1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (Paul.Marquess@btinternet.com)
4 # last modified 1st March 2002
7 # Copyright (c) 1995-2002 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, $splice_end_array);
156 local $SIG{__WARN__} = sub {$splice_end_array = "@_";};
157 my @a =(1); splice(@a, 3);
159 ($splice_end_array =~ /^splice\(\) offset past end of array at /);
162 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
163 $DB_BTREE = new DB_File::BTREEINFO ;
164 $DB_HASH = new DB_File::HASHINFO ;
165 $DB_RECNO = new DB_File::RECNOINFO ;
172 { local $SIG{__DIE__} ; eval { require XSLoader } ; }
177 @ISA = qw(DynaLoader);
181 push @ISA, qw(Tie::Hash Exporter);
183 $DB_BTREE $DB_HASH $DB_RECNO
218 ($constname = $AUTOLOAD) =~ s/.*:://;
219 my ($error, $val) = constant($constname);
220 Carp::croak $error if $error;
222 *{$AUTOLOAD} = sub { $val };
228 # Make all Fcntl O_XXX constants available for importing
230 my @O = grep /^O_/, @Fcntl::EXPORT;
231 Fcntl->import(@O); # first we import what we want to export
236 { XSLoader::load("DB_File", $VERSION)}
238 { bootstrap DB_File $VERSION }
240 # Preloaded methods go here. Autoload methods go after __END__, and are
241 # processed by the autosplit program.
243 sub tie_hash_or_array
246 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
248 $arg[4] = tied %{ $arg[4] }
249 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
251 # make recno in Berkeley DB version 2 work like recno in version 1.
252 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
253 $arg[1] and ! -e $arg[1]) {
254 open(FH, ">$arg[1]") or return undef ;
256 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
259 DoTie_($tieHASH, @arg) ;
264 tie_hash_or_array(@_) ;
269 tie_hash_or_array(@_) ;
277 my $status = $self->seq($key, $value, R_FIRST());
280 while ($status == 0) {
282 $status = $self->seq($key, $value, R_NEXT());
284 foreach $key (reverse @keys) {
285 my $s = $self->del($key);
295 my $current_length = $self->length() ;
297 if ($length < $current_length) {
299 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
302 elsif ($length > $current_length) {
303 $self->put($length-1, "") ;
312 if (not defined $offset) {
313 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
317 my $length = @_ ? shift : 0;
318 # Carping about definedness comes _after_ the OFFSET sanity check.
319 # This is so we get the same error messages as Perl's splice().
324 my $size = $self->FETCHSIZE();
326 # 'If OFFSET is negative then it start that far from the end of
330 my $new_offset = $size + $offset;
331 if ($new_offset < 0) {
332 die "Modification of non-creatable array value attempted, "
333 . "subscript $offset";
335 $offset = $new_offset;
338 if (not defined $length) {
339 warnings::warnif('uninitialized', 'Use of uninitialized value in splice');
343 if ($offset > $size) {
345 warnings::warnif('misc', 'splice() offset past end of array')
346 if $splice_end_array;
349 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
350 if (not defined $length) {
351 $length = $size - $offset;
354 # 'If LENGTH is negative, leave that many elements off the end of
358 $length = $size - $offset + $length;
361 # The user must have specified a length bigger than the
362 # length of the array passed in. But perl's splice()
363 # doesn't catch this, it just behaves as for length=0.
369 if ($length > $size - $offset) {
370 $length = $size - $offset;
373 # $num_elems holds the current number of elements in the database.
374 my $num_elems = $size;
376 # 'Removes the elements designated by OFFSET and LENGTH from an
380 foreach (0 .. $length - 1) {
382 my $status = $self->get($offset, $old);
384 my $msg = "error from Berkeley DB on get($offset, \$old)";
386 $msg .= ' (no such element?)';
389 $msg .= ": error status $status";
390 if (defined $! and $! ne '') {
391 $msg .= ", message $!";
398 $status = $self->del($offset);
400 my $msg = "error from Berkeley DB on del($offset)";
402 $msg .= ' (no such element?)';
405 $msg .= ": error status $status";
406 if (defined $! and $! ne '') {
407 $msg .= ", message $!";
416 # ...'and replaces them with the elements of LIST, if any.'
