1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (Paul.Marquess@btinternet.com)
4 # last modified 23rd Nov 2001
7 # Copyright (c) 1995-2001 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);
35 bless { VALID => { map {$_, 1}
36 qw( bsize ffactor nelem cachesize hash lorder)
48 return $self->{GOT}{$key} if exists $self->{VALID}{$key} ;
51 croak "${pkg}::FETCH - Unknown element '$key'" ;
61 if ( exists $self->{VALID}{$key} )
63 $self->{GOT}{$key} = $value ;
68 croak "${pkg}::STORE - Unknown element '$key'" ;
76 if ( exists $self->{VALID}{$key} )
78 delete $self->{GOT}{$key} ;
83 croak "DB_File::HASHINFO::DELETE - Unknown element '$key'" ;
91 exists $self->{VALID}{$key} ;
99 croak ref($self) . " does not define the method ${method}" ;
102 sub FIRSTKEY { my $self = shift ; $self->NotHere("FIRSTKEY") }
103 sub NEXTKEY { my $self = shift ; $self->NotHere("NEXTKEY") }
104 sub CLEAR { my $self = shift ; $self->NotHere("CLEAR") }
106 package DB_File::RECNOINFO ;
111 @DB_File::RECNOINFO::ISA = qw(DB_File::HASHINFO) ;
117 bless { VALID => { map {$_, 1}
118 qw( bval cachesize psize flags lorder reclen bfname )
124 package DB_File::BTREEINFO ;
129 @DB_File::BTREEINFO::ISA = qw(DB_File::HASHINFO) ;
135 bless { VALID => { map {$_, 1}
136 qw( flags cachesize maxkeypage minkeypage psize
137 compare prefix lorder )
148 our ($VERSION, @ISA, @EXPORT, $AUTOLOAD, $DB_BTREE, $DB_HASH, $DB_RECNO);
149 our ($db_version, $use_XSLoader);
155 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
156 $DB_BTREE = new DB_File::BTREEINFO ;
157 $DB_HASH = new DB_File::HASHINFO ;
158 $DB_RECNO = new DB_File::RECNOINFO ;
165 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
170 @ISA = qw(DynaLoader);
174 push @ISA, qw(Tie::Hash Exporter);
176 $DB_BTREE $DB_HASH $DB_RECNO
211 ($constname = $AUTOLOAD) =~ s/.*:://;
212 my ($error, $val) = constant($constname);
213 Carp::croak $error if $error;
215 *{$AUTOLOAD} = sub { $val };
221 # Make all Fcntl O_XXX constants available for importing
223 my @O = grep /^O_/, @Fcntl::EXPORT;
224 Fcntl->import(@O); # first we import what we want to export
229 { XSLoader::load("DB_File", $VERSION)}
231 { bootstrap DB_File $VERSION }
233 # Preloaded methods go here. Autoload methods go after __END__, and are
234 # processed by the autosplit program.
236 sub tie_hash_or_array
239 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
241 $arg[4] = tied %{ $arg[4] }
242 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
244 # make recno in Berkeley DB version 2 work like recno in version 1.
245 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
246 $arg[1] and ! -e $arg[1]) {
247 open(FH, ">$arg[1]") or return undef ;
249 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
252 DoTie_($tieHASH, @arg) ;
257 tie_hash_or_array(@_) ;
262 tie_hash_or_array(@_) ;
270 my $status = $self->seq($key, $value, R_FIRST());
273 while ($status == 0) {
275 $status = $self->seq($key, $value, R_NEXT());
277 foreach $key (reverse @keys) {
278 my $s = $self->del($key);
288 my $current_length = $self->length() ;
290 if ($length < $current_length) {
292 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
295 elsif ($length > $current_length) {
296 $self->put($length-1, "") ;
305 if (not defined $offset) {
306 carp 'Use of uninitialized value in splice';
310 my $length = @_ ? shift : 0;
311 # Carping about definedness comes _after_ the OFFSET sanity check.
312 # This is so we get the same error messages as Perl's splice().
317 my $size = $self->FETCHSIZE();
319 # 'If OFFSET is negative then it start that far from the end of
323 my $new_offset = $size + $offset;
324 if ($new_offset < 0) {
325 die "Modification of non-creatable array value attempted, "
326 . "subscript $offset";
328 $offset = $new_offset;
331 if ($offset > $size) {
335 if (not defined $length) {
336 carp 'Use of uninitialized value in splice';
340 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
341 if (not defined $length) {
342 $length = $size - $offset;
345 # 'If LENGTH is negative, leave that many elements off the end of
349 $length = $size - $offset + $length;
352 # The user must have specified a length bigger than the
353 # length of the array passed in. But perl's splice()
354 # doesn't catch this, it just behaves as for length=0.
360 if ($length > $size - $offset) {
361 $length = $size - $offset;
364 # $num_elems holds the current number of elements in the database.
365 my $num_elems = $size;
367 # 'Removes the elements designated by OFFSET and LENGTH from an
371 foreach (0 .. $length - 1) {
373 my $status = $self->get($offset, $old);
375 my $msg = "error from Berkeley DB on get($offset, \$old)";
377 $msg .= ' (no such element?)';
380 $msg .= ": error status $status";
381 if (defined $! and $! ne '') {
382 $msg .= ", message $!";
389 $status = $self->del($offset);
391 my $msg = "error from Berkeley DB on del($offset)";
393 $msg .= ' (no such element?)';
396 $msg .= ": error status $status";
397 if (defined $! and $! ne '') {
398 $msg .= ", message $!";
407 # ...'and replaces them with the elements of LIST, if any.'
409 while (defined (my $elem = shift @list)) {
412 if ($pos >= $num_elems) {
413 $status = $self->put($pos, $elem);
416 $status = $self->put($pos, $elem, $self->R_IBEFORE);
420 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
422 $msg .= ' (no such element?)';
425 $msg .= ", error status $status";
426 if (defined $! and $! ne '') {
427 $msg .= ", message $!";
433 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
441 # 'In list context, returns the elements removed from the
446 elsif (defined wantarray and not wantarray) {
447 # 'In scalar context, returns the last element removed, or
448 # undef if no elements are removed.'
