is possible to write your own in Perl and have
648 B<DB_File> use it instead.
652 The btree format allows arbitrary key/value pairs to be stored in a
653 sorted, balanced binary tree.
655 As with the DB_HASH format, it is possible to provide a user defined
656 Perl routine to perform the comparison of keys. By default, though, the
657 keys are stored in lexical order.
661 DB_RECNO allows both fixed-length and variable-length flat text files
662 to be manipulated using the same key/value pair interface as in DB_HASH
663 and DB_BTREE. In this case the key will consist of a record (line)
668 =head2 Using DB_File with Berkeley DB version 2 or greater
670 Although B<DB_File> is intended to be used with Berkeley DB version 1,
671 it can also be used with version 2, 3 or 4. In this case the interface is
672 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
673 version 2 or greater interface differs, B<DB_File> arranges for it to work
674 like version 1. This feature allows B<DB_File> scripts that were built
675 with version 1 to be migrated to version 2 or greater without any changes.
677 If you want to make use of the new features available in Berkeley DB
678 2.x or greater, use the Perl module B<BerkeleyDB> instead.
680 B<Note:> The database file format has changed multiple times in Berkeley
681 DB version 2, 3 and 4. If you cannot recreate your databases, you
682 must dump any existing databases with either the C<db_dump> or the
683 C<db_dump185> utility that comes with Berkeley DB.
684 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
685 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
686 documentation for further details.
688 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
691 =head2 Interface to Berkeley DB
693 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
694 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
695 allows B<DB_File> to access Berkeley DB files using either an
696 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
697 array (for the DB_RECNO file type).
699 In addition to the tie() interface, it is also possible to access most
700 of the functions provided in the Berkeley DB API directly.
701 See L<THE API INTERFACE>.
703 =head2 Opening a Berkeley DB Database File
705 Berkeley DB uses the function dbopen() to open or create a database.
706 Here is the C prototype for dbopen():
709 dbopen (const char * file, int flags, int mode,
710 DBTYPE type, const void * openinfo)
712 The parameter C<type> is an enumeration which specifies which of the 3
713 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
714 Depending on which of these is actually chosen, the final parameter,
715 I<openinfo> points to a data structure which allows tailoring of the
716 specific interface method.
718 This interface is handled slightly differently in B<DB_File>. Here is
719 an equivalent call using B<DB_File>:
721 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
723 The C<filename>, C<flags> and C<mode> parameters are the direct
724 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
725 performs the function of both the C<type> and C<openinfo> parameters in
728 In the example above $DB_HASH is actually a pre-defined reference to a
729 hash object. B<DB_File> has three of these pre-defined references.
730 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
732 The keys allowed in each of these pre-defined references is limited to
733 the names used in the equivalent C structure. So, for example, the
734 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
735 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
737 To change one of these elements, just assign to it like this:
739 $DB_HASH->{'cachesize'} = 10000 ;
741 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
742 usually adequate for most applications. If you do need to create extra
743 instances of these objects, constructors are available for each file
746 Here are examples of the constructors and the valid options available
747 for DB_HASH, DB_BTREE and DB_RECNO respectively.
749 $a = new DB_File::HASHINFO ;
757 $b = new DB_File::BTREEINFO ;
767 $c = new DB_File::RECNOINFO ;
776 The values stored in the hashes above are mostly the direct equivalent
777 of their C counterpart. Like their C counterparts, all are set to a
778 default values - that means you don't have to set I<all> of the
779 values when you only want to change one. Here is an example:
781 $a = new DB_File::HASHINFO ;
782 $a->{'cachesize'} = 12345 ;
783 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
785 A few of the options need extra discussion here. When used, the C
786 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
787 to C functions. In B<DB_File> these keys are used to store references
788 to Perl subs. Below are templates for each of the subs:
794 # return the hash value for $data
800 my ($key, $key2) = @_ ;
802 # return 0 if $key1 eq $key2
803 # -1 if $key1 lt $key2
804 # 1 if $key1 gt $key2
805 return (-1 , 0 or 1) ;
810 my ($key, $key2) = @_ ;
812 # return number of bytes of $key2 which are
813 # necessary to determine that it is greater than $key1
817 See L<Changing the BTREE sort order> for an example of using the
820 If you are using the DB_RECNO interface and you intend making use of
821 C<bval>, you should check out L<The 'bval' Option>.
823 =head2 Default Parameters
825 It is possible to omit some or all of the final 4 parameters in the
826 call to C<tie> and let them take default values. As DB_HASH is the most
827 common file format used, the call:
829 tie %A, "DB_File", "filename" ;
833 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
835 It is also possible to omit the filename parameter as well, so the
842 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
844 See L<In Memory Databases> for a discussion on the use of C<undef>
845 in place of a filename.
847 =head2 In Memory Databases
849 Berkeley DB allows the creation of in-memory databases by using NULL
850 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
851 uses C<undef> instead of NULL to provide this functionality.