418 while (defined (my $elem = shift @list)) {
421 if ($pos >= $num_elems) {
422 $status = $self->put($pos, $elem);
425 $status = $self->put($pos, $elem, $self->R_IBEFORE);
429 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
431 $msg .= ' (no such element?)';
434 $msg .= ", error status $status";
435 if (defined $! and $! ne '') {
436 $msg .= ", message $!";
442 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
450 # 'In list context, returns the elements removed from the
455 elsif (defined wantarray and not wantarray) {
456 # 'In scalar context, returns the last element removed, or
457 # undef if no elements are removed.'
460 my $last = pop @removed;
467 elsif (not defined wantarray) {
472 sub ::DB_File::splice { &SPLICE }
476 croak "Usage: \$db->find_dup(key,value)\n"
480 my ($origkey, $value_wanted) = @_ ;
481 my ($key, $value) = ($origkey, 0);
484 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
486 $status = $db->seq($key, $value, R_NEXT() ) ) {
488 return 0 if $key eq $origkey and $value eq $value_wanted ;
496 croak "Usage: \$db->del_dup(key,value)\n"
500 my ($key, $value) = @_ ;
501 my ($status) = $db->find_dup($key, $value) ;
502 return $status if $status != 0 ;
504 $status = $db->del($key, R_CURSOR() ) ;
510 croak "Usage: \$db->get_dup(key [,flag])\n"
511 unless @_ == 2 or @_ == 3 ;
518 my $wantarray = wantarray ;
524 # iterate through the database until either EOF ($status == 0)
525 # or a different key is encountered ($key ne $origkey).
526 for ($status = $db->seq($key, $value, R_CURSOR()) ;
527 $status == 0 and $key eq $origkey ;
528 $status = $db->seq($key, $value, R_NEXT()) ) {
530 # save the value or count number of matches
533 { ++ $values{$value} }
535 { push (@values, $value) }
542 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
551 DB_File - Perl5 access to Berkeley DB version 1.x
557 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
558 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
559 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
561 $status = $X->del($key [, $flags]) ;
562 $status = $X->put($key, $value [, $flags]) ;
563 $status = $X->get($key, $value [, $flags]) ;
564 $status = $X->seq($key, $value, $flags) ;
565 $status = $X->sync([$flags]) ;
569 $count = $X->get_dup($key) ;
570 @list = $X->get_dup($key) ;
571 %list = $X->get_dup($key, 1) ;
572 $status = $X->find_dup($key, $value) ;
573 $status = $X->del_dup($key, $value) ;
581 @r = $X->splice(offset, length, elements);
584 $old_filter = $db->filter_store_key ( sub { ... } ) ;
585 $old_filter = $db->filter_store_value( sub { ... } ) ;
586 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
587 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
594 B<DB_File> is a module which allows Perl programs to make use of the
595 facilities provided by Berkeley DB version 1.x (if you have a newer
596 version of DB, see L<Using DB_File with Berkeley DB version 2 or greater>).
597 It is assumed that you have a copy of the Berkeley DB manual pages at
598 hand when reading this documentation. The interface defined here
599 mirrors the Berkeley DB interface closely.
601 Berkeley DB is a C library which provides a consistent interface to a
602 number of database formats. B<DB_File> provides an interface to all
603 three of the database types currently supported by Berkeley DB.
611 This database type allows arbitrary key/value pairs to be stored in data
612 files. This is equivalent to the functionality provided by other
613 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
614 the files created using DB_HASH are not compatible with any of the
615 other packages mentioned.
617 A default hashing algorithm, which will be adequate for most
618 applications, is built into Berkeley DB. If you do need to use your own
619 hashing algorithm it is possible to write your own in Perl and have
620 B<DB_File> use it instead.
624 The btree format allows arbitrary key/value pairs to be stored in a
625 sorted, balanced binary tree.