451 my $last = pop @removed;
458 elsif (not defined wantarray) {
463 sub ::DB_File::splice { &SPLICE }
467 croak "Usage: \$db->find_dup(key,value)\n"
471 my ($origkey, $value_wanted) = @_ ;
472 my ($key, $value) = ($origkey, 0);
475 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
477 $status = $db->seq($key, $value, R_NEXT() ) ) {
479 return 0 if $key eq $origkey and $value eq $value_wanted ;
487 croak "Usage: \$db->del_dup(key,value)\n"
491 my ($key, $value) = @_ ;
492 my ($status) = $db->find_dup($key, $value) ;
493 return $status if $status != 0 ;
495 $status = $db->del($key, R_CURSOR() ) ;
501 croak "Usage: \$db->get_dup(key [,flag])\n"
502 unless @_ == 2 or @_ == 3 ;
509 my $wantarray = wantarray ;
515 # iterate through the database until either EOF ($status == 0)
516 # or a different key is encountered ($key ne $origkey).
517 for ($status = $db->seq($key, $value, R_CURSOR()) ;
518 $status == 0 and $key eq $origkey ;
519 $status = $db->seq($key, $value, R_NEXT()) ) {
521 # save the value or count number of matches
524 { ++ $values{$value} }
526 { push (@values, $value) }
533 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
542 DB_File - Perl5 access to Berkeley DB version 1.x
548 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
549 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
550 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
552 $status = $X->del($key [, $flags]) ;
553 $status = $X->put($key, $value [, $flags]) ;
554 $status = $X->get($key, $value [, $flags]) ;
555 $status = $X->seq($key, $value, $flags) ;
556 $status = $X->sync([$flags]) ;
560 $count = $X->get_dup($key) ;
561 @list = $X->get_dup($key) ;
562 %list = $X->get_dup($key, 1) ;
563 $status = $X->find_dup($key, $value) ;
564 $status = $X->del_dup($key, $value) ;
572 @r = $X->splice(offset, length, elements);
575 $old_filter = $db->filter_store_key ( sub { ... } ) ;
576 $old_filter = $db->filter_store_value( sub { ... } ) ;
577 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
578 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
585 B<DB_File> is a module which allows Perl programs to make use of the
586 facilities provided by Berkeley DB version 1.x (if you have a newer
587 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
588 It is assumed that you have a copy of the Berkeley DB manual pages at
589 hand when reading this documentation. The interface defined here
590 mirrors the Berkeley DB interface closely.
592 Berkeley DB is a C library which provides a consistent interface to a
593 number of database formats. B<DB_File> provides an interface to all
594 three of the database types currently supported by Berkeley DB.
602 This database type allows arbitrary key/value pairs to be stored in data
603 files. This is equivalent to the functionality provided by other
604 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
605 the files created using DB_HASH are not compatible with any of the
606 other packages mentioned.
608 A default hashing algorithm, which will be adequate for most
609 applications, is built into Berkeley DB. If you do need to use your own
610 hashing algorithm it is possible to write your own in Perl and have
611 B<DB_File> use it instead.
615 The btree format allows arbitrary key/value pairs to be stored in a
616 sorted, balanced binary tree.
618 As with the DB_HASH format, it is possible to provide a user defined
619 Perl routine to perform the comparison of keys. By default, though, the
620 keys are stored in lexical order.
624 DB_RECNO allows both fixed-length and variable-length flat text files
625 to be manipulated using the same key/value pair interface as in DB_HASH
626 and DB_BTREE. In this case the key will consist of a record (line)
631 =head2 Using DB_File with Berkeley DB version 2 or greater
633 Although B<DB_File> is intended to be used with Berkeley DB version 1,
634 it can also be used with version 2, 3 or 4. In this case the interface is
635 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
636 version 2 or greater interface differs, B<DB_File> arranges for it to work
637 like version 1. This feature allows B<DB_File> scripts that were built
638 with version 1 to be migrated to version 2 or greater without any changes.
640 If you want to make use of the new features available in Berkeley DB
641 2.x or greater, use the Perl module B<BerkeleyDB> instead.
643 B<Note:> The database file format has changed multiple times in Berkeley
644 DB version 2, 3 and 4. If you cannot recreate your databases, you
645 must dump any existing databases with either the C<db_dump> or the
646 C<db_dump185> utility that comes with Berkeley DB.
647 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
648 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
649 documentation for further details.
651 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
654 =head2 Interface to Berkeley DB
656 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
657 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
658 allows B<DB_File> to access Berkeley DB files using either an
659 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
660 array (for the DB_RECNO file type).
662 In addition to the tie() interface, it is also possible to access most
663 of the functions provided in the Berkeley DB API directly.
664 See L<THE API INTERFACE>.
666 =head2 Opening a Berkeley DB Database File
668 Berkeley DB uses the function dbopen() to open or create a database.
669 Here is the C prototype for dbopen():
672 dbopen (const char * file, int flags, int mode,
673 DBTYPE type, const void * openinfo)
675 The parameter C<type> is an enumeration which specifies which of the 3
676 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
677 Depending on which of these is actually chosen, the final parameter,
678 I<openinfo> points to a data structure which allows tailoring of the
679 specific interface method.
681 This interface is handled slightly differently in B<DB_File>. Here is
682 an equivalent call using B<DB_File>:
684 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
686 The C<filename>, C<flags> and C<mode> parameters are the direct
687 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
688 performs the function of both the C<type> and C<openinfo> parameters in
691 In the example above $DB_HASH is actually a pre-defined reference to a
692 hash object. B<DB_File> has three of these pre-defined references.
693 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
695 The keys allowed in each of these pre-defined references is limited to
696 the names used in the equivalent C structure. So, for example, the
697 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
698 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
700 To change one of these elements, just assign to it like this:
702 $DB_HASH->{'cachesize'} = 10000 ;
704 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
705 usually adequate for most applications. If you do need to create extra
706 instances of these objects, constructors are available for each file
709 Here are examples of the constructors and the valid options available
710 for DB_HASH, DB_BTREE and DB_RECNO respectively.