855 The DB_HASH file format is probably the most commonly used of the three
856 file formats that B<DB_File> supports. It is also very straightforward
859 =head2 A Simple Example
861 This example shows how to create a database, add key/value pairs to the
862 database, delete keys/value pairs and finally how to enumerate the
863 contents of the database.
871 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
872 or die "Cannot open file 'fruit': $!\n";
874 # Add a few key/value pairs to the file
875 $h{"apple"} = "red" ;
876 $h{"orange"} = "orange" ;
877 $h{"banana"} = "yellow" ;
878 $h{"tomato"} = "red" ;
880 # Check for existence of a key
881 print "Banana Exists\n\n" if $h{"banana"} ;
883 # Delete a key/value pair.
886 # print the contents of the file
887 while (($k, $v) = each %h)
888 { print "$k -> $v\n" }
900 Note that the like ordinary associative arrays, the order of the keys
901 retrieved is in an apparently random order.
905 The DB_BTREE format is useful when you want to store data in a given
906 order. By default the keys will be stored in lexical order, but as you
907 will see from the example shown in the next section, it is very easy to
908 define your own sorting function.
910 =head2 Changing the BTREE sort order
912 This script shows how to override the default sorting algorithm that
913 BTREE uses. Instead of using the normal lexical ordering, a case
914 insensitive compare function will be used.
924 my ($key1, $key2) = @_ ;
925 "\L$key1" cmp "\L$key2" ;
928 # specify the Perl sub that will do the comparison
929 $DB_BTREE->{'compare'} = \&Compare ;
932 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
933 or die "Cannot open file 'tree': $!\n" ;
935 # Add a key/value pair to the file
936 $h{'Wall'} = 'Larry' ;
937 $h{'Smith'} = 'John' ;
938 $h{'mouse'} = 'mickey' ;
939 $h{'duck'} = 'donald' ;
944 # Cycle through the keys printing them in order.
945 # Note it is not necessary to sort the keys as
946 # the btree will have kept them in order automatically.
952 Here is the output from the code above.
958 There are a few point to bear in mind if you want to change the
959 ordering in a BTREE database:
965 The new compare function must be specified when you create the database.
969 You cannot change the ordering once the database has been created. Thus
970 you must use the same compare function every time you access the
975 Duplicate keys are entirely defined by the comparison function.
976 In the case-insensitive example above, the keys: 'KEY' and 'key'
977 would be considered duplicates, and assigning to the second one
978 would overwrite the first. If duplicates are allowed for (with the
979 R_DUP flag discussed below), only a single copy of duplicate keys
980 is stored in the database --- so (again with example above) assigning
981 three values to the keys: 'KEY', 'Key', and 'key' would leave just
982 the first key: 'KEY' in the database with three values. For some
983 situations this results in information loss, so care should be taken
984 to provide fully qualified comparison functions when necessary.
985 For example, the above comparison routine could be modified to
986 additionally compare case-sensitively if two keys are equal in the
987 case insensitive comparison:
990 my($key1, $key2) = @_;
991 lc $key1 cmp lc $key2 ||
995 And now you will only have duplicates when the keys themselves
996 are truly the same. (note: in versions of the db library prior to
997 about November 1996, such duplicate keys were retained so it was
998 possible to recover the original keys in sets of keys that
1004 =head2 Handling Duplicate Keys
1006 The BTREE file type optionally allows a single key to be associated
1007 with an arbitrary number of values. This option is enabled by setting
1008 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
1010 There are some difficulties in using the tied hash interface if you
1011 want to manipulate a BTREE database with duplicate keys. Consider this
1018 my ($filename, %h) ;
1020 $filename = "tree" ;
1023 # Enable duplicate records
1024 $DB_BTREE->{'flags'} = R_DUP ;
1026 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1027 or die "Cannot open $filename: $!\n";
1029 # Add some key/value pairs to the file
1030 $h{'Wall'} = 'Larry' ;
1031 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1032 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1033 $h{'Smith'} = 'John' ;
1034 $h{'mouse'} = 'mickey' ;
1036 # iterate through the associative array
1037 # and print each key/value pair.
1038 foreach (sort keys %h)
1039 { print "$_ -> $h{$_}\n" }
1051 As you can see 3 records have been successfully created with key C<Wall>
1052 - the only thing is, when they are retrieved from the database they
1053 I<seem> to have the same value, namely C<Larry>. The problem is caused
1054 by the way that the associative array interface works. Basically, when
1055 the associative array interface is used to fetch the value associated
1056 with a given key, it will only ever retrieve the first value.
1058 Although it may not be immediately obvious from the code above, the
1059 associative array interface can be used to write values with duplicate
1060 keys, but it cannot be used to read them back from the database.
1062 The way to get around this problem is to use the Berkeley DB API method
1063 called C<seq>. This method allows sequential access to key/value
1064 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1065 and the API in general.
1067 Here is the script above rewritten using the C<seq> API method.
1073 my ($filename, $x, %h, $status, $key, $value) ;
1075 $filename = "tree" ;
1078 # Enable duplicate records
1079 $DB_BTREE->{'flags'} = R_DUP ;
1081 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1082 or die "Cannot open $filename: $!\n";
1084 # Add some key/value pairs to the file
1085 $h{'Wall'} = 'Larry' ;
1086 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1087 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1088 $h{'Smith'} = 'John' ;
1089 $h{'mouse'} = 'mickey' ;
1091 # iterate through the btree using seq
1092 # and print each key/value pair.