627 As with the DB_HASH format, it is possible to provide a user defined
628 Perl routine to perform the comparison of keys. By default, though, the
629 keys are stored in lexical order.
633 DB_RECNO allows both fixed-length and variable-length flat text files
634 to be manipulated using the same key/value pair interface as in DB_HASH
635 and DB_BTREE. In this case the key will consist of a record (line)
640 =head2 Using DB_File with Berkeley DB version 2 or greater
642 Although B<DB_File> is intended to be used with Berkeley DB version 1,
643 it can also be used with version 2, 3 or 4. In this case the interface is
644 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
645 version 2 or greater interface differs, B<DB_File> arranges for it to work
646 like version 1. This feature allows B<DB_File> scripts that were built
647 with version 1 to be migrated to version 2 or greater without any changes.
649 If you want to make use of the new features available in Berkeley DB
650 2.x or greater, use the Perl module B<BerkeleyDB> instead.
652 B<Note:> The database file format has changed multiple times in Berkeley
653 DB version 2, 3 and 4. If you cannot recreate your databases, you
654 must dump any existing databases with either the C<db_dump> or the
655 C<db_dump185> utility that comes with Berkeley DB.
656 Once you have rebuilt DB_File to use Berkeley DB version 2 or greater,
657 your databases can be recreated using C<db_load>. Refer to the Berkeley DB
658 documentation for further details.
660 Please read L<"COPYRIGHT"> before using version 2.x or greater of Berkeley
663 =head2 Interface to Berkeley DB
665 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
666 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
667 allows B<DB_File> to access Berkeley DB files using either an
668 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
669 array (for the DB_RECNO file type).
671 In addition to the tie() interface, it is also possible to access most
672 of the functions provided in the Berkeley DB API directly.
673 See L<THE API INTERFACE>.
675 =head2 Opening a Berkeley DB Database File
677 Berkeley DB uses the function dbopen() to open or create a database.
678 Here is the C prototype for dbopen():
681 dbopen (const char * file, int flags, int mode,
682 DBTYPE type, const void * openinfo)
684 The parameter C<type> is an enumeration which specifies which of the 3
685 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
686 Depending on which of these is actually chosen, the final parameter,
687 I<openinfo> points to a data structure which allows tailoring of the
688 specific interface method.
690 This interface is handled slightly differently in B<DB_File>. Here is
691 an equivalent call using B<DB_File>:
693 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
695 The C<filename>, C<flags> and C<mode> parameters are the direct
696 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
697 performs the function of both the C<type> and C<openinfo> parameters in
700 In the example above $DB_HASH is actually a pre-defined reference to a
701 hash object. B<DB_File> has three of these pre-defined references.
702 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
704 The keys allowed in each of these pre-defined references is limited to
705 the names used in the equivalent C structure. So, for example, the
706 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
707 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
709 To change one of these elements, just assign to it like this:
711 $DB_HASH->{'cachesize'} = 10000 ;
713 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
714 usually adequate for most applications. If you do need to create extra
715 instances of these objects, constructors are available for each file
718 Here are examples of the constructors and the valid options available
719 for DB_HASH, DB_BTREE and DB_RECNO respectively.
721 $a = new DB_File::HASHINFO ;
729 $b = new DB_File::BTREEINFO ;
739 $c = new DB_File::RECNOINFO ;
748 The values stored in the hashes above are mostly the direct equivalent
749 of their C counterpart. Like their C counterparts, all are set to a
750 default values - that means you don't have to set I<all> of the
751 values when you only want to change one. Here is an example:
753 $a = new DB_File::HASHINFO ;
754 $a->{'cachesize'} = 12345 ;
755 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
757 A few of the options need extra discussion here. When used, the C
758 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
759 to C functions. In B<DB_File> these keys are used to store references
760 to Perl subs. Below are templates for each of the subs:
766 # return the hash value for $data
772 my ($key, $key2) = @_ ;
774 # return 0 if $key1 eq $key2
775 # -1 if $key1 lt $key2
776 # 1 if $key1 gt $key2
777 return (-1 , 0 or 1) ;
782 my ($key, $key2) = @_ ;