712 $a = new DB_File::HASHINFO ;
720 $b = new DB_File::BTREEINFO ;
730 $c = new DB_File::RECNOINFO ;
739 The values stored in the hashes above are mostly the direct equivalent
740 of their C counterpart. Like their C counterparts, all are set to a
741 default values - that means you don't have to set I<all> of the
742 values when you only want to change one. Here is an example:
744 $a = new DB_File::HASHINFO ;
745 $a->{'cachesize'} = 12345 ;
746 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
748 A few of the options need extra discussion here. When used, the C
749 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
750 to C functions. In B<DB_File> these keys are used to store references
751 to Perl subs. Below are templates for each of the subs:
757 # return the hash value for $data
763 my ($key, $key2) = @_ ;
765 # return 0 if $key1 eq $key2
766 # -1 if $key1 lt $key2
767 # 1 if $key1 gt $key2
768 return (-1 , 0 or 1) ;
773 my ($key, $key2) = @_ ;
775 # return number of bytes of $key2 which are
776 # necessary to determine that it is greater than $key1
780 See L<Changing the BTREE sort order> for an example of using the
783 If you are using the DB_RECNO interface and you intend making use of
784 C<bval>, you should check out L<The 'bval' Option>.
786 =head2 Default Parameters
788 It is possible to omit some or all of the final 4 parameters in the
789 call to C<tie> and let them take default values. As DB_HASH is the most
790 common file format used, the call:
792 tie %A, "DB_File", "filename" ;
796 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
798 It is also possible to omit the filename parameter as well, so the
805 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
807 See L<In Memory Databases> for a discussion on the use of C<undef>
808 in place of a filename.
810 =head2 In Memory Databases
812 Berkeley DB allows the creation of in-memory databases by using NULL
813 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
814 uses C<undef> instead of NULL to provide this functionality.
818 The DB_HASH file format is probably the most commonly used of the three
819 file formats that B<DB_File> supports. It is also very straightforward
822 =head2 A Simple Example
824 This example shows how to create a database, add key/value pairs to the
825 database, delete keys/value pairs and finally how to enumerate the
826 contents of the database.
834 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
835 or die "Cannot open file 'fruit': $!\n";
837 # Add a few key/value pairs to the file
838 $h{"apple"} = "red" ;
839 $h{"orange"} = "orange" ;
840 $h{"banana"} = "yellow" ;
841 $h{"tomato"} = "red" ;
843 # Check for existence of a key
844 print "Banana Exists\n\n" if $h{"banana"} ;
846 # Delete a key/value pair.
849 # print the contents of the file
850 while (($k, $v) = each %h)
851 { print "$k -> $v\n" }
863 Note that the like ordinary associative arrays, the order of the keys
864 retrieved is in an apparently random order.
868 The DB_BTREE format is useful when you want to store data in a given
869 order. By default the keys will be stored in lexical order, but as you
870 will see from the example shown in the next section, it is very easy to
871 define your own sorting function.
873 =head2 Changing the BTREE sort order
875 This script shows how to override the default sorting algorithm that
876 BTREE uses. Instead of using the normal lexical ordering, a case
877 insensitive compare function will be used.
887 my ($key1, $key2) = @_ ;
888 "\L$key1" cmp "\L$key2" ;
891 # specify the Perl sub that will do the comparison
892 $DB_BTREE->{'compare'} = \&Compare ;
895 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
896 or die "Cannot open file 'tree': $!\n" ;
898 # Add a key/value pair to the file
899 $h{'Wall'} = 'Larry' ;
900 $h{'Smith'} = 'John' ;
901 $h{'mouse'} = 'mickey' ;
902 $h{'duck'} = 'donald' ;
907 # Cycle through the keys printing them in order.
908 # Note it is not necessary to sort the keys as
909 # the btree will have kept them in order automatically.
915 Here is the output from the code above.
921 There are a few point to bear in mind if you want to change the
922 ordering in a BTREE database:
928 The new compare function must be specified when you create the database.
932 You cannot change the ordering once the database has been created. Thus
933 you must use the same compare function every time you access the
938 Duplicate keys are entirely defined by the comparison function.
939 In the case-insensitive example above, the keys: 'KEY' and 'key'
940 would be considered duplicates, and assigning to the second one
941 would overwrite the first. If duplicates are allowed for (with the
942 R_DUPS flag discussed below), only a single copy of duplicate keys
943 is stored in the database --- so (again with example above) assigning
944 three values to the keys: 'KEY', 'Key', and 'key' would leave just
945 the first key: 'KEY' in the database with three values. For some
946 situations this results in information loss, so care should be taken
947 to provide fully qualified comparison functions when necessary.
948 For example, the above comparison routine could be modified to
949 additionally compare case-sensitively if two keys are equal in the
950 case insensitive comparison:
953 my($key1, $key2) = @_;
954 lc $key1 cmp lc $key2 ||
958 And now you will only have duplicates when the keys themselves
959 are truly the same. (note: in versions of the db library prior to
960 about November 1996, such duplicate keys were retained so it was
961 possible to recover the original keys in sets of keys that
967 =head2 Handling Duplicate Keys
969 The BTREE file type optionally allows a single key to be associated
970 with an arbitrary number of values. This option is enabled by setting
971 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
973 There are some difficulties in using the tied hash interface if you
974 want to manipulate a BTREE database with duplicate keys. Consider this
981 our ($filename, %h) ;
986 # Enable duplicate records
987 $DB_BTREE->{'flags'} = R_DUP ;
989 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
990 or die "Cannot open $filename: $!\n";
992 # Add some key/value pairs to the file
993 $h{'Wall'} = 'Larry' ;
994 $h{'Wall'} = 'Brick' ; # Note the duplicate key
995 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
996 $h{'Smith'} = 'John' ;
997 $h{'mouse'} = 'mickey' ;
999 # iterate through the associative array
1000 # and print each key/value pair.
1001 foreach (sort keys %h)
1002 { print "$_ -> $h{$_}\n" }
1014 As you can see 3 records have been successfully created with key C<Wall>
1015 - the only thing is, when they are retrieved from the database they
1016 I<seem> to have the same value, namely C<Larry>. The problem is caused
1017 by the way that the associative array interface works. Basically, when
1018 the associative array interface is used to fetch the value associated
1019 with a given key, it will only ever retrieve the first value.
