1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (pmqs@cpan.org)
4 # last modified 12th March 2005
7 # Copyright (c) 1995-2005 Paul Marquess. All rights reserved.
8 # This program is free software; you can redistribute it and/or
9 # modify it under the same terms as Perl itself.
12 package DB_File::HASHINFO ;
20 @DB_File::HASHINFO::ISA = qw(Tie::Hash);
53 return $self->{GOT}{$key} if exists $self->{VALID}{$key} ;
56 croak "${pkg}::FETCH - Unknown element '$key'" ;
66 my $type = $self->{VALID}{$key};
70 croak "Key '$key' not associated with a code reference"
71 if $type == 2 && !ref $value && ref $value ne 'CODE';
72 $self->{GOT}{$key} = $value ;
77 croak "${pkg}::STORE - Unknown element '$key'" ;
85 if ( exists $self->{VALID}{$key} )
87 delete $self->{GOT}{$key} ;
92 croak "DB_File::HASHINFO::DELETE - Unknown element '$key'" ;
100 exists $self->{VALID}{$key} ;
108 croak ref($self) . " does not define the method ${method}" ;
111 sub FIRSTKEY { my $self = shift ; $self->NotHere("FIRSTKEY") }
112 sub NEXTKEY { my $self = shift ; $self->NotHere("NEXTKEY") }
113 sub CLEAR { my $self = shift ; $self->NotHere("CLEAR") }
115 package DB_File::RECNOINFO ;
120 @DB_File::RECNOINFO::ISA = qw(DB_File::HASHINFO) ;
126 bless { VALID => { map {$_, 1}
127 qw( bval cachesize psize flags lorder reclen bfname )
133 package DB_File::BTREEINFO ;
138 @DB_File::BTREEINFO::ISA = qw(DB_File::HASHINFO) ;
163 our ($VERSION, @ISA, @EXPORT, $AUTOLOAD, $DB_BTREE, $DB_HASH, $DB_RECNO);
164 our ($db_version, $use_XSLoader, $splice_end_array);
171 local $SIG{__WARN__} = sub {$splice_end_array = "@_";};
172 my @a =(1); splice(@a, 3);
174 ($splice_end_array =~ /^splice\(\) offset past end of array at /);
177 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
178 $DB_BTREE = new DB_File::BTREEINFO ;
179 $DB_HASH = new DB_File::HASHINFO ;
180 $DB_RECNO = new DB_File::RECNOINFO ;
187 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
192 @ISA = qw(DynaLoader);
196 push @ISA, qw(Tie::Hash Exporter);
198 $DB_BTREE $DB_HASH $DB_RECNO
233 ($constname = $AUTOLOAD) =~ s/.*:://;
234 my ($error, $val) = constant($constname);
235 Carp::croak $error if $error;
237 *{$AUTOLOAD} = sub { $val };
243 # Make all Fcntl O_XXX constants available for importing
245 my @O = grep /^O_/, @Fcntl::EXPORT;
246 Fcntl->import(@O); # first we import what we want to export
251 { XSLoader::load("DB_File", $VERSION)}
253 { bootstrap DB_File $VERSION }
255 # Preloaded methods go here. Autoload methods go after __END__, and are
256 # processed by the autosplit program.
258 sub tie_hash_or_array
261 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
263 $arg[4] = tied %{ $arg[4] }
264 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
266 $arg[2] = O_CREAT()|O_RDWR() if @arg >=3 && ! defined $arg[2];
267 $arg[3] = 0666 if @arg >=4 && ! defined $arg[3];
269 # make recno in Berkeley DB version 2 (or better) work like
270 # recno in version 1.
271 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
272 $arg[1] and ! -e $arg[1]) {
273 open(FH, ">$arg[1]") or return undef ;
275 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
278 DoTie_($tieHASH, @arg) ;
283 tie_hash_or_array(@_) ;
288 tie_hash_or_array(@_) ;
296 my $status = $self->seq($key, $value, R_FIRST());
299 while ($status == 0) {
301 $status = $self->seq($key, $value, R_NEXT());
303 foreach $key (reverse @keys) {
304 my $s = $self->del($key);
314 my $current_length = $self->length() ;
316 if ($length < $current_length) {
318 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
321 elsif ($length > $current_length) {
322 $self->put($length-1, "") ;
331 if (not defined $offset) {
332 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
336 my $length = @_ ? shift : 0;
337 # Carping about definedness comes _after_ the OFFSET sanity check.
338 # This is so we get the same error messages as Perl's splice().
343 my $size = $self->FETCHSIZE();
345 # 'If OFFSET is negative then it start that far from the end of
349 my $new_offset = $size + $offset;
350 if ($new_offset < 0) {
351 die "Modification of non-creatable array value attempted, "
352 . "subscript $offset";
354 $offset = $new_offset;
357 if (not defined $length) {
358 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
362 if ($offset > $size) {
364 warnings::warnif('misc', 'splice() offset past end of array')
365 if $splice_end_array;
368 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
369 if (not defined $length) {
370 $length = $size - $offset;
373 # 'If LENGTH is negative, leave that many elements off the end of
377 $length = $size - $offset + $length;
380 # The user must have specified a length bigger than the
381 # length of the array passed in. But perl's splice()
382 # doesn't catch this, it just behaves as for length=0.
388 if ($length > $size - $offset) {
389 $length = $size - $offset;
392 # $num_elems holds the current number of elements in the database.
393 my $num_elems = $size;
395 # 'Removes the elements designated by OFFSET and LENGTH from an
399 foreach (0 .. $length - 1) {
401 my $status = $self->get($offset, $old);
403 my $msg = "error from Berkeley DB on get($offset, \$old)";
405 $msg .= ' (no such element?)';
408 $msg .= ": error status $status";
409 if (defined $! and $! ne '') {
410 $msg .= ", message $!";
417 $status = $self->del($offset);
419 my $msg = "error from Berkeley DB on del($offset)";
421 $msg .= ' (no such element?)';
424 $msg .= ": error status $status";
425 if (defined $! and $! ne '') {
426 $msg .= ", message $!";
435 # ...'and replaces them with the elements of LIST, if any.'