784 # return number of bytes of $key2 which are
785 # necessary to determine that it is greater than $key1
789 See L<Changing the BTREE sort order> for an example of using the
792 If you are using the DB_RECNO interface and you intend making use of
793 C<bval>, you should check out L<The 'bval' Option>.
795 =head2 Default Parameters
797 It is possible to omit some or all of the final 4 parameters in the
798 call to C<tie> and let them take default values. As DB_HASH is the most
799 common file format used, the call:
801 tie %A, "DB_File", "filename" ;
805 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
807 It is also possible to omit the filename parameter as well, so the
814 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
816 See L<In Memory Databases> for a discussion on the use of C<undef>
817 in place of a filename.
819 =head2 In Memory Databases
821 Berkeley DB allows the creation of in-memory databases by using NULL
822 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
823 uses C<undef> instead of NULL to provide this functionality.
827 The DB_HASH file format is probably the most commonly used of the three
828 file formats that B<DB_File> supports. It is also very straightforward
831 =head2 A Simple Example
833 This example shows how to create a database, add key/value pairs to the
834 database, delete keys/value pairs and finally how to enumerate the
835 contents of the database.
843 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
844 or die "Cannot open file 'fruit': $!\n";
846 # Add a few key/value pairs to the file
847 $h{"apple"} = "red" ;
848 $h{"orange"} = "orange" ;
849 $h{"banana"} = "yellow" ;
850 $h{"tomato"} = "red" ;
852 # Check for existence of a key
853 print "Banana Exists\n\n" if $h{"banana"} ;
855 # Delete a key/value pair.
858 # print the contents of the file
859 while (($k, $v) = each %h)
860 { print "$k -> $v\n" }
872 Note that the like ordinary associative arrays, the order of the keys
873 retrieved is in an apparently random order.
877 The DB_BTREE format is useful when you want to store data in a given
878 order. By default the keys will be stored in lexical order, but as you
879 will see from the example shown in the next section, it is very easy to
880 define your own sorting function.
882 =head2 Changing the BTREE sort order
884 This script shows how to override the default sorting algorithm that
885 BTREE uses. Instead of using the normal lexical ordering, a case
886 insensitive compare function will be used.
896 my ($key1, $key2) = @_ ;
897 "\L$key1" cmp "\L$key2" ;
900 # specify the Perl sub that will do the comparison
901 $DB_BTREE->{'compare'} = \&Compare ;
904 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
905 or die "Cannot open file 'tree': $!\n" ;
907 # Add a key/value pair to the file
908 $h{'Wall'} = 'Larry' ;
909 $h{'Smith'} = 'John' ;
910 $h{'mouse'} = 'mickey' ;
911 $h{'duck'} = 'donald' ;
916 # Cycle through the keys printing them in order.
917 # Note it is not necessary to sort the keys as
918 # the btree will have kept them in order automatically.
924 Here is the output from the code above.
930 There are a few point to bear in mind if you want to change the
931 ordering in a BTREE database:
937 The new compare function must be specified when you create the database.
941 You cannot change the ordering once the database has been created. Thus
942 you must use the same compare function every time you access the
947 Duplicate keys are entirely defined by the comparison function.
948 In the case-insensitive example above, the keys: 'KEY' and 'key'
949 would be considered duplicates, and assigning to the second one
950 would overwrite the first. If duplicates are allowed for (with the
951 R_DUPS flag discussed below), only a single copy of duplicate keys
952 is stored in the database --- so (again with example above) assigning
953 three values to the keys: 'KEY', 'Key', and 'key' would leave just
954 the first key: 'KEY' in the database with three values. For some
955 situations this results in information loss, so care should be taken
956 to provide fully qualified comparison functions when necessary.