1021 Although it may not be immediately obvious from the code above, the
1022 associative array interface can be used to write values with duplicate
1023 keys, but it cannot be used to read them back from the database.
1025 The way to get around this problem is to use the Berkeley DB API method
1026 called C<seq>. This method allows sequential access to key/value
1027 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1028 and the API in general.
1030 Here is the script above rewritten using the C<seq> API method.
1036 our ($filename, $x, %h, $status, $key, $value) ;
1038 $filename = "tree" ;
1041 # Enable duplicate records
1042 $DB_BTREE->{'flags'} = R_DUP ;
1044 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1045 or die "Cannot open $filename: $!\n";
1047 # Add some key/value pairs to the file
1048 $h{'Wall'} = 'Larry' ;
1049 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1050 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1051 $h{'Smith'} = 'John' ;
1052 $h{'mouse'} = 'mickey' ;
1054 # iterate through the btree using seq
1055 # and print each key/value pair.
1057 for ($status = $x->seq($key, $value, R_FIRST) ;
1059 $status = $x->seq($key, $value, R_NEXT) )
1060 { print "$key -> $value\n" }
1073 This time we have got all the key/value pairs, including the multiple
1074 values associated with the key C<Wall>.
1076 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1077 a few utility methods.
1079 =head2 The get_dup() Method
1081 The C<get_dup> method assists in
1082 reading duplicate values from BTREE databases. The method can take the
1085 $count = $x->get_dup($key) ;
1086 @list = $x->get_dup($key) ;
1087 %list = $x->get_dup($key, 1) ;
1089 In a scalar context the method returns the number of values associated
1090 with the key, C<$key>.
1092 In list context, it returns all the values which match C<$key>. Note
1093 that the values will be returned in an apparently random order.
1095 In list context, if the second parameter is present and evaluates
1096 TRUE, the method returns an associative array. The keys of the
1097 associative array correspond to the values that matched in the BTREE
1098 and the values of the array are a count of the number of times that
1099 particular value occurred in the BTREE.
1101 So assuming the database created above, we can use C<get_dup> like
1108 our ($filename, $x, %h) ;
1110 $filename = "tree" ;
1112 # Enable duplicate records
1113 $DB_BTREE->{'flags'} = R_DUP ;
1115 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1116 or die "Cannot open $filename: $!\n";
1118 my $cnt = $x->get_dup("Wall") ;
1119 print "Wall occurred $cnt times\n" ;
1121 my %hash = $x->get_dup("Wall", 1) ;
1122 print "Larry is there\n" if $hash{'Larry'} ;
1123 print "There are $hash{'Brick'} Brick Walls\n" ;
1125 my @list = sort $x->get_dup("Wall") ;
1126 print "Wall => [@list]\n" ;
1128 @list = $x->get_dup("Smith") ;
1129 print "Smith => [@list]\n" ;
1131 @list = $x->get_dup("Dog") ;
1132 print "Dog => [@list]\n" ;
1137 Wall occurred 3 times
1139 There are 2 Brick Walls
1140 Wall => [Brick Brick Larry]
1144 =head2 The find_dup() Method
1146 $status = $X->find_dup($key, $value) ;
1148 This method checks for the existence of a specific key/value pair. If the
1149 pair exists, the cursor is left pointing to the pair and the method
1150 returns 0. Otherwise the method returns a non-zero value.
1152 Assuming the database from the previous example:
1158 our ($filename, $x, %h, $found) ;
1160 $filename = "tree" ;
1162 # Enable duplicate records
1163 $DB_BTREE->{'flags'} = R_DUP ;
1165 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1166 or die "Cannot open $filename: $!\n";
1168 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1169 print "Larry Wall is $found there\n" ;
1171 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1172 print "Harry Wall is $found there\n" ;
1180 Harry Wall is not there
1183 =head2 The del_dup() Method
1185 $status = $X->del_dup($key, $value) ;
1187 This method deletes a specific key/value pair. It returns
1188 0 if they exist and have been deleted successfully.
1189 Otherwise the method returns a non-zero value.
1191 Again assuming the existence of the C<tree> database
1197 our ($filename, $x, %h, $found) ;
1199 $filename = "tree" ;
1201 # Enable duplicate records
1202 $DB_BTREE->{'flags'} = R_DUP ;
1204 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1205 or die "Cannot open $filename: $!\n";
1207 $x->del_dup("Wall", "Larry") ;
1209 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1210 print "Larry Wall is $found there\n" ;
1217 Larry Wall is not there
1219 =head2 Matching Partial Keys
1221 The BTREE interface has a feature which allows partial keys to be
1222 matched. This functionality is I<only> available when the C<seq> method
1223 is used along with the R_CURSOR flag.
1225 $x->seq($key, $value, R_CURSOR) ;
1227 Here is the relevant quote from the dbopen man page where it defines
1228 the use of the R_CURSOR flag with seq:
1230 Note, for the DB_BTREE access method, the returned key is not
1231 necessarily an exact match for the specified key. The returned key
1232 is the smallest key greater than or equal to the specified key,
1233 permitting partial key matches and range searches.
1235 In the example script below, the C<match> sub uses this feature to find
1236 and print the first matching key/value pair given a partial key.
1243 our ($filename, $x, %h, $st, $key, $value) ;
1249 my $orig_key = $key ;
1250 $x->seq($key, $value, R_CURSOR) ;
1251 print "$orig_key\t-> $key\t-> $value\n" ;
1254 $filename = "tree" ;
1257 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1258 or die "Cannot open $filename: $!\n";
1260 # Add some key/value pairs to the file
1261 $h{'mouse'} = 'mickey' ;
1262 $h{'Wall'} = 'Larry' ;
1263 $h{'Walls'} = 'Brick' ;
1264 $h{'Smith'} = 'John' ;
1268 print "IN ORDER\n" ;
1269 for ($st = $x->seq($key, $value, R_FIRST) ;
1271 $st = $x->seq($key, $value, R_NEXT) )
1273 { print "$key -> $value\n" }
1275 print "\nPARTIAL MATCH\n" ;
1295 a -> mouse -> mickey
1299 DB_RECNO provides an interface to flat text files. Both variable and
1300 fixed length records are supported.