437 while (defined (my $elem = shift @list)) {
440 if ($pos >= $num_elems) {
441 $status = $self->put($pos, $elem);
444 $status = $self->put($pos, $elem, $self->R_IBEFORE);
448 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
450 $msg .= ' (no such element?)';
453 $msg .= ", error status $status";
454 if (defined $! and $! ne '') {
455 $msg .= ", message $!";
461 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
469 # 'In list context, returns the elements removed from the
474 elsif (defined wantarray and not wantarray) {
475 # 'In scalar context, returns the last element removed, or
476 # undef if no elements are removed.'
479 my $last = pop @removed;
486 elsif (not defined wantarray) {
491 sub ::DB_File::splice { &SPLICE }
495 croak "Usage: \$db->find_dup(key,value)\n"
499 my ($origkey, $value_wanted) = @_ ;
500 my ($key, $value) = ($origkey, 0);
503 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
505 $status = $db->seq($key, $value, R_NEXT() ) ) {
507 return 0 if $key eq $origkey and $value eq $value_wanted ;
515 croak "Usage: \$db->del_dup(key,value)\n"
519 my ($key, $value) = @_ ;
520 my ($status) = $db->find_dup($key, $value) ;
521 return $status if $status != 0 ;
523 $status = $db->del($key, R_CURSOR() ) ;
529 croak "Usage: \$db->get_dup(key [,flag])\n"
530 unless @_ == 2 or @_ == 3 ;
537 my $wantarray = wantarray ;
543 # iterate through the database until either EOF ($status == 0)
544 # or a different key is encountered ($key ne $origkey).
545 for ($status = $db->seq($key, $value, R_CURSOR()) ;
546 $status == 0 and $key eq $origkey ;
547 $status = $db->seq($key, $value, R_NEXT()) ) {
549 # save the value or count number of matches
552 { ++ $values{$value} }
554 { push (@values, $value) }
561 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
570 DB_File - Perl5 access to Berkeley DB version 1.x
576 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
577 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
578 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
580 $status = $X->del($key [, $flags]) ;
581 $status = $X->put($key, $value [, $flags]) ;
582 $status = $X->get($key, $value [, $flags]) ;
583 $status = $X->seq($key, $value, $flags) ;
584 $status = $X->sync([$flags]) ;
588 $count = $X->get_dup($key) ;
589 @list = $X->get_dup($key) ;
590 %list = $X->get_dup($key, 1) ;
591 $status = $X->find_dup($key, $value) ;
592 $status = $X->del_dup($key, $value) ;
600 @r = $X->splice(offset, length, elements);
603 $old_filter = $db->filter_store_key ( sub { ... } ) ;
604 $old_filter = $db->filter_store_value( sub { ... } ) ;
605 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
606 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
613 B<DB_File> is a module which allows Perl programs to make use of the
614 facilities provided by Berkeley DB version 1.x (if you have a newer
615 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
616 It is assumed that you have a copy of the Berkeley DB manual pages at
617 hand when reading this documentation. The interface defined here
618 mirrors the Berkeley DB interface closely.
620 Berkeley DB is a C library which provides a consistent interface to a
621 number of database formats. B<DB_File> provides an interface to all
622 three of the database types currently supported by Berkeley DB.
630 This database type allows arbitrary key/value pairs to be stored in data
631 files. This is equivalent to the functionality provided by other
632 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
633 the files created using DB_HASH are not compatible with any of the
634 other packages mentioned.
636 A default hashing algorithm, which will be adequate for most
637 applications, is built into Berkeley DB. If you do need to use your own
638 hashing algorithm it is possible to write your own in Perl and have
639 B<DB_File> use it instead.
643 The btree format allows arbitrary key/value pairs to be stored in a
644 sorted, balanced binary tree.
646 As with the DB_HASH format, it is possible to provide a user defined
647 Perl routine to perform the comparison of keys. By default, though, the
648 keys are stored in lexical order.
652 DB_RECNO allows both fixed-length and variable-length flat text files
653 to be manipulated using the same key/value pair interface as in DB_HASH
654 and DB_BTREE. In this case the key will consist of a record (line)
659 =head2 Using DB_File with Berkeley DB version 2 or greater
661 Although B<DB_File> is intended to be used with Berkeley DB version 1,
662 it can also be used with version 2, 3 or 4. In this case the interface is
663 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
664 version 2 or greater interface differs, B<DB_File> arranges for it to work
665 like version 1. This feature allows B<DB_File> scripts that were built
666 with version 1 to be migrated to version 2 or greater without any changes.
668 If you want to make use of the new features available in Berkeley DB
669 2.x or greater, use the Perl module B<BerkeleyDB> instead.
671 B<Note:> The database file format has changed multiple times in Berkeley
672 DB version 2, 3 and 4. If you cannot recreate your databases, you
673 must dump any existing databases with either the C<db_dump> or the
674 C<db_dump185> utility that comes with Berkeley DB.
675 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
676 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
677 documentation for further details.
679 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
682 =head2 Interface to Berkeley DB
684 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
685 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
686 allows B<DB_File> to access Berkeley DB files using either an
687 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
688 array (for the DB_RECNO file type).
690 In addition to the tie() interface, it is also possible to access most
691 of the functions provided in the Berkeley DB API directly.
692 See L<THE API INTERFACE>.
694 =head2 Opening a Berkeley DB Database File
696 Berkeley DB uses the function dbopen() to open or create a database.
697 Here is the C prototype for dbopen():
700 dbopen (const char * file, int flags, int mode,
701 DBTYPE type, const void * openinfo)
703 The parameter C<type> is an enumeration which specifies which of the 3
704 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
705 Depending on which of these is actually chosen, the final parameter,
706 I<openinfo> points to a data structure which allows tailoring of the
707 specific interface method.
709 This interface is handled slightly differently in B<DB_File>. Here is
710 an equivalent call using B<DB_File>:
712 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
714 The C<filename>, C<flags> and C<mode> parameters are the direct
715 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
716 performs the function of both the C<type> and C<openinfo> parameters in
719 In the example above $DB_HASH is actually a pre-defined reference to a
720 hash object. B<DB_File> has three of these pre-defined references.