1094 for ($status = $x->seq($key, $value, R_FIRST) ;
1096 $status = $x->seq($key, $value, R_NEXT) )
1097 { print "$key -> $value\n" }
1110 This time we have got all the key/value pairs, including the multiple
1111 values associated with the key C<Wall>.
1113 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1114 a few utility methods.
1116 =head2 The get_dup() Method
1118 The C<get_dup> method assists in
1119 reading duplicate values from BTREE databases. The method can take the
1122 $count = $x->get_dup($key) ;
1123 @list = $x->get_dup($key) ;
1124 %list = $x->get_dup($key, 1) ;
1126 In a scalar context the method returns the number of values associated
1127 with the key, C<$key>.
1129 In list context, it returns all the values which match C<$key>. Note
1130 that the values will be returned in an apparently random order.
1132 In list context, if the second parameter is present and evaluates
1133 TRUE, the method returns an associative array. The keys of the
1134 associative array correspond to the values that matched in the BTREE
1135 and the values of the array are a count of the number of times that
1136 particular value occurred in the BTREE.
1138 So assuming the database created above, we can use C<get_dup> like
1145 my ($filename, $x, %h) ;
1147 $filename = "tree" ;
1149 # Enable duplicate records
1150 $DB_BTREE->{'flags'} = R_DUP ;
1152 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1153 or die "Cannot open $filename: $!\n";
1155 my $cnt = $x->get_dup("Wall") ;
1156 print "Wall occurred $cnt times\n" ;
1158 my %hash = $x->get_dup("Wall", 1) ;
1159 print "Larry is there\n" if $hash{'Larry'} ;
1160 print "There are $hash{'Brick'} Brick Walls\n" ;
1162 my @list = sort $x->get_dup("Wall") ;
1163 print "Wall => [@list]\n" ;
1165 @list = $x->get_dup("Smith") ;
1166 print "Smith => [@list]\n" ;
1168 @list = $x->get_dup("Dog") ;
1169 print "Dog => [@list]\n" ;
1174 Wall occurred 3 times
1176 There are 2 Brick Walls
1177 Wall => [Brick Brick Larry]
1181 =head2 The find_dup() Method
1183 $status = $X->find_dup($key, $value) ;
1185 This method checks for the existence of a specific key/value pair. If the
1186 pair exists, the cursor is left pointing to the pair and the method
1187 returns 0. Otherwise the method returns a non-zero value.
1189 Assuming the database from the previous example:
1195 my ($filename, $x, %h, $found) ;
1197 $filename = "tree" ;
1199 # Enable duplicate records
1200 $DB_BTREE->{'flags'} = R_DUP ;
1202 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1203 or die "Cannot open $filename: $!\n";
1205 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1206 print "Larry Wall is $found there\n" ;
1208 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1209 print "Harry Wall is $found there\n" ;
1217 Harry Wall is not there
1220 =head2 The del_dup() Method
1222 $status = $X->del_dup($key, $value) ;
1224 This method deletes a specific key/value pair. It returns
1225 0 if they exist and have been deleted successfully.
1226 Otherwise the method returns a non-zero value.
1228 Again assuming the existence of the C<tree> database
1234 my ($filename, $x, %h, $found) ;
1236 $filename = "tree" ;
1238 # Enable duplicate records
1239 $DB_BTREE->{'flags'} = R_DUP ;
1241 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1242 or die "Cannot open $filename: $!\n";
1244 $x->del_dup("Wall", "Larry") ;
1246 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1247 print "Larry Wall is $found there\n" ;
1254 Larry Wall is not there
1256 =head2 Matching Partial Keys
1258 The BTREE interface has a feature which allows partial keys to be
1259 matched. This functionality is I<only> available when the C<seq> method
1260 is used along with the R_CURSOR flag.
1262 $x->seq($key, $value, R_CURSOR) ;
1264 Here is the relevant quote from the dbopen man page where it defines
1265 the use of the R_CURSOR flag with seq:
1267 Note, for the DB_BTREE access method, the returned key is not
1268 necessarily an exact match for the specified key. The returned key
1269 is the smallest key greater than or equal to the specified key,
1270 permitting partial key matches and range searches.
1272 In the example script below, the C<match> sub uses this feature to find
1273 and print the first matching key/value pair given a partial key.
1280 my ($filename, $x, %h, $st, $key, $value) ;
1286 my $orig_key = $key ;
1287 $x->seq($key, $value, R_CURSOR) ;
1288 print "$orig_key\t-> $key\t-> $value\n" ;
1291 $filename = "tree" ;
1294 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1295 or die "Cannot open $filename: $!\n";
1297 # Add some key/value pairs to the file
1298 $h{'mouse'} = 'mickey' ;
1299 $h{'Wall'} = 'Larry' ;
1300 $h{'Walls'} = 'Brick' ;
1301 $h{'Smith'} = 'John' ;
1305 print "IN ORDER\n" ;
1306 for ($st = $x->seq($key, $value, R_FIRST) ;
1308 $st = $x->seq($key, $value, R_NEXT) )
1310 { print "$key -> $value\n" }
1312 print "\nPARTIAL MATCH\n" ;
1332 a -> mouse -> mickey
1336 DB_RECNO provides an interface to flat text files. Both variable and
1337 fixed length records are supported.