1302 In order to make RECNO more compatible with Perl, the array offset for
1303 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1305 As with normal Perl arrays, a RECNO array can be accessed using
1306 negative indexes. The index -1 refers to the last element of the array,
1307 -2 the second last, and so on. Attempting to access an element before
1308 the start of the array will raise a fatal run-time error.
1310 =head2 The 'bval' Option
1312 The operation of the bval option warrants some discussion. Here is the
1313 definition of bval from the Berkeley DB 1.85 recno manual page:
1315 The delimiting byte to be used to mark the end of a
1316 record for variable-length records, and the pad charac-
1317 ter for fixed-length records. If no value is speci-
1318 fied, newlines (``\n'') are used to mark the end of
1319 variable-length records and fixed-length records are
1322 The second sentence is wrong. In actual fact bval will only default to
1323 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1324 openinfo parameter is used at all, the value that happens to be in bval
1325 will be used. That means you always have to specify bval when making
1326 use of any of the options in the openinfo parameter. This documentation
1327 error will be fixed in the next release of Berkeley DB.
1329 That clarifies the situation with regards Berkeley DB itself. What
1330 about B<DB_File>? Well, the behavior defined in the quote above is
1331 quite useful, so B<DB_File> conforms to it.
1333 That means that you can specify other options (e.g. cachesize) and
1334 still have bval default to C<"\n"> for variable length records, and
1335 space for fixed length records.
1337 Also note that the bval option only allows you to specify a single byte
1340 =head2 A Simple Example
1342 Here is a simple example that uses RECNO (if you are using a version
1343 of Perl earlier than 5.004_57 this example won't work -- see
1344 L<Extra RECNO Methods> for a workaround).
1350 my $filename = "text" ;
1354 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1355 or die "Cannot open file 'text': $!\n" ;
1357 # Add a few key/value pairs to the file
1362 push @h, "green", "black" ;
1364 my $elements = scalar @h ;
1365 print "The array contains $elements entries\n" ;
1368 print "popped $last\n" ;
1370 unshift @h, "white" ;
1371 my $first = shift @h ;
1372 print "shifted $first\n" ;
1374 # Check for existence of a key
1375 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1377 # use a negative index
1378 print "The last element is $h[-1]\n" ;
1379 print "The 2nd last element is $h[-2]\n" ;
1383 Here is the output from the script:
1385 The array contains 5 entries
1388 Element 1 Exists with value blue
1389 The last element is green
1390 The 2nd last element is yellow
1392 =head2 Extra RECNO Methods
1394 If you are using a version of Perl earlier than 5.004_57, the tied
1395 array interface is quite limited. In the example script above
1396 C<push>, C<pop>, C<shift>, C<unshift>
1397 or determining the array length will not work with a tied array.
1399 To make the interface more useful for older versions of Perl, a number
1400 of methods are supplied with B<DB_File> to simulate the missing array
1401 operations. All these methods are accessed via the object returned from
1404 Here are the methods:
1408 =item B<$X-E<gt>push(list) ;>
1410 Pushes the elements of C<list> to the end of the array.
1412 =item B<$value = $X-E<gt>pop ;>
1414 Removes and returns the last element of the array.
1416 =item B<$X-E<gt>shift>
1418 Removes and returns the first element of the array.
1420 =item B<$X-E<gt>unshift(list) ;>
1422 Pushes the elements of C<list> to the start of the array.
1424 =item B<$X-E<gt>length>
1426 Returns the number of elements in the array.
1428 =item B<$X-E<gt>splice(offset, length, elements);>
1430 Returns a splice of the the array.
1434 =head2 Another Example
1436 Here is a more complete example that makes use of some of the methods
1437 described above. It also makes use of the API interface directly (see
1438 L<THE API INTERFACE>).
1442 our (@h, $H, $file, $i) ;
1450 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1451 or die "Cannot open file $file: $!\n" ;
1453 # first create a text file to play with
1461 # Print the records in order.
1463 # The length method is needed here because evaluating a tied
1464 # array in a scalar context does not return the number of
1465 # elements in the array.
1467 print "\nORIGINAL\n" ;
1468 foreach $i (0 .. $H->length - 1) {
1469 print "$i: $h[$i]\n" ;
1472 # use the push & pop methods
1475 print "\nThe last record was [$a]\n" ;
1477 # and the shift & unshift methods
1479 $H->unshift("first") ;
1480 print "The first record was [$a]\n" ;
1482 # Use the API to add a new record after record 2.
1484 $H->put($i, "Newbie", R_IAFTER) ;
1486 # and a new record before record 1.
1488 $H->put($i, "New One", R_IBEFORE) ;
1493 # now print the records in reverse order
1494 print "\nREVERSE\n" ;
1495 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1496 { print "$i: $h[$i]\n" }
1498 # same again, but use the API functions instead
1499 print "\nREVERSE again\n" ;
1500 my ($s, $k, $v) = (0, 0, 0) ;
1501 for ($s = $H->seq($k, $v, R_LAST) ;
1503 $s = $H->seq($k, $v, R_PREV))
1504 { print "$k: $v\n" }
1509 and this is what it outputs:
1518 The last record was [four]
1519 The first record was [zero]
1543 Rather than iterating through the array, C<@h> like this:
1547 it is necessary to use either this:
1549 foreach $i (0 .. $H->length - 1)
1553 for ($a = $H->get($k, $v, R_FIRST) ;
1555 $a = $H->get($k, $v, R_NEXT) )
1559 Notice that both times the C<put> method was used the record index was
1560 specified using a variable, C<$i>, rather than the literal value
1561 itself. This is because C<put> will return the record number of the
1562 inserted line via that parameter.