957 For example, the above comparison routine could be modified to
958 additionally compare case-sensitively if two keys are equal in the
959 case insensitive comparison:
962 my($key1, $key2) = @_;
963 lc $key1 cmp lc $key2 ||
967 And now you will only have duplicates when the keys themselves
968 are truly the same. (note: in versions of the db library prior to
969 about November 1996, such duplicate keys were retained so it was
970 possible to recover the original keys in sets of keys that
976 =head2 Handling Duplicate Keys
978 The BTREE file type optionally allows a single key to be associated
979 with an arbitrary number of values. This option is enabled by setting
980 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
982 There are some difficulties in using the tied hash interface if you
983 want to manipulate a BTREE database with duplicate keys. Consider this
995 # Enable duplicate records
996 $DB_BTREE->{'flags'} = R_DUP ;
998 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
999 or die "Cannot open $filename: $!\n";
1001 # Add some key/value pairs to the file
1002 $h{'Wall'} = 'Larry' ;
1003 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1004 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1005 $h{'Smith'} = 'John' ;
1006 $h{'mouse'} = 'mickey' ;
1008 # iterate through the associative array
1009 # and print each key/value pair.
1010 foreach (sort keys %h)
1011 { print "$_ -> $h{$_}\n" }
1023 As you can see 3 records have been successfully created with key C<Wall>
1024 - the only thing is, when they are retrieved from the database they
1025 I<seem> to have the same value, namely C<Larry>. The problem is caused
1026 by the way that the associative array interface works. Basically, when
1027 the associative array interface is used to fetch the value associated
1028 with a given key, it will only ever retrieve the first value.
1030 Although it may not be immediately obvious from the code above, the
1031 associative array interface can be used to write values with duplicate
1032 keys, but it cannot be used to read them back from the database.
1034 The way to get around this problem is to use the Berkeley DB API method
1035 called C<seq>. This method allows sequential access to key/value
1036 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1037 and the API in general.
1039 Here is the script above rewritten using the C<seq> API method.
1045 my ($filename, $x, %h, $status, $key, $value) ;
1047 $filename = "tree" ;
1050 # Enable duplicate records
1051 $DB_BTREE->{'flags'} = R_DUP ;
1053 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1054 or die "Cannot open $filename: $!\n";
1056 # Add some key/value pairs to the file
1057 $h{'Wall'} = 'Larry' ;
1058 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1059 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1060 $h{'Smith'} = 'John' ;
1061 $h{'mouse'} = 'mickey' ;
1063 # iterate through the btree using seq
1064 # and print each key/value pair.
1066 for ($status = $x->seq($key, $value, R_FIRST) ;
1068 $status = $x->seq($key, $value, R_NEXT) )
1069 { print "$key -> $value\n" }
1082 This time we have got all the key/value pairs, including the multiple
1083 values associated with the key C<Wall>.
1085 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1086 a few utility methods.
1088 =head2 The get_dup() Method
1090 The C<get_dup> method assists in
1091 reading duplicate values from BTREE databases. The method can take the
1094 $count = $x->get_dup($key) ;
1095 @list = $x->get_dup($key) ;
1096 %list = $x->get_dup($key, 1) ;
1098 In a scalar context the method returns the number of values associated
1099 with the key, C<$key>.
1101 In list context, it returns all the values which match C<$key>. Note
1102 that the values will be returned in an apparently random order.
1104 In list context, if the second parameter is present and evaluates
1105 TRUE, the method returns an associative array. The keys of the
1106 associative array correspond to the values that matched in the BTREE
1107 and the values of the array are a count of the number of times that
1108 particular value occurred in the BTREE.