721 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
723 The keys allowed in each of these pre-defined references is limited to
724 the names used in the equivalent C structure. So, for example, the
725 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
726 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
728 To change one of these elements, just assign to it like this:
730 $DB_HASH->{'cachesize'} = 10000 ;
732 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
733 usually adequate for most applications. If you do need to create extra
734 instances of these objects, constructors are available for each file
737 Here are examples of the constructors and the valid options available
738 for DB_HASH, DB_BTREE and DB_RECNO respectively.
740 $a = new DB_File::HASHINFO ;
748 $b = new DB_File::BTREEINFO ;
758 $c = new DB_File::RECNOINFO ;
767 The values stored in the hashes above are mostly the direct equivalent
768 of their C counterpart. Like their C counterparts, all are set to a
769 default values - that means you don't have to set I<all> of the
770 values when you only want to change one. Here is an example:
772 $a = new DB_File::HASHINFO ;
773 $a->{'cachesize'} = 12345 ;
774 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
776 A few of the options need extra discussion here. When used, the C
777 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
778 to C functions. In B<DB_File> these keys are used to store references
779 to Perl subs. Below are templates for each of the subs:
785 # return the hash value for $data
791 my ($key, $key2) = @_ ;
793 # return 0 if $key1 eq $key2
794 # -1 if $key1 lt $key2
795 # 1 if $key1 gt $key2
796 return (-1 , 0 or 1) ;
801 my ($key, $key2) = @_ ;
803 # return number of bytes of $key2 which are
804 # necessary to determine that it is greater than $key1
808 See L<Changing the BTREE sort order> for an example of using the
811 If you are using the DB_RECNO interface and you intend making use of
812 C<bval>, you should check out L<The 'bval' Option>.
814 =head2 Default Parameters
816 It is possible to omit some or all of the final 4 parameters in the
817 call to C<tie> and let them take default values. As DB_HASH is the most
818 common file format used, the call:
820 tie %A, "DB_File", "filename" ;
824 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
826 It is also possible to omit the filename parameter as well, so the
833 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
835 See L<In Memory Databases> for a discussion on the use of C<undef>
836 in place of a filename.
838 =head2 In Memory Databases
840 Berkeley DB allows the creation of in-memory databases by using NULL
841 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
842 uses C<undef> instead of NULL to provide this functionality.
846 The DB_HASH file format is probably the most commonly used of the three
847 file formats that B<DB_File> supports. It is also very straightforward
850 =head2 A Simple Example
852 This example shows how to create a database, add key/value pairs to the
853 database, delete keys/value pairs and finally how to enumerate the
854 contents of the database.
862 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
863 or die "Cannot open file 'fruit': $!\n";
865 # Add a few key/value pairs to the file
866 $h{"apple"} = "red" ;
867 $h{"orange"} = "orange" ;
868 $h{"banana"} = "yellow" ;
869 $h{"tomato"} = "red" ;
871 # Check for existence of a key
872 print "Banana Exists\n\n" if $h{"banana"} ;
874 # Delete a key/value pair.
877 # print the contents of the file
878 while (($k, $v) = each %h)
879 { print "$k -> $v\n" }
891 Note that the like ordinary associative arrays, the order of the keys
892 retrieved is in an apparently random order.
896 The DB_BTREE format is useful when you want to store data in a given
897 order. By default the keys will be stored in lexical order, but as you
898 will see from the example shown in the next section, it is very easy to
899 define your own sorting function.
901 =head2 Changing the BTREE sort order
903 This script shows how to override the default sorting algorithm that
904 BTREE uses. Instead of using the normal lexical ordering, a case
905 insensitive compare function will be used.
915 my ($key1, $key2) = @_ ;
916 "\L$key1" cmp "\L$key2" ;
919 # specify the Perl sub that will do the comparison
920 $DB_BTREE->{'compare'} = \&Compare ;
923 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
924 or die "Cannot open file 'tree': $!\n" ;
926 # Add a key/value pair to the file
927 $h{'Wall'} = 'Larry' ;
928 $h{'Smith'} = 'John' ;
929 $h{'mouse'} = 'mickey' ;
930 $h{'duck'} = 'donald' ;
935 # Cycle through the keys printing them in order.
936 # Note it is not necessary to sort the keys as
937 # the btree will have kept them in order automatically.
943 Here is the output from the code above.
949 There are a few point to bear in mind if you want to change the
950 ordering in a BTREE database:
956 The new compare function must be specified when you create the database.
960 You cannot change the ordering once the database has been created. Thus
961 you must use the same compare function every time you access the
966 Duplicate keys are entirely defined by the comparison function.
967 In the case-insensitive example above, the keys: 'KEY' and 'key'
968 would be considered duplicates, and assigning to the second one
969 would overwrite the first. If duplicates are allowed for (with the
970 R_DUP flag discussed below), only a single copy of duplicate keys
971 is stored in the database --- so (again with example above) assigning
972 three values to the keys: 'KEY', 'Key', and 'key' would leave just
973 the first key: 'KEY' in the database with three values. For some
974 situations this results in information loss, so care should be taken
975 to provide fully qualified comparison functions when necessary.
976 For example, the above comparison routine could be modified to
977 additionally compare case-sensitively if two keys are equal in the
978 case insensitive comparison:
981 my($key1, $key2) = @_;
982 lc $key1 cmp lc $key2 ||
986 And now you will only have duplicates when the keys themselves
987 are truly the same. (note: in versions of the db library prior to
988 about November 1996, such duplicate keys were retained so it was
989 possible to recover the original keys in sets of keys that
995 =head2 Handling Duplicate Keys
997 The BTREE file type optionally allows a single key to be associated
998 with an arbitrary number of values. This option is enabled by setting
999 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
1001 There are some difficulties in using the tied hash interface if you
1002 want to manipulate a BTREE database with duplicate keys. Consider this
1009 my ($filename, %h) ;
1011 $filename = "tree" ;
1014 # Enable duplicate records
1015 $DB_BTREE->{'flags'} = R_DUP ;
1017 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1018 or die "Cannot open $filename: $!\n";
1020 # Add some key/value pairs to the file
1021 $h{'Wall'} = 'Larry' ;
1022 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1023 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1024 $h{'Smith'} = 'John' ;
1025 $h{'mouse'} = 'mickey' ;
1027 # iterate through the associative array
1028 # and print each key/value pair.