1566 =head1 THE API INTERFACE
1568 As well as accessing Berkeley DB using a tied hash or array, it is also
1569 possible to make direct use of most of the API functions defined in the
1570 Berkeley DB documentation.
1572 To do this you need to store a copy of the object returned from the tie.
1574 $db = tie %hash, "DB_File", "filename" ;
1576 Once you have done that, you can access the Berkeley DB API functions
1577 as B<DB_File> methods directly like this:
1579 $db->put($key, $value, R_NOOVERWRITE) ;
1581 B<Important:> If you have saved a copy of the object returned from
1582 C<tie>, the underlying database file will I<not> be closed until both
1583 the tied variable is untied and all copies of the saved object are
1587 $db = tie %hash, "DB_File", "filename"
1588 or die "Cannot tie filename: $!" ;
1593 See L<The untie() Gotcha> for more details.
1595 All the functions defined in L<dbopen> are available except for
1596 close() and dbopen() itself. The B<DB_File> method interface to the
1597 supported functions have been implemented to mirror the way Berkeley DB
1598 works whenever possible. In particular note that:
1604 The methods return a status value. All return 0 on success.
1605 All return -1 to signify an error and set C<$!> to the exact
1606 error code. The return code 1 generally (but not always) means that the
1607 key specified did not exist in the database.
1609 Other return codes are defined. See below and in the Berkeley DB
1610 documentation for details. The Berkeley DB documentation should be used
1611 as the definitive source.
1615 Whenever a Berkeley DB function returns data via one of its parameters,
1616 the equivalent B<DB_File> method does exactly the same.
1620 If you are careful, it is possible to mix API calls with the tied
1621 hash/array interface in the same piece of code. Although only a few of
1622 the methods used to implement the tied interface currently make use of
1623 the cursor, you should always assume that the cursor has been changed
1624 any time the tied hash/array interface is used. As an example, this
1625 code will probably not do what you expect:
1627 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1628 or die "Cannot tie $filename: $!" ;
1630 # Get the first key/value pair and set the cursor
1631 $X->seq($key, $value, R_FIRST) ;
1633 # this line will modify the cursor
1634 $count = scalar keys %x ;
1636 # Get the second key/value pair.
1637 # oops, it didn't, it got the last key/value pair!
1638 $X->seq($key, $value, R_NEXT) ;
1640 The code above can be rearranged to get around the problem, like this:
1642 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1643 or die "Cannot tie $filename: $!" ;
1645 # this line will modify the cursor
1646 $count = scalar keys %x ;
1648 # Get the first key/value pair and set the cursor
1649 $X->seq($key, $value, R_FIRST) ;
1651 # Get the second key/value pair.
1653 $X->seq($key, $value, R_NEXT) ;
1657 All the constants defined in L<dbopen> for use in the flags parameters
1658 in the methods defined below are also available. Refer to the Berkeley
1659 DB documentation for the precise meaning of the flags values.
1661 Below is a list of the methods available.
1665 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1667 Given a key (C<$key>) this method reads the value associated with it
1668 from the database. The value read from the database is returned in the
1669 C<$value> parameter.
1671 If the key does not exist the method returns 1.
1673 No flags are currently defined for this method.
1675 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1677 Stores the key/value pair in the database.
1679 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1680 will have the record number of the inserted key/value pair set.
1682 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1685 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1687 Removes all key/value pairs with key C<$key> from the database.
1689 A return code of 1 means that the requested key was not in the
1692 R_CURSOR is the only valid flag at present.
1694 =item B<$status = $X-E<gt>fd ;>
1696 Returns the file descriptor for the underlying database.
1698 See L<Locking: The Trouble with fd> for an explanation for why you should
1699 not use C<fd> to lock your database.
1701 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1703 This interface allows sequential retrieval from the database. See
1704 L<dbopen> for full details.
1706 Both the C<$key> and C<$value> parameters will be set to the key/value
1707 pair read from the database.
1709 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1710 R_FIRST, R_LAST, R_NEXT and R_PREV.
1712 =item B<$status = $X-E<gt>sync([$flags]) ;>
1714 Flushes any cached buffers to disk.
1716 R_RECNOSYNC is the only valid flag at present.
1722 A DBM Filter is a piece of code that is be used when you I<always>
1723 want to make the same transformation to all keys and/or values in a
1726 There are four methods associated with DBM Filters. All work identically,
1727 and each is used to install (or uninstall) a single DBM Filter. Each
1728 expects a single parameter, namely a reference to a sub. The only
1729 difference between them is the place that the filter is installed.
1735 =item B<filter_store_key>
1737 If a filter has been installed with this method, it will be invoked
1738 every time you write a key to a DBM database.
1740 =item B<filter_store_value>
1742 If a filter has been installed with this method, it will be invoked
1743 every time you write a value to a DBM database.
1746 =item B<filter_fetch_key>
1748 If a filter has been installed with this method, it will be invoked
1749 every time you read a key from a DBM database.
1751 =item B<filter_fetch_value>
1753 If a filter has been installed with this method, it will be invoked
1754 every time you read a value from a DBM database.
1758 You can use any combination of the methods, from none, to all four.
1760 All filter methods return the existing filter, if present, or C<undef>
1763 To delete a filter pass C<undef> to it.
1767 When each filter is called by Perl, a local copy of C<$_> will contain
1768 the key or value to be filtered. Filtering is achieved by modifying
1769 the contents of C<$_>. The return code from the filter is ignored.
1771 =head2 An Example -- the NULL termination problem.
1773 Consider the following scenario. You have a DBM database
1774 that you need to share with a third-party C application. The C application
1775 assumes that I<all> keys and values are NULL terminated. Unfortunately
1776 when Perl writes to DBM databases it doesn't use NULL termination, so
1777 your Perl application will have to manage NULL termination itself. When
1778 you write to the database you will have to use something like this:
1780 $hash{"$key\0"} = "$value\0" ;
1782 Similarly the NULL needs to be taken into account when you are considering
1783 the length of existing keys/values.
1785 It would be much better if you could ignore the NULL terminations issue
1786 in the main application code and have a mechanism that automatically
1787 added the terminating NULL to all keys and values whenever you write to
1788 the database and have them removed when you read from the database. As I'm
1789 sure you have already guessed, this is a problem that DBM Filters can
1797 my $filename = "/tmp/filt" ;
1800 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1801 or die "Cannot open $filename: $!\n" ;
1803 # Install DBM Filters
1804 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1805 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1806 $db->filter_fetch_value( sub { s/\0$// } ) ;
1807 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1809 $hash{"abc"} = "def" ;
1810 my $a = $hash{"ABC"} ;
1815 Hopefully the contents of each of the filters should be
1816 self-explanatory. Both "fetch" filters remove the terminating NULL,
1817 and both "store" filters add a terminating NULL.