1339 In order to make RECNO more compatible with Perl, the array offset for
1340 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1342 As with normal Perl arrays, a RECNO array can be accessed using
1343 negative indexes. The index -1 refers to the last element of the array,
1344 -2 the second last, and so on. Attempting to access an element before
1345 the start of the array will raise a fatal run-time error.
1347 =head2 The 'bval' Option
1349 The operation of the bval option warrants some discussion. Here is the
1350 definition of bval from the Berkeley DB 1.85 recno manual page:
1352 The delimiting byte to be used to mark the end of a
1353 record for variable-length records, and the pad charac-
1354 ter for fixed-length records. If no value is speci-
1355 fied, newlines (``\n'') are used to mark the end of
1356 variable-length records and fixed-length records are
1359 The second sentence is wrong. In actual fact bval will only default to
1360 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1361 openinfo parameter is used at all, the value that happens to be in bval
1362 will be used. That means you always have to specify bval when making
1363 use of any of the options in the openinfo parameter. This documentation
1364 error will be fixed in the next release of Berkeley DB.
1366 That clarifies the situation with regards Berkeley DB itself. What
1367 about B<DB_File>? Well, the behavior defined in the quote above is
1368 quite useful, so B<DB_File> conforms to it.
1370 That means that you can specify other options (e.g. cachesize) and
1371 still have bval default to C<"\n"> for variable length records, and
1372 space for fixed length records.
1374 Also note that the bval option only allows you to specify a single byte
1377 =head2 A Simple Example
1379 Here is a simple example that uses RECNO (if you are using a version
1380 of Perl earlier than 5.004_57 this example won't work -- see
1381 L<Extra RECNO Methods> for a workaround).
1387 my $filename = "text" ;
1391 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1392 or die "Cannot open file 'text': $!\n" ;
1394 # Add a few key/value pairs to the file
1399 push @h, "green", "black" ;
1401 my $elements = scalar @h ;
1402 print "The array contains $elements entries\n" ;
1405 print "popped $last\n" ;
1407 unshift @h, "white" ;
1408 my $first = shift @h ;
1409 print "shifted $first\n" ;
1411 # Check for existence of a key
1412 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1414 # use a negative index
1415 print "The last element is $h[-1]\n" ;
1416 print "The 2nd last element is $h[-2]\n" ;
1420 Here is the output from the script:
1422 The array contains 5 entries
1425 Element 1 Exists with value blue
1426 The last element is green
1427 The 2nd last element is yellow
1429 =head2 Extra RECNO Methods
1431 If you are using a version of Perl earlier than 5.004_57, the tied
1432 array interface is quite limited. In the example script above
1433 C<push>, C<pop>, C<shift>, C<unshift>
1434 or determining the array length will not work with a tied array.
1436 To make the interface more useful for older versions of Perl, a number
1437 of methods are supplied with B<DB_File> to simulate the missing array
1438 operations. All these methods are accessed via the object returned from
1441 Here are the methods:
1445 =item B<$X-E<gt>push(list) ;>
1447 Pushes the elements of C<list> to the end of the array.
1449 =item B<$value = $X-E<gt>pop ;>
1451 Removes and returns the last element of the array.
1453 =item B<$X-E<gt>shift>
1455 Removes and returns the first element of the array.
1457 =item B<$X-E<gt>unshift(list) ;>
1459 Pushes the elements of C<list> to the start of the array.
1461 =item B<$X-E<gt>length>
1463 Returns the number of elements in the array.
1465 =item B<$X-E<gt>splice(offset, length, elements);>
1467 Returns a splice of the array.
1471 =head2 Another Example
1473 Here is a more complete example that makes use of some of the methods
1474 described above. It also makes use of the API interface directly (see
1475 L<THE API INTERFACE>).
1479 my (@h, $H, $file, $i) ;
1487 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1488 or die "Cannot open file $file: $!\n" ;
1490 # first create a text file to play with
1498 # Print the records in order.
1500 # The length method is needed here because evaluating a tied
1501 # array in a scalar context does not return the number of
1502 # elements in the array.
1504 print "\nORIGINAL\n" ;
1505 foreach $i (0 .. $H->length - 1) {
1506 print "$i: $h[$i]\n" ;
1509 # use the push & pop methods
1512 print "\nThe last record was [$a]\n" ;
1514 # and the shift & unshift methods
1516 $H->unshift("first") ;
1517 print "The first record was [$a]\n" ;
1519 # Use the API to add a new record after record 2.
1521 $H->put($i, "Newbie", R_IAFTER) ;
1523 # and a new record before record 1.