1029 foreach (sort keys %h)
1030 { print "$_ -> $h{$_}\n" }
1042 As you can see 3 records have been successfully created with key C<Wall>
1043 - the only thing is, when they are retrieved from the database they
1044 I<seem> to have the same value, namely C<Larry>. The problem is caused
1045 by the way that the associative array interface works. Basically, when
1046 the associative array interface is used to fetch the value associated
1047 with a given key, it will only ever retrieve the first value.
1049 Although it may not be immediately obvious from the code above, the
1050 associative array interface can be used to write values with duplicate
1051 keys, but it cannot be used to read them back from the database.
1053 The way to get around this problem is to use the Berkeley DB API method
1054 called C<seq>. This method allows sequential access to key/value
1055 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1056 and the API in general.
1058 Here is the script above rewritten using the C<seq> API method.
1064 my ($filename, $x, %h, $status, $key, $value) ;
1066 $filename = "tree" ;
1069 # Enable duplicate records
1070 $DB_BTREE->{'flags'} = R_DUP ;
1072 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1073 or die "Cannot open $filename: $!\n";
1075 # Add some key/value pairs to the file
1076 $h{'Wall'} = 'Larry' ;
1077 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1078 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1079 $h{'Smith'} = 'John' ;
1080 $h{'mouse'} = 'mickey' ;
1082 # iterate through the btree using seq
1083 # and print each key/value pair.
1085 for ($status = $x->seq($key, $value, R_FIRST) ;
1087 $status = $x->seq($key, $value, R_NEXT) )
1088 { print "$key -> $value\n" }
1101 This time we have got all the key/value pairs, including the multiple
1102 values associated with the key C<Wall>.
1104 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1105 a few utility methods.
1107 =head2 The get_dup() Method
1109 The C<get_dup> method assists in
1110 reading duplicate values from BTREE databases. The method can take the
1113 $count = $x->get_dup($key) ;
1114 @list = $x->get_dup($key) ;
1115 %list = $x->get_dup($key, 1) ;
1117 In a scalar context the method returns the number of values associated
1118 with the key, C<$key>.
1120 In list context, it returns all the values which match C<$key>. Note
1121 that the values will be returned in an apparently random order.
1123 In list context, if the second parameter is present and evaluates
1124 TRUE, the method returns an associative array. The keys of the
1125 associative array correspond to the values that matched in the BTREE
1126 and the values of the array are a count of the number of times that
1127 particular value occurred in the BTREE.
1129 So assuming the database created above, we can use C<get_dup> like
1136 my ($filename, $x, %h) ;
1138 $filename = "tree" ;
1140 # Enable duplicate records
1141 $DB_BTREE->{'flags'} = R_DUP ;
1143 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1144 or die "Cannot open $filename: $!\n";
1146 my $cnt = $x->get_dup("Wall") ;
1147 print "Wall occurred $cnt times\n" ;
1149 my %hash = $x->get_dup("Wall", 1) ;
1150 print "Larry is there\n" if $hash{'Larry'} ;
1151 print "There are $hash{'Brick'} Brick Walls\n" ;
1153 my @list = sort $x->get_dup("Wall") ;
1154 print "Wall => [@list]\n" ;
1156 @list = $x->get_dup("Smith") ;
1157 print "Smith => [@list]\n" ;
1159 @list = $x->get_dup("Dog") ;
1160 print "Dog => [@list]\n" ;
1165 Wall occurred 3 times
1167 There are 2 Brick Walls
1168 Wall => [Brick Brick Larry]
1172 =head2 The find_dup() Method
1174 $status = $X->find_dup($key, $value) ;
1176 This method checks for the existence of a specific key/value pair. If the
1177 pair exists, the cursor is left pointing to the pair and the method
1178 returns 0. Otherwise the method returns a non-zero value.
1180 Assuming the database from the previous example:
1186 my ($filename, $x, %h, $found) ;
1188 $filename = "tree" ;
1190 # Enable duplicate records
1191 $DB_BTREE->{'flags'} = R_DUP ;
1193 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1194 or die "Cannot open $filename: $!\n";
1196 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1197 print "Larry Wall is $found there\n" ;
1199 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1200 print "Harry Wall is $found there\n" ;
1208 Harry Wall is not there
1211 =head2 The del_dup() Method
1213 $status = $X->del_dup($key, $value) ;
1215 This method deletes a specific key/value pair. It returns
1216 0 if they exist and have been deleted successfully.
1217 Otherwise the method returns a non-zero value.
1219 Again assuming the existence of the C<tree> database
1225 my ($filename, $x, %h, $found) ;
1227 $filename = "tree" ;
1229 # Enable duplicate records
1230 $DB_BTREE->{'flags'} = R_DUP ;
1232 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1233 or die "Cannot open $filename: $!\n";
1235 $x->del_dup("Wall", "Larry") ;
1237 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1238 print "Larry Wall is $found there\n" ;
1245 Larry Wall is not there
1247 =head2 Matching Partial Keys
1249 The BTREE interface has a feature which allows partial keys to be
1250 matched. This functionality is I<only> available when the C<seq> method
1251 is used along with the R_CURSOR flag.
1253 $x->seq($key, $value, R_CURSOR) ;
1255 Here is the relevant quote from the dbopen man page where it defines
1256 the use of the R_CURSOR flag with seq:
1258 Note, for the DB_BTREE access method, the returned key is not
1259 necessarily an exact match for the specified key. The returned key
1260 is the smallest key greater than or equal to the specified key,
1261 permitting partial key matches and range searches.
1263 In the example script below, the C<match> sub uses this feature to find
1264 and print the first matching key/value pair given a partial key.
1271 my ($filename, $x, %h, $st, $key, $value) ;
1277 my $orig_key = $key ;
1278 $x->seq($key, $value, R_CURSOR) ;
1279 print "$orig_key\t-> $key\t-> $value\n" ;
1282 $filename = "tree" ;
1285 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1286 or die "Cannot open $filename: $!\n";
1288 # Add some key/value pairs to the file
1289 $h{'mouse'} = 'mickey' ;
1290 $h{'Wall'} = 'Larry' ;
1291 $h{'Walls'} = 'Brick' ;
1292 $h{'Smith'} = 'John' ;
1296 print "IN ORDER\n" ;
1297 for ($st = $x->seq($key, $value, R_FIRST) ;
1299 $st = $x->seq($key, $value, R_NEXT) )
1301 { print "$key -> $value\n" }
1303 print "\nPARTIAL MATCH\n" ;
1323 a -> mouse -> mickey
1327 DB_RECNO provides an interface to flat text files. Both variable and
1328 fixed length records are supported.