1820 =head2 Another Example -- Key is a C int.
1822 Here is another real-life example. By default, whenever Perl writes to
1823 a DBM database it always writes the key and value as strings. So when
1826 $hash{12345} = "soemthing" ;
1828 the key 12345 will get stored in the DBM database as the 5 byte string
1829 "12345". If you actually want the key to be stored in the DBM database
1830 as a C int, you will have to use C<pack> when writing, and C<unpack>
1833 Here is a DBM Filter that does it:
1839 my $filename = "/tmp/filt" ;
1843 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1844 or die "Cannot open $filename: $!\n" ;
1846 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1847 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1848 $hash{123} = "def" ;
1853 This time only two filters have been used -- we only need to manipulate
1854 the contents of the key, so it wasn't necessary to install any value
1857 =head1 HINTS AND TIPS
1860 =head2 Locking: The Trouble with fd
1862 Until version 1.72 of this module, the recommended technique for locking
1863 B<DB_File> databases was to flock the filehandle returned from the "fd"
1864 function. Unfortunately this technique has been shown to be fundamentally
1865 flawed (Kudos to David Harris for tracking this down). Use it at your own
1868 The locking technique went like this.
1870 $db = tie(%db, 'DB_File', '/tmp/foo.db', O_CREAT|O_RDWR, 0666)
1871 || die "dbcreat /tmp/foo.db $!";
1873 open(DB_FH, "+<&=$fd") || die "dup $!";
1874 flock (DB_FH, LOCK_EX) || die "flock: $!";
1876 $db{"Tom"} = "Jerry" ;
1878 flock(DB_FH, LOCK_UN);
1883 In simple terms, this is what happens:
1889 Use "tie" to open the database.
1893 Lock the database with fd & flock.
1897 Read & Write to the database.
1901 Unlock and close the database.
1905 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1906 database in step 2 is that an initial block from the database will get
1907 read from disk and cached in memory.
1909 To see why this is a problem, consider what can happen when two processes,
1910 say "A" and "B", both want to update the same B<DB_File> database
1911 using the locking steps outlined above. Assume process "A" has already
1912 opened the database and has a write lock, but it hasn't actually updated
1913 the database yet (it has finished step 2, but not started step 3 yet). Now
1914 process "B" tries to open the same database - step 1 will succeed,
1915 but it will block on step 2 until process "A" releases the lock. The
1916 important thing to notice here is that at this point in time both
1917 processes will have cached identical initial blocks from the database.
1919 Now process "A" updates the database and happens to change some of the
1920 data held in the initial buffer. Process "A" terminates, flushing
1921 all cached data to disk and releasing the database lock. At this point
1922 the database on disk will correctly reflect the changes made by process
1925 With the lock released, process "B" can now continue. It also updates the
1926 database and unfortunately it too modifies the data that was in its
1927 initial buffer. Once that data gets flushed to disk it will overwrite
1928 some/all of the changes process "A" made to the database.
1930 The result of this scenario is at best a database that doesn't contain
1931 what you expect. At worst the database will corrupt.
1933 The above won't happen every time competing process update the same
1934 B<DB_File> database, but it does illustrate why the technique should
1937 =head2 Safe ways to lock a database
1939 Starting with version 2.x, Berkeley DB has internal support for locking.
1940 The companion module to this one, B<BerkeleyDB>, provides an interface
1941 to this locking functionality. If you are serious about locking
1942 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1944 If using B<BerkeleyDB> isn't an option, there are a number of modules
1945 available on CPAN that can be used to implement locking. Each one
1946 implements locking differently and has different goals in mind. It is
1947 therefore worth knowing the difference, so that you can pick the right
1948 one for your application. Here are the three locking wrappers:
1952 =item B<Tie::DB_Lock>
1954 A B<DB_File> wrapper which creates copies of the database file for
1955 read access, so that you have a kind of a multiversioning concurrent read
1956 system. However, updates are still serial. Use for databases where reads
1957 may be lengthy and consistency problems may occur.
1959 =item B<Tie::DB_LockFile>
1961 A B<DB_File> wrapper that has the ability to lock and unlock the database
1962 while it is being used. Avoids the tie-before-flock problem by simply
1963 re-tie-ing the database when you get or drop a lock. Because of the
1964 flexibility in dropping and re-acquiring the lock in the middle of a
1965 session, this can be massaged into a system that will work with long
1966 updates and/or reads if the application follows the hints in the POD
1969 =item B<DB_File::Lock>
1971 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
1972 before tie-ing the database and drops the lock after the untie. Allows
1973 one to use the same lockfile for multiple databases to avoid deadlock
1974 problems, if desired. Use for databases where updates are reads are
1975 quick and simple flock locking semantics are enough.
1979 =head2 Sharing Databases With C Applications
1981 There is no technical reason why a Berkeley DB database cannot be
1982 shared by both a Perl and a C application.
1984 The vast majority of problems that are reported in this area boil down
1985 to the fact that C strings are NULL terminated, whilst Perl strings are
1986 not. See L<DBM FILTERS> for a generic way to work around this problem.
1988 Here is a real example. Netscape 2.0 keeps a record of the locations you
1989 visit along with the time you last visited them in a DB_HASH database.
1990 This is usually stored in the file F<~/.netscape/history.db>. The key
1991 field in the database is the location string and the value field is the
1992 time the location was last visited stored as a 4 byte binary value.
1994 If you haven't already guessed, the location string is stored with a
1995 terminating NULL. This means you need to be careful when accessing the
1998 Here is a snippet of code that is loosely based on Tom Christiansen's
1999 I<ggh> script (available from your nearest CPAN archive in
2000 F<authors/id/TOMC/scripts/nshist.gz>).
2007 our ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2008 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2010 $HISTORY = "$dotdir/.netscape/history.db";
2012 tie %hist_db, 'DB_File', $HISTORY
2013 or die "Cannot open $HISTORY: $!\n" ;;
2015 # Dump the complete database
2016 while ( ($href, $binary_time) = each %hist_db ) {
2018 # remove the terminating NULL
2019 $href =~ s/\x00$// ;
2021 # convert the binary time into a user friendly string
2022 $date = localtime unpack("V", $binary_time);
2023 print "$date $href\n" ;
2026 # check for the existence of a specific key
2027 # remember to add the NULL
2028 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2029 $date = localtime unpack("V", $binary_time) ;
2030 print "Last visited mox.perl.com on $date\n" ;
2033 print "Never visited mox.perl.com\n"
2038 =head2 The untie() Gotcha
2040 If you make use of the Berkeley DB API, it is I<very> strongly
2041 recommended that you read L<perltie/The untie Gotcha>.