1525 $H->put($i, "New One", R_IBEFORE) ;
1530 # now print the records in reverse order
1531 print "\nREVERSE\n" ;
1532 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1533 { print "$i: $h[$i]\n" }
1535 # same again, but use the API functions instead
1536 print "\nREVERSE again\n" ;
1537 my ($s, $k, $v) = (0, 0, 0) ;
1538 for ($s = $H->seq($k, $v, R_LAST) ;
1540 $s = $H->seq($k, $v, R_PREV))
1541 { print "$k: $v\n" }
1546 and this is what it outputs:
1555 The last record was [four]
1556 The first record was [zero]
1580 Rather than iterating through the array, C<@h> like this:
1584 it is necessary to use either this:
1586 foreach $i (0 .. $H->length - 1)
1590 for ($a = $H->get($k, $v, R_FIRST) ;
1592 $a = $H->get($k, $v, R_NEXT) )
1596 Notice that both times the C<put> method was used the record index was
1597 specified using a variable, C<$i>, rather than the literal value
1598 itself. This is because C<put> will return the record number of the
1599 inserted line via that parameter.
1603 =head1 THE API INTERFACE
1605 As well as accessing Berkeley DB using a tied hash or array, it is also
1606 possible to make direct use of most of the API functions defined in the
1607 Berkeley DB documentation.
1609 To do this you need to store a copy of the object returned from the tie.
1611 $db = tie %hash, "DB_File", "filename" ;
1613 Once you have done that, you can access the Berkeley DB API functions
1614 as B<DB_File> methods directly like this:
1616 $db->put($key, $value, R_NOOVERWRITE) ;
1618 B<Important:> If you have saved a copy of the object returned from
1619 C<tie>, the underlying database file will I<not> be closed until both
1620 the tied variable is untied and all copies of the saved object are
1624 $db = tie %hash, "DB_File", "filename"
1625 or die "Cannot tie filename: $!" ;
1630 See L<The untie() Gotcha> for more details.
1632 All the functions defined in L<dbopen> are available except for
1633 close() and dbopen() itself. The B<DB_File> method interface to the
1634 supported functions have been implemented to mirror the way Berkeley DB
1635 works whenever possible. In particular note that:
1641 The methods return a status value. All return 0 on success.
1642 All return -1 to signify an error and set C<$!> to the exact
1643 error code. The return code 1 generally (but not always) means that the
1644 key specified did not exist in the database.
1646 Other return codes are defined. See below and in the Berkeley DB
1647 documentation for details. The Berkeley DB documentation should be used
1648 as the definitive source.
1652 Whenever a Berkeley DB function returns data via one of its parameters,
1653 the equivalent B<DB_File> method does exactly the same.
1657 If you are careful, it is possible to mix API calls with the tied
1658 hash/array interface in the same piece of code. Although only a few of
1659 the methods used to implement the tied interface currently make use of
1660 the cursor, you should always assume that the cursor has been changed
1661 any time the tied hash/array interface is used. As an example, this
1662 code will probably not do what you expect:
1664 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1665 or die "Cannot tie $filename: $!" ;
1667 # Get the first key/value pair and set the cursor
1668 $X->seq($key, $value, R_FIRST) ;
1670 # this line will modify the cursor
1671 $count = scalar keys %x ;
1673 # Get the second key/value pair.
1674 # oops, it didn't, it got the last key/value pair!
1675 $X->seq($key, $value, R_NEXT) ;
1677 The code above can be rearranged to get around the problem, like this:
1679 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1680 or die "Cannot tie $filename: $!" ;
1682 # this line will modify the cursor
1683 $count = scalar keys %x ;
1685 # Get the first key/value pair and set the cursor
1686 $X->seq($key, $value, R_FIRST) ;
1688 # Get the second key/value pair.
1690 $X->seq($key, $value, R_NEXT) ;
1694 All the constants defined in L<dbopen> for use in the flags parameters
1695 in the methods defined below are also available. Refer to the Berkeley
1696 DB documentation for the precise meaning of the flags values.
1698 Below is a list of the methods available.
1702 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1704 Given a key (C<$key>) this method reads the value associated with it
1705 from the database. The value read from the database is returned in the
1706 C<$value> parameter.
1708 If the key does not exist the method returns 1.
1710 No flags are currently defined for this method.
1712 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1714 Stores the key/value pair in the database.
1716 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1717 will have the record number of the inserted key/value pair set.
1719 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1722 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1724 Removes all key/value pairs with key C<$key> from the database.
1726 A return code of 1 means that the requested key was not in the
1729 R_CURSOR is the only valid flag at present.
1731 =item B<$status = $X-E<gt>fd ;>
1733 Returns the file descriptor for the underlying database.
1735 See L<Locking: The Trouble with fd> for an explanation for why you should
1736 not use C<fd> to lock your database.
1738 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1740 This interface allows sequential retrieval from the database. See
1741 L<dbopen> for full details.
1743 Both the C<$key> and C<$value> parameters will be set to the key/value
1744 pair read from the database.
1746 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1747 R_FIRST, R_LAST, R_NEXT and R_PREV.