1330 In order to make RECNO more compatible with Perl, the array offset for
1331 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1333 As with normal Perl arrays, a RECNO array can be accessed using
1334 negative indexes. The index -1 refers to the last element of the array,
1335 -2 the second last, and so on. Attempting to access an element before
1336 the start of the array will raise a fatal run-time error.
1338 =head2 The 'bval' Option
1340 The operation of the bval option warrants some discussion. Here is the
1341 definition of bval from the Berkeley DB 1.85 recno manual page:
1343 The delimiting byte to be used to mark the end of a
1344 record for variable-length records, and the pad charac-
1345 ter for fixed-length records. If no value is speci-
1346 fied, newlines (``\n'') are used to mark the end of
1347 variable-length records and fixed-length records are
1350 The second sentence is wrong. In actual fact bval will only default to
1351 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1352 openinfo parameter is used at all, the value that happens to be in bval
1353 will be used. That means you always have to specify bval when making
1354 use of any of the options in the openinfo parameter. This documentation
1355 error will be fixed in the next release of Berkeley DB.
1357 That clarifies the situation with regards Berkeley DB itself. What
1358 about B<DB_File>? Well, the behavior defined in the quote above is
1359 quite useful, so B<DB_File> conforms to it.
1361 That means that you can specify other options (e.g. cachesize) and
1362 still have bval default to C<"\n"> for variable length records, and
1363 space for fixed length records.
1365 Also note that the bval option only allows you to specify a single byte
1368 =head2 A Simple Example
1370 Here is a simple example that uses RECNO (if you are using a version
1371 of Perl earlier than 5.004_57 this example won't work -- see
1372 L<Extra RECNO Methods> for a workaround).
1378 my $filename = "text" ;
1382 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1383 or die "Cannot open file 'text': $!\n" ;
1385 # Add a few key/value pairs to the file
1390 push @h, "green", "black" ;
1392 my $elements = scalar @h ;
1393 print "The array contains $elements entries\n" ;
1396 print "popped $last\n" ;
1398 unshift @h, "white" ;
1399 my $first = shift @h ;
1400 print "shifted $first\n" ;
1402 # Check for existence of a key
1403 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1405 # use a negative index
1406 print "The last element is $h[-1]\n" ;
1407 print "The 2nd last element is $h[-2]\n" ;
1411 Here is the output from the script:
1413 The array contains 5 entries
1416 Element 1 Exists with value blue
1417 The last element is green
1418 The 2nd last element is yellow
1420 =head2 Extra RECNO Methods
1422 If you are using a version of Perl earlier than 5.004_57, the tied
1423 array interface is quite limited. In the example script above
1424 C<push>, C<pop>, C<shift>, C<unshift>
1425 or determining the array length will not work with a tied array.
1427 To make the interface more useful for older versions of Perl, a number
1428 of methods are supplied with B<DB_File> to simulate the missing array
1429 operations. All these methods are accessed via the object returned from
1432 Here are the methods:
1436 =item B<$X-E<gt>push(list) ;>
1438 Pushes the elements of C<list> to the end of the array.
1440 =item B<$value = $X-E<gt>pop ;>
1442 Removes and returns the last element of the array.
1444 =item B<$X-E<gt>shift>
1446 Removes and returns the first element of the array.
1448 =item B<$X-E<gt>unshift(list) ;>
1450 Pushes the elements of C<list> to the start of the array.
1452 =item B<$X-E<gt>length>
1454 Returns the number of elements in the array.
1456 =item B<$X-E<gt>splice(offset, length, elements);>
1458 Returns a splice of the array.
1462 =head2 Another Example
1464 Here is a more complete example that makes use of some of the methods
1465 described above. It also makes use of the API interface directly (see
1466 L<THE API INTERFACE>).
1470 my (@h, $H, $file, $i) ;
1478 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1479 or die "Cannot open file $file: $!\n" ;
1481 # first create a text file to play with
1489 # Print the records in order.
1491 # The length method is needed here because evaluating a tied
1492 # array in a scalar context does not return the number of
1493 # elements in the array.
1495 print "\nORIGINAL\n" ;
1496 foreach $i (0 .. $H->length - 1) {
1497 print "$i: $h[$i]\n" ;
1500 # use the push & pop methods
1503 print "\nThe last record was [$a]\n" ;
1505 # and the shift & unshift methods
1507 $H->unshift("first") ;
1508 print "The first record was [$a]\n" ;
1510 # Use the API to add a new record after record 2.
1512 $H->put($i, "Newbie", R_IAFTER) ;
1514 # and a new record before record 1.
1516 $H->put($i, "New One", R_IBEFORE) ;
1521 # now print the records in reverse order
1522 print "\nREVERSE\n" ;
1523 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1524 { print "$i: $h[$i]\n" }
1526 # same again, but use the API functions instead
1527 print "\nREVERSE again\n" ;
1528 my ($s, $k, $v) = (0, 0, 0) ;
1529 for ($s = $H->seq($k, $v, R_LAST) ;
1531 $s = $H->seq($k, $v, R_PREV))
1532 { print "$k: $v\n" }
1537 and this is what it outputs:
1546 The last record was [four]
1547 The first record was [zero]
1571 Rather than iterating through the array, C<@h> like this:
1575 it is necessary to use either this:
1577 foreach $i (0 .. $H->length - 1)
1581 for ($a = $H->get($k, $v, R_FIRST) ;
1583 $a = $H->get($k, $v, R_NEXT) )
1587 Notice that both times the C<put> method was used the record index was
1588 specified using a variable, C<$i>, rather than the literal value
1589 itself. This is because C<put> will return the record number of the
1590 inserted line via that parameter.