2043 Even if you don't currently make use of the API interface, it is still
2046 Here is an example which illustrates the problem from a B<DB_File>
2055 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2056 or die "Cannot tie first time: $!" ;
2062 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2063 or die "Cannot tie second time: $!" ;
2067 When run, the script will produce this error message:
2069 Cannot tie second time: Invalid argument at bad.file line 14.
2071 Although the error message above refers to the second tie() statement
2072 in the script, the source of the problem is really with the untie()
2073 statement that precedes it.
2075 Having read L<perltie> you will probably have already guessed that the
2076 error is caused by the extra copy of the tied object stored in C<$X>.
2077 If you haven't, then the problem boils down to the fact that the
2078 B<DB_File> destructor, DESTROY, will not be called until I<all>
2079 references to the tied object are destroyed. Both the tied variable,
2080 C<%x>, and C<$X> above hold a reference to the object. The call to
2081 untie() will destroy the first, but C<$X> still holds a valid
2082 reference, so the destructor will not get called and the database file
2083 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2084 attempt to open a database that is already open via the catch-all
2085 "Invalid argument" doesn't help.
2087 If you run the script with the C<-w> flag the error message becomes:
2089 untie attempted while 1 inner references still exist at bad.file line 12.
2090 Cannot tie second time: Invalid argument at bad.file line 14.
2092 which pinpoints the real problem. Finally the script can now be
2093 modified to fix the original problem by destroying the API object
2102 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2106 =head1 COMMON QUESTIONS
2108 =head2 Why is there Perl source in my database?
2110 If you look at the contents of a database file created by DB_File,
2111 there can sometimes be part of a Perl script included in it.
2113 This happens because Berkeley DB uses dynamic memory to allocate
2114 buffers which will subsequently be written to the database file. Being
2115 dynamic, the memory could have been used for anything before DB
2116 malloced it. As Berkeley DB doesn't clear the memory once it has been
2117 allocated, the unused portions will contain random junk. In the case
2118 where a Perl script gets written to the database, the random junk will
2119 correspond to an area of dynamic memory that happened to be used during
2120 the compilation of the script.
2122 Unless you don't like the possibility of there being part of your Perl
2123 scripts embedded in a database file, this is nothing to worry about.
2125 =head2 How do I store complex data structures with DB_File?
2127 Although B<DB_File> cannot do this directly, there is a module which
2128 can layer transparently over B<DB_File> to accomplish this feat.
2130 Check out the MLDBM module, available on CPAN in the directory
2131 F<modules/by-module/MLDBM>.
2133 =head2 What does "Invalid Argument" mean?
2135 You will get this error message when one of the parameters in the
2136 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2137 get wrong, so it can be difficult to figure out which one it is.
2139 Here are a couple of possibilities:
2145 Attempting to reopen a database without closing it.
2149 Using the O_WRONLY flag.
2153 =head2 What does "Bareword 'DB_File' not allowed" mean?
2155 You will encounter this particular error message when you have the
2156 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2157 Consider this script:
2163 tie %x, DB_File, "filename" ;
2165 Running it produces the error in question:
2167 Bareword "DB_File" not allowed while "strict subs" in use
2169 To get around the error, place the word C<DB_File> in either single or
2170 double quotes, like this:
2172 tie %x, "DB_File", "filename" ;
2174 Although it might seem like a real pain, it is really worth the effort
2175 of having a C<use strict> in all your scripts.
2179 Articles that are either about B<DB_File> or make use of it.
2185 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2186 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2192 Moved to the Changes file.
2196 Some older versions of Berkeley DB had problems with fixed length
2197 records using the RECNO file format. This problem has been fixed since
2198 version 1.85 of Berkeley DB.
2200 I am sure there are bugs in the code. If you do find any, or can
2201 suggest any enhancements, I would welcome your comments.
2205 B<DB_File> comes with the standard Perl source distribution. Look in
2206 the directory F<ext/DB_File>. Given the amount of time between releases
2207 of Perl the version that ships with Perl is quite likely to be out of
2208 date, so the most recent version can always be found on CPAN (see
2209 L<perlmod/CPAN> for details), in the directory
2210 F<modules/by-module/DB_File>.
2212 This version of B<DB_File> will work with either version 1.x, 2.x or
2213 3.x of Berkeley DB, but is limited to the functionality provided by
2216 The official web site for Berkeley DB is F<http://www.sleepycat.com>.
2217 All versions of Berkeley DB are available there.
2219 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2220 archive in F<src/misc/db.1.85.tar.gz>.
2222 If you are running IRIX, then get Berkeley DB version 1 from
2223 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2224 compile properly on IRIX 5.3.
2228 Copyright (c) 1995-2001 Paul Marquess. All rights reserved. This program
2229 is free software; you can redistribute it and/or modify it under the
2230 same terms as Perl itself.
2232 Although B<DB_File> is covered by the Perl license, the library it
2233 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2234 copyright and its own license. Please take the time to read it.
2236 Here are are few words taken from the Berkeley DB FAQ (at
2237 F<http://www.sleepycat.com>) regarding the license:
2239 Do I have to license DB to use it in Perl scripts?
2241 No. The Berkeley DB license requires that software that uses
2242 Berkeley DB be freely redistributable. In the case of Perl, that
2243 software is Perl, and not your scripts. Any Perl scripts that you
2244 write are your property, including scripts that make use of
2245 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2246 place any restriction on what you may do with them.
2248 If you are in any doubt about the license situation, contact either the
2249 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2254 L<perl(1)>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2259 The DB_File interface was written by Paul Marquess
2260 E<lt>Paul.Marquess@btinternet.comE<gt>.
2261 Questions about the DB system itself may be addressed to
2262 E<lt>db@sleepycat.com<gt>.