1749 =item B<$status = $X-E<gt>sync([$flags]) ;>
1751 Flushes any cached buffers to disk.
1753 R_RECNOSYNC is the only valid flag at present.
1759 A DBM Filter is a piece of code that is be used when you I<always>
1760 want to make the same transformation to all keys and/or values in a
1763 There are four methods associated with DBM Filters. All work identically,
1764 and each is used to install (or uninstall) a single DBM Filter. Each
1765 expects a single parameter, namely a reference to a sub. The only
1766 difference between them is the place that the filter is installed.
1772 =item B<filter_store_key>
1774 If a filter has been installed with this method, it will be invoked
1775 every time you write a key to a DBM database.
1777 =item B<filter_store_value>
1779 If a filter has been installed with this method, it will be invoked
1780 every time you write a value to a DBM database.
1783 =item B<filter_fetch_key>
1785 If a filter has been installed with this method, it will be invoked
1786 every time you read a key from a DBM database.
1788 =item B<filter_fetch_value>
1790 If a filter has been installed with this method, it will be invoked
1791 every time you read a value from a DBM database.
1795 You can use any combination of the methods, from none, to all four.
1797 All filter methods return the existing filter, if present, or C<undef>
1800 To delete a filter pass C<undef> to it.
1804 When each filter is called by Perl, a local copy of C<$_> will contain
1805 the key or value to be filtered. Filtering is achieved by modifying
1806 the contents of C<$_>. The return code from the filter is ignored.
1808 =head2 An Example -- the NULL termination problem.
1810 Consider the following scenario. You have a DBM database
1811 that you need to share with a third-party C application. The C application
1812 assumes that I<all> keys and values are NULL terminated. Unfortunately
1813 when Perl writes to DBM databases it doesn't use NULL termination, so
1814 your Perl application will have to manage NULL termination itself. When
1815 you write to the database you will have to use something like this:
1817 $hash{"$key\0"} = "$value\0" ;
1819 Similarly the NULL needs to be taken into account when you are considering
1820 the length of existing keys/values.
1822 It would be much better if you could ignore the NULL terminations issue
1823 in the main application code and have a mechanism that automatically
1824 added the terminating NULL to all keys and values whenever you write to
1825 the database and have them removed when you read from the database. As I'm
1826 sure you have already guessed, this is a problem that DBM Filters can
1834 my $filename = "filt" ;
1837 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1838 or die "Cannot open $filename: $!\n" ;
1840 # Install DBM Filters
1841 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1842 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1843 $db->filter_fetch_value( sub { s/\0$// } ) ;
1844 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1846 $hash{"abc"} = "def" ;
1847 my $a = $hash{"ABC"} ;
1852 Hopefully the contents of each of the filters should be
1853 self-explanatory. Both "fetch" filters remove the terminating NULL,
1854 and both "store" filters add a terminating NULL.
1857 =head2 Another Example -- Key is a C int.
1859 Here is another real-life example. By default, whenever Perl writes to
1860 a DBM database it always writes the key and value as strings. So when
1863 $hash{12345} = "something" ;
1865 the key 12345 will get stored in the DBM database as the 5 byte string
1866 "12345". If you actually want the key to be stored in the DBM database
1867 as a C int, you will have to use C<pack> when writing, and C<unpack>
1870 Here is a DBM Filter that does it:
1876 my $filename = "filt" ;
1880 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1881 or die "Cannot open $filename: $!\n" ;
1883 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1884 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1885 $hash{123} = "def" ;
1890 This time only two filters have been used -- we only need to manipulate
1891 the contents of the key, so it wasn't necessary to install any value
1894 =head1 HINTS AND TIPS
1897 =head2 Locking: The Trouble with fd
1899 Until version 1.72 of this module, the recommended technique for locking
1900 B<DB_File> databases was to flock the filehandle returned from the "fd"
1901 function. Unfortunately this technique has been shown to be fundamentally
1902 flawed (Kudos to David Harris for tracking this down). Use it at your own
1905 The locking technique went like this.
1907 $db = tie(%db, 'DB_File', 'foo.db', O_CREAT|O_RDWR, 0644)
1908 || die "dbcreat foo.db $!";
1910 open(DB_FH, "+<&=$fd") || die "dup $!";
1911 flock (DB_FH, LOCK_EX) || die "flock: $!";
1913 $db{"Tom"} = "Jerry" ;
1915 flock(DB_FH, LOCK_UN);
1920 In simple terms, this is what happens:
1926 Use "tie" to open the database.
1930 Lock the database with fd & flock.
1934 Read & Write to the database.
1938 Unlock and close the database.
1942 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1943 database in step 2 is that an initial block from the database will get
1944 read from disk and cached in memory.
1946 To see why this is a problem, consider what can happen when two processes,
1947 say "A" and "B", both want to update the same B<DB_File> database
1948 using the locking steps outlined above. Assume process "A" has already
1949 opened the database and has a write lock, but it hasn't actually updated
1950 the database yet (it has finished step 2, but not started step 3 yet). Now
1951 process "B" tries to open the same database - step 1 will succeed,
1952 but it will block on step 2 until process "A" releases the lock. The
1953 important thing to notice here is that at this point in time both
1954 processes will have cached identical initial blocks from the database.