1594 =head1 THE API INTERFACE
1596 As well as accessing Berkeley DB using a tied hash or array, it is also
1597 possible to make direct use of most of the API functions defined in the
1598 Berkeley DB documentation.
1600 To do this you need to store a copy of the object returned from the tie.
1602 $db = tie %hash, "DB_File", "filename" ;
1604 Once you have done that, you can access the Berkeley DB API functions
1605 as B<DB_File> methods directly like this:
1607 $db->put($key, $value, R_NOOVERWRITE) ;
1609 B<Important:> If you have saved a copy of the object returned from
1610 C<tie>, the underlying database file will I<not> be closed until both
1611 the tied variable is untied and all copies of the saved object are
1615 $db = tie %hash, "DB_File", "filename"
1616 or die "Cannot tie filename: $!" ;
1621 See L<The untie() Gotcha> for more details.
1623 All the functions defined in L<dbopen> are available except for
1624 close() and dbopen() itself. The B<DB_File> method interface to the
1625 supported functions have been implemented to mirror the way Berkeley DB
1626 works whenever possible. In particular note that:
1632 The methods return a status value. All return 0 on success.
1633 All return -1 to signify an error and set C<$!> to the exact
1634 error code. The return code 1 generally (but not always) means that the
1635 key specified did not exist in the database.
1637 Other return codes are defined. See below and in the Berkeley DB
1638 documentation for details. The Berkeley DB documentation should be used
1639 as the definitive source.
1643 Whenever a Berkeley DB function returns data via one of its parameters,
1644 the equivalent B<DB_File> method does exactly the same.
1648 If you are careful, it is possible to mix API calls with the tied
1649 hash/array interface in the same piece of code. Although only a few of
1650 the methods used to implement the tied interface currently make use of
1651 the cursor, you should always assume that the cursor has been changed
1652 any time the tied hash/array interface is used. As an example, this
1653 code will probably not do what you expect:
1655 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1656 or die "Cannot tie $filename: $!" ;
1658 # Get the first key/value pair and set the cursor
1659 $X->seq($key, $value, R_FIRST) ;
1661 # this line will modify the cursor
1662 $count = scalar keys %x ;
1664 # Get the second key/value pair.
1665 # oops, it didn't, it got the last key/value pair!
1666 $X->seq($key, $value, R_NEXT) ;
1668 The code above can be rearranged to get around the problem, like this:
1670 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1671 or die "Cannot tie $filename: $!" ;
1673 # this line will modify the cursor
1674 $count = scalar keys %x ;
1676 # Get the first key/value pair and set the cursor
1677 $X->seq($key, $value, R_FIRST) ;
1679 # Get the second key/value pair.
1681 $X->seq($key, $value, R_NEXT) ;
1685 All the constants defined in L<dbopen> for use in the flags parameters
1686 in the methods defined below are also available. Refer to the Berkeley
1687 DB documentation for the precise meaning of the flags values.
1689 Below is a list of the methods available.
1693 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1695 Given a key (C<$key>) this method reads the value associated with it
1696 from the database. The value read from the database is returned in the
1697 C<$value> parameter.
1699 If the key does not exist the method returns 1.
1701 No flags are currently defined for this method.
1703 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1705 Stores the key/value pair in the database.
1707 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1708 will have the record number of the inserted key/value pair set.
1710 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1713 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1715 Removes all key/value pairs with key C<$key> from the database.
1717 A return code of 1 means that the requested key was not in the
1720 R_CURSOR is the only valid flag at present.
1722 =item B<$status = $X-E<gt>fd ;>
1724 Returns the file descriptor for the underlying database.
1726 See L<Locking: The Trouble with fd> for an explanation for why you should
1727 not use C<fd> to lock your database.
1729 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1731 This interface allows sequential retrieval from the database. See
1732 L<dbopen> for full details.
1734 Both the C<$key> and C<$value> parameters will be set to the key/value
1735 pair read from the database.
1737 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1738 R_FIRST, R_LAST, R_NEXT and R_PREV.
1740 =item B<$status = $X-E<gt>sync([$flags]) ;>
1742 Flushes any cached buffers to disk.
1744 R_RECNOSYNC is the only valid flag at present.
1750 A DBM Filter is a piece of code that is be used when you I<always>
1751 want to make the same transformation to all keys and/or values in a
1754 There are four methods associated with DBM Filters. All work identically,
1755 and each is used to install (or uninstall) a single DBM Filter. Each
1756 expects a single parameter, namely a reference to a sub. The only
1757 difference between them is the place that the filter is installed.
1763 =item B<filter_store_key>
1765 If a filter has been installed with this method, it will be invoked
1766 every time you write a key to a DBM database.
1768 =item B<filter_store_value>
1770 If a filter has been installed with this method, it will be invoked
1771 every time you write a value to a DBM database.
1774 =item B<filter_fetch_key>
1776 If a filter has been installed with this method, it will be invoked
1777 every time you read a key from a DBM database.
1779 =item B<filter_fetch_value>
1781 If a filter has been installed with this method, it will be invoked
1782 every time you read a value from a DBM database.
1786 You can use any combination of the methods, from none, to all four.
1788 All filter methods return the existing filter, if present, or C<undef>
1791 To delete a filter pass C<undef> to it.
1795 When each filter is called by Perl, a local copy of C<$_> will contain
1796 the key or value to be filtered. Filtering is achieved by modifying
1797 the contents of C<$_>. The return code from the filter is ignored.
1799 =head2 An Example -- the NULL termination problem.
1801 Consider the following scenario. You have a DBM database
1802 that you need to share with a third-party C application. The C application
1803 assumes that I<all> keys and values are NULL terminated. Unfortunately
1804 when Perl writes to DBM databases it doesn't use NULL termination, so
1805 your Perl application will have to manage NULL termination itself. When
1806 you write to the database you will have to use something like this:
1808 $hash{"$key\0"} = "$value\0" ;
1810 Similarly the NULL needs to be taken into account when you are considering
1811 the length of existing keys/values.
1813 It would be much better if you could ignore the NULL terminations issue
1814 in the main application code and have a mechanism that automatically
1815 added the terminating NULL to all keys and values whenever you write to
1816 the database and have them removed when you read from the database. As I'm
1817 sure you have already guessed, this is a problem that DBM Filters can
1825 my $filename = "filt" ;
1828 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1829 or die "Cannot open $filename: $!\n" ;
1831 # Install DBM Filters
1832 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1833 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1834 $db->filter_fetch_value( sub { s/\0$// } ) ;
1835 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1837 $hash{"abc"} = "def" ;
1838 my $a = $hash{"ABC"} ;
1843 Hopefully the contents of each of the filters should be
1844 self-explanatory. Both "fetch" filters remove the terminating NULL,
1845 and both "store" filters add a terminating NULL.
1848 =head2 Another Example -- Key is a C int.
1850 Here is another real-life example. By default, whenever Perl writes to
1851 a DBM database it always writes the key and value as strings. So when
1854 $hash{12345} = "something" ;
1856 the key 12345 will get stored in the DBM database as the 5 byte string
1857 "12345". If you actually want the key to be stored in the DBM database
1858 as a C int, you will have to use C<pack> when writing, and C<unpack>
1861 Here is a DBM Filter that does it:
1867 my $filename = "filt" ;
1871 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1872 or die "Cannot open $filename: $!\n" ;
1874 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1875 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1876 $hash{123} = "def" ;
1881 This time only two filters have been used -- we only need to manipulate
1882 the contents of the key, so it wasn't necessary to install any value
1885 =head1 HINTS AND TIPS
1888 =head2 Locking: The Trouble with fd
1890 Until version 1.72 of this module, the recommended technique for locking
1891 B<DB_File> databases was to flock the filehandle returned from the "fd"
1892 function. Unfortunately this technique has been shown to be fundamentally
1893 flawed (Kudos to David Harris for tracking this down). Use it at your own
1896 The locking technique went like this.
1898 $db = tie(%db, 'DB_File', 'foo.db', O_CREAT|O_RDWR, 0644)
1899 || die "dbcreat foo.db $!";
1901 open(DB_FH, "+<&=$fd") || die "dup $!";
1902 flock (DB_FH, LOCK_EX) || die "flock: $!";
1904 $db{"Tom"} = "Jerry" ;
1906 flock(DB_FH, LOCK_UN);
1911 In simple terms, this is what happens:
1917 Use "tie" to open the database.
1921 Lock the database with fd & flock.
1925 Read & Write to the database.
1929 Unlock and close the database.
1933 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1934 database in step 2 is that an initial block from the database will get
1935 read from disk and cached in memory.
1937 To see why this is a problem, consider what can happen when two processes,
1938 say "A" and "B", both want to update the same B<DB_File> database
1939 using the locking steps outlined above. Assume process "A" has already
1940 opened the database and has a write lock, but it hasn't actually updated
1941 the database yet (it has finished step 2, but not started step 3 yet). Now
1942 process "B" tries to open the same database - step 1 will succeed,
1943 but it will block on step 2 until process "A" releases the lock. The
1944 important thing to notice here is that at this point in time both
1945 processes will have cached identical initial blocks from the database.
1947 Now process "A" updates the database and happens to change some of the
1948 data held in the initial buffer. Process "A" terminates, flushing
1949 all cached data to disk and releasing the database lock. At this point
1950 the database on disk will correctly reflect the changes made by process
1953 With the lock released, process "B" can now continue. It also updates the
1954 database and unfortunately it too modifies the data that was in its
1955 initial buffer. Once that data gets flushed to disk it will overwrite
1956 some/all of the changes process "A" made to the database.
1958 The result of this scenario is at best a database that doesn't contain
1959 what you expect. At worst the database will corrupt.
1961 The above won't happen every time competing process update the same
1962 B<DB_File> database, but it does illustrate why the technique should
1965 =head2 Safe ways to lock a database
1967 Starting with version 2.x, Berkeley DB has internal support for locking.
1968 The companion module to this one, B<BerkeleyDB>, provides an interface
1969 to this locking functionality. If you are serious about locking
1970 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1972 If using B<BerkeleyDB> isn't an option, there are a number of modules
1973 available on CPAN that can be used to implement locking. Each one
1974 implements locking differently and has different goals in mind. It is
1975 therefore worth knowing the difference, so that you can pick the right
1976 one for your application. Here are the three locking wrappers:
1980 =item B<Tie::DB_Lock>
1982 A B<DB_File> wrapper which creates copies of the database file for
1983 read access, so that you have a kind of a multiversioning concurrent read
1984 system. However, updates are still serial. Use for databases where reads
1985 may be lengthy and consistency problems may occur.
1987 =item B<Tie::DB_LockFile>
1989 A B<DB_File> wrapper that has the ability to lock and unlock the database
1990 while it is being used. Avoids the tie-before-flock problem by simply
1991 re-tie-ing the database when you get or drop a lock. Because of the
1992 flexibility in dropping and re-acquiring the lock in the middle of a
1993 session, this can be massaged into a system that will work with long
1994 updates and/or reads if the application follows the hints in the POD
1997 =item B<DB_File::Lock>
1999 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
2000 before tie-ing the database and drops the lock after the untie. Allows
2001 one to use the same lockfile for multiple databases to avoid deadlock
2002 problems, if desired. Use for databases where updates are reads are
2003 quick and simple flock locking semantics are enough.
2007 =head2 Sharing Databases With C Applications
2009 There is no technical reason why a Berkeley DB database cannot be
2010 shared by both a Perl and a C application.
2012 The vast majority of problems that are reported in this area boil down
2013 to the fact that C strings are NULL terminated, whilst Perl strings are
2014 not. See L<DBM FILTERS> for a generic way to work around this problem.
2016 Here is a real example. Netscape 2.0 keeps a record of the locations you
2017 visit along with the time you last visited them in a DB_HASH database.
2018 This is usually stored in the file F<~/.netscape/history.db>. The key
2019 field in the database is the location string and the value field is the
2020 time the location was last visited stored as a 4 byte binary value.
2022 If you haven't already guessed, the location string is stored with a
2023 terminating NULL. This means you need to be careful when accessing the
2026 Here is a snippet of code that is loosely based on Tom Christiansen's
2027 I<ggh> script (available from your nearest CPAN archive in
2028 F<authors/id/TOMC/scripts/nshist.gz>).
2035 my ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2036 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2038 $HISTORY = "$dotdir/.netscape/history.db";
2040 tie %hist_db, 'DB_File', $HISTORY
2041 or die "Cannot open $HISTORY: $!\n" ;;
2043 # Dump the complete database
2044 while ( ($href, $binary_time) = each %hist_db ) {
2046 # remove the terminating NULL
2047 $href =~ s/\x00$// ;
2049 # convert the binary time into a user friendly string
2050 $date = localtime unpack("V", $binary_time);
2051 print "$date $href\n" ;
2054 # check for the existence of a specific key
2055 # remember to add the NULL
2056 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2057 $date = localtime unpack("V", $binary_time) ;
2058 print "Last visited mox.perl.com on $date\n" ;
2061 print "Never visited mox.perl.com\n"
2066 =head2 The untie() Gotcha
2068 If you make use of the Berkeley DB API, it is I<very> strongly
2069 recommended that you read L<perltie/The untie Gotcha>.