1956 Now process "A" updates the database and happens to change some of the
1957 data held in the initial buffer. Process "A" terminates, flushing
1958 all cached data to disk and releasing the database lock. At this point
1959 the database on disk will correctly reflect the changes made by process
1962 With the lock released, process "B" can now continue. It also updates the
1963 database and unfortunately it too modifies the data that was in its
1964 initial buffer. Once that data gets flushed to disk it will overwrite
1965 some/all of the changes process "A" made to the database.
1967 The result of this scenario is at best a database that doesn't contain
1968 what you expect. At worst the database will corrupt.
1970 The above won't happen every time competing process update the same
1971 B<DB_File> database, but it does illustrate why the technique should
1974 =head2 Safe ways to lock a database
1976 Starting with version 2.x, Berkeley DB has internal support for locking.
1977 The companion module to this one, B<BerkeleyDB>, provides an interface
1978 to this locking functionality. If you are serious about locking
1979 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1981 If using B<BerkeleyDB> isn't an option, there are a number of modules
1982 available on CPAN that can be used to implement locking. Each one
1983 implements locking differently and has different goals in mind. It is
1984 therefore worth knowing the difference, so that you can pick the right
1985 one for your application. Here are the three locking wrappers:
1989 =item B<Tie::DB_Lock>
1991 A B<DB_File> wrapper which creates copies of the database file for
1992 read access, so that you have a kind of a multiversioning concurrent read
1993 system. However, updates are still serial. Use for databases where reads
1994 may be lengthy and consistency problems may occur.
1996 =item B<Tie::DB_LockFile>
1998 A B<DB_File> wrapper that has the ability to lock and unlock the database
1999 while it is being used. Avoids the tie-before-flock problem by simply
2000 re-tie-ing the database when you get or drop a lock. Because of the
2001 flexibility in dropping and re-acquiring the lock in the middle of a
2002 session, this can be massaged into a system that will work with long
2003 updates and/or reads if the application follows the hints in the POD
2006 =item B<DB_File::Lock>
2008 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
2009 before tie-ing the database and drops the lock after the untie. Allows
2010 one to use the same lockfile for multiple databases to avoid deadlock
2011 problems, if desired. Use for databases where updates are reads are
2012 quick and simple flock locking semantics are enough.
2016 =head2 Sharing Databases With C Applications
2018 There is no technical reason why a Berkeley DB database cannot be
2019 shared by both a Perl and a C application.
2021 The vast majority of problems that are reported in this area boil down
2022 to the fact that C strings are NULL terminated, whilst Perl strings are
2023 not. See L<DBM FILTERS> for a generic way to work around this problem.
2025 Here is a real example. Netscape 2.0 keeps a record of the locations you
2026 visit along with the time you last visited them in a DB_HASH database.
2027 This is usually stored in the file F<~/.netscape/history.db>. The key
2028 field in the database is the location string and the value field is the
2029 time the location was last visited stored as a 4 byte binary value.
2031 If you haven't already guessed, the location string is stored with a
2032 terminating NULL. This means you need to be careful when accessing the
2035 Here is a snippet of code that is loosely based on Tom Christiansen's
2036 I<ggh> script (available from your nearest CPAN archive in
2037 F<authors/id/TOMC/scripts/nshist.gz>).
2044 my ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2045 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2047 $HISTORY = "$dotdir/.netscape/history.db";
2049 tie %hist_db, 'DB_File', $HISTORY
2050 or die "Cannot open $HISTORY: $!\n" ;;
2052 # Dump the complete database
2053 while ( ($href, $binary_time) = each %hist_db ) {
2055 # remove the terminating NULL
2056 $href =~ s/\x00$// ;
2058 # convert the binary time into a user friendly string
2059 $date = localtime unpack("V", $binary_time);
2060 print "$date $href\n" ;
2063 # check for the existence of a specific key
2064 # remember to add the NULL
2065 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2066 $date = localtime unpack("V", $binary_time) ;
2067 print "Last visited mox.perl.com on $date\n" ;
2070 print "Never visited mox.perl.com\n"
2075 =head2 The untie() Gotcha
2077 If you make use of the Berkeley DB API, it is I<very> strongly
2078 recommended that you read L<perltie/The untie Gotcha>.
2080 Even if you don't currently make use of the API interface, it is still
2083 Here is an example which illustrates the problem from a B<DB_File>
2092 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2093 or die "Cannot tie first time: $!" ;
2099 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2100 or die "Cannot tie second time: $!" ;
2104 When run, the script will produce this error message:
2106 Cannot tie second time: Invalid argument at bad.file line 14.
2108 Although the error message above refers to the second tie() statement
2109 in the script, the source of the problem is really with the untie()
2110 statement that precedes it.
2112 Having read L<perltie> you will probably have already guessed that the
2113 error is caused by the extra copy of the tied object stored in C<$X>.
2114 If you haven't, then the problem boils down to the fact that the
2115 B<DB_File> destructor, DESTROY, will not be called until I<all>
2116 references to the tied object are destroyed. Both the tied variable,
2117 C<%x>, and C<$X> above hold a reference to the object. The call to
2118 untie() will destroy the first, but C<$X> still holds a valid
2119 reference, so the destructor will not get called and the database file
2120 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2121 attempt to open a database that is already open via the catch-all
2122 "Invalid argument" doesn't help.
2124 If you run the script with the C<-w> flag the error message becomes:
2126 untie attempted while 1 inner references still exist at bad.file line 12.
2127 Cannot tie second time: Invalid argument at bad.file line 14.
2129 which pinpoints the real problem. Finally the script can now be
2130 modified to fix the original problem by destroying the API object
2139 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2143 =head1 COMMON QUESTIONS
2145 =head2 Why is there Perl source in my database?
2147 If you look at the contents of a database file created by DB_File,
2148 there can sometimes be part of a Perl script included in it.
2150 This happens because Berkeley DB uses dynamic memory to allocate
2151 buffers which will subsequently be written to the database file. Being
2152 dynamic, the memory could have been used for anything before DB
2153 malloced it. As Berkeley DB doesn't clear the memory once it has been
2154 allocated, the unused portions will contain random junk. In the case
2155 where a Perl script gets written to the database, the random junk will
2156 correspond to an area of dynamic memory that happened to be used during
2157 the compilation of the script.
2159 Unless you don't like the possibility of there being part of your Perl
2160 scripts embedded in a database file, this is nothing to worry about.
2162 =head2 How do I store complex data structures with DB_File?
2164 Although B<DB_File> cannot do this directly, there is a module which
2165 can layer transparently over B<DB_File> to accomplish this feat.
2167 Check out the MLDBM module, available on CPAN in the directory
2168 F<modules/by-module/MLDBM>.
2170 =head2 What does "Invalid Argument" mean?
2172 You will get this error message when one of the parameters in the
2173 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2174 get wrong, so it can be difficult to figure out which one it is.
2176 Here are a couple of possibilities:
2182 Attempting to reopen a database without closing it.
2186 Using the O_WRONLY flag.
2190 =head2 What does "Bareword 'DB_File' not allowed" mean?
2192 You will encounter this particular error message when you have the
2193 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2194 Consider this script:
2200 tie %x, DB_File, "filename" ;
2202 Running it produces the error in question:
2204 Bareword "DB_File" not allowed while "strict subs" in use
2206 To get around the error, place the word C<DB_File> in either single or
2207 double quotes, like this:
2209 tie %x, "DB_File", "filename" ;
2211 Although it might seem like a real pain, it is really worth the effort
2212 of having a C<use strict> in all your scripts.
2216 Articles that are either about B<DB_File> or make use of it.
2222 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2223 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2229 Moved to the Changes file.
2233 Some older versions of Berkeley DB had problems with fixed length
2234 records using the RECNO file format. This problem has been fixed since
2235 version 1.85 of Berkeley DB.
2237 I am sure there are bugs in the code. If you do find any, or can
2238 suggest any enhancements, I would welcome your comments.
2242 B<DB_File> comes with the standard Perl source distribution. Look in
2243 the directory F<ext/DB_File>. Given the amount of time between releases
2244 of Perl the version that ships with Perl is quite likely to be out of
2245 date, so the most recent version can always be found on CPAN (see
2246 L<perlmodlib/CPAN> for details), in the directory
2247 F<modules/by-module/DB_File>.
2249 This version of B<DB_File> will work with either version 1.x, 2.x or
2250 3.x of Berkeley DB, but is limited to the functionality provided by
2253 The official web site for Berkeley DB is F<http://www.oracle.com/technology/products/berkeley-db/db/index.html>.
2254 All versions of Berkeley DB are available there.
2256 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2257 archive in F<src/misc/db.1.85.tar.gz>.
2259 If you are running IRIX, then get Berkeley DB version 1 from
2260 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2261 compile properly on IRIX 5.3.
2265 Copyright (c) 1995-2007 Paul Marquess. All rights reserved. This program
2266 is free software; you can redistribute it and/or modify it under the
2267 same terms as Perl itself.
2269 Although B<DB_File> is covered by the Perl license, the library it
2270 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2271 copyright and its own license. Please take the time to read it.
2273 Here are are few words taken from the Berkeley DB FAQ (at
2274 F<http://www.oracle.com/technology/products/berkeley-db/db/index.html>) regarding the license:
2276 Do I have to license DB to use it in Perl scripts?
2278 No. The Berkeley DB license requires that software that uses
2279 Berkeley DB be freely redistributable. In the case of Perl, that
2280 software is Perl, and not your scripts. Any Perl scripts that you
2281 write are your property, including scripts that make use of
2282 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2283 place any restriction on what you may do with them.
2285 If you are in any doubt about the license situation, contact either the
2286 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2291 L<perl>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2296 The DB_File interface was written by Paul Marquess
2297 E<lt>pmqs@cpan.orgE<gt>.