2071 Even if you don't currently make use of the API interface, it is still
2074 Here is an example which illustrates the problem from a B<DB_File>
2083 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2084 or die "Cannot tie first time: $!" ;
2090 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2091 or die "Cannot tie second time: $!" ;
2095 When run, the script will produce this error message:
2097 Cannot tie second time: Invalid argument at bad.file line 14.
2099 Although the error message above refers to the second tie() statement
2100 in the script, the source of the problem is really with the untie()
2101 statement that precedes it.
2103 Having read L<perltie> you will probably have already guessed that the
2104 error is caused by the extra copy of the tied object stored in C<$X>.
2105 If you haven't, then the problem boils down to the fact that the
2106 B<DB_File> destructor, DESTROY, will not be called until I<all>
2107 references to the tied object are destroyed. Both the tied variable,
2108 C<%x>, and C<$X> above hold a reference to the object. The call to
2109 untie() will destroy the first, but C<$X> still holds a valid
2110 reference, so the destructor will not get called and the database file
2111 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2112 attempt to open a database that is already open via the catch-all
2113 "Invalid argument" doesn't help.
2115 If you run the script with the C<-w> flag the error message becomes:
2117 untie attempted while 1 inner references still exist at bad.file line 12.
2118 Cannot tie second time: Invalid argument at bad.file line 14.
2120 which pinpoints the real problem. Finally the script can now be
2121 modified to fix the original problem by destroying the API object
2130 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2134 =head1 COMMON QUESTIONS
2136 =head2 Why is there Perl source in my database?
2138 If you look at the contents of a database file created by DB_File,
2139 there can sometimes be part of a Perl script included in it.
2141 This happens because Berkeley DB uses dynamic memory to allocate
2142 buffers which will subsequently be written to the database file. Being
2143 dynamic, the memory could have been used for anything before DB
2144 malloced it. As Berkeley DB doesn't clear the memory once it has been
2145 allocated, the unused portions will contain random junk. In the case
2146 where a Perl script gets written to the database, the random junk will
2147 correspond to an area of dynamic memory that happened to be used during
2148 the compilation of the script.
2150 Unless you don't like the possibility of there being part of your Perl
2151 scripts embedded in a database file, this is nothing to worry about.
2153 =head2 How do I store complex data structures with DB_File?
2155 Although B<DB_File> cannot do this directly, there is a module which
2156 can layer transparently over B<DB_File> to accomplish this feat.
2158 Check out the MLDBM module, available on CPAN in the directory
2159 F<modules/by-module/MLDBM>.
2161 =head2 What does "Invalid Argument" mean?
2163 You will get this error message when one of the parameters in the
2164 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2165 get wrong, so it can be difficult to figure out which one it is.
2167 Here are a couple of possibilities:
2173 Attempting to reopen a database without closing it.
2177 Using the O_WRONLY flag.
2181 =head2 What does "Bareword 'DB_File' not allowed" mean?
2183 You will encounter this particular error message when you have the
2184 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2185 Consider this script:
2191 tie %x, DB_File, "filename" ;
2193 Running it produces the error in question:
2195 Bareword "DB_File" not allowed while "strict subs" in use
2197 To get around the error, place the word C<DB_File> in either single or
2198 double quotes, like this:
2200 tie %x, "DB_File", "filename" ;
2202 Although it might seem like a real pain, it is really worth the effort
2203 of having a C<use strict> in all your scripts.
2207 Articles that are either about B<DB_File> or make use of it.
2213 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2214 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2220 Moved to the Changes file.
2224 Some older versions of Berkeley DB had problems with fixed length
2225 records using the RECNO file format. This problem has been fixed since
2226 version 1.85 of Berkeley DB.
2228 I am sure there are bugs in the code. If you do find any, or can
2229 suggest any enhancements, I would welcome your comments.
2233 B<DB_File> comes with the standard Perl source distribution. Look in
2234 the directory F<ext/DB_File>. Given the amount of time between releases
2235 of Perl the version that ships with Perl is quite likely to be out of
2236 date, so the most recent version can always be found on CPAN (see
2237 L<perlmodlib/CPAN> for details), in the directory
2238 F<modules/by-module/DB_File>.
2240 This version of B<DB_File> will work with either version 1.x, 2.x or
2241 3.x of Berkeley DB, but is limited to the functionality provided by
2244 The official web site for Berkeley DB is F<http://www.sleepycat.com>.
2245 All versions of Berkeley DB are available there.
2247 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2248 archive in F<src/misc/db.1.85.tar.gz>.
2250 If you are running IRIX, then get Berkeley DB version 1 from
2251 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2252 compile properly on IRIX 5.3.
2256 Copyright (c) 1995-2005 Paul Marquess. All rights reserved. This program
2257 is free software; you can redistribute it and/or modify it under the
2258 same terms as Perl itself.
2260 Although B<DB_File> is covered by the Perl license, the library it
2261 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2262 copyright and its own license. Please take the time to read it.
2264 Here are are few words taken from the Berkeley DB FAQ (at
2265 F<http://www.sleepycat.com>) regarding the license:
2267 Do I have to license DB to use it in Perl scripts?
2269 No. The Berkeley DB license requires that software that uses
2270 Berkeley DB be freely redistributable. In the case of Perl, that
2271 software is Perl, and not your scripts. Any Perl scripts that you
2272 write are your property, including scripts that make use of
2273 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2274 place any restriction on what you may do with them.
2276 If you are in any doubt about the license situation, contact either the
2277 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2282 L<perl>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2287 The DB_File interface was written by Paul Marquess
2288 E<lt>pmqs@cpan.orgE<gt>.
2289 Questions about the DB system itself may be addressed to
2290 E<lt>db@sleepycat.comE<gt>.