1110 So assuming the database created above, we can use C<get_dup> like
1117 my ($filename, $x, %h) ;
1119 $filename = "tree" ;
1121 # Enable duplicate records
1122 $DB_BTREE->{'flags'} = R_DUP ;
1124 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1125 or die "Cannot open $filename: $!\n";
1127 my $cnt = $x->get_dup("Wall") ;
1128 print "Wall occurred $cnt times\n" ;
1130 my %hash = $x->get_dup("Wall", 1) ;
1131 print "Larry is there\n" if $hash{'Larry'} ;
1132 print "There are $hash{'Brick'} Brick Walls\n" ;
1134 my @list = sort $x->get_dup("Wall") ;
1135 print "Wall => [@list]\n" ;
1137 @list = $x->get_dup("Smith") ;
1138 print "Smith => [@list]\n" ;
1140 @list = $x->get_dup("Dog") ;
1141 print "Dog => [@list]\n" ;
1146 Wall occurred 3 times
1148 There are 2 Brick Walls
1149 Wall => [Brick Brick Larry]
1153 =head2 The find_dup() Method
1155 $status = $X->find_dup($key, $value) ;
1157 This method checks for the existence of a specific key/value pair. If the
1158 pair exists, the cursor is left pointing to the pair and the method
1159 returns 0. Otherwise the method returns a non-zero value.
1161 Assuming the database from the previous example:
1167 my ($filename, $x, %h, $found) ;
1169 $filename = "tree" ;
1171 # Enable duplicate records
1172 $DB_BTREE->{'flags'} = R_DUP ;
1174 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1175 or die "Cannot open $filename: $!\n";
1177 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1178 print "Larry Wall is $found there\n" ;
1180 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1181 print "Harry Wall is $found there\n" ;
1189 Harry Wall is not there
1192 =head2 The del_dup() Method
1194 $status = $X->del_dup($key, $value) ;
1196 This method deletes a specific key/value pair. It returns
1197 0 if they exist and have been deleted successfully.
1198 Otherwise the method returns a non-zero value.
1200 Again assuming the existence of the C<tree> database
1206 my ($filename, $x, %h, $found) ;
1208 $filename = "tree" ;
1210 # Enable duplicate records
1211 $DB_BTREE->{'flags'} = R_DUP ;
1213 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1214 or die "Cannot open $filename: $!\n";
1216 $x->del_dup("Wall", "Larry") ;
1218 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1219 print "Larry Wall is $found there\n" ;
1226 Larry Wall is not there
1228 =head2 Matching Partial Keys
1230 The BTREE interface has a feature which allows partial keys to be
1231 matched. This functionality is I<only> available when the C<seq> method
1232 is used along with the R_CURSOR flag.
1234 $x->seq($key, $value, R_CURSOR) ;
1236 Here is the relevant quote from the dbopen man page where it defines
1237 the use of the R_CURSOR flag with seq:
1239 Note, for the DB_BTREE access method, the returned key is not
1240 necessarily an exact match for the specified key. The returned key
1241 is the smallest key greater than or equal to the specified key,
1242 permitting partial key matches and range searches.
1244 In the example script below, the C<match> sub uses this feature to find
1245 and print the first matching key/value pair given a partial key.
1252 my ($filename, $x, %h, $st, $key, $value) ;
1258 my $orig_key = $key ;
1259 $x->seq($key, $value, R_CURSOR) ;
1260 print "$orig_key\t-> $key\t-> $value\n" ;
1263 $filename = "tree" ;
1266 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1267 or die "Cannot open $filename: $!\n";
1269 # Add some key/value pairs to the file
1270 $h{'mouse'} = 'mickey' ;
1271 $h{'Wall'} = 'Larry' ;
1272 $h{'Walls'} = 'Brick' ;
1273 $h{'Smith'} = 'John' ;
1277 print "IN ORDER\n" ;
1278 for ($st = $x->seq($key, $value, R_FIRST) ;
1280 $st = $x->seq($key, $value, R_NEXT) )
1282 { print "$key -> $value\n" }
1284 print "\nPARTIAL MATCH\n" ;
1304 a -> mouse -> mickey
1308 DB_RECNO provides an interface to flat text files. Both variable and
1309 fixed length records are supported.
1311 In order to make RECNO more compatible with Perl, the array offset for
1312 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1314 As with normal Perl arrays, a RECNO array can be accessed using
1315 negative indexes. The index -1 refers to the last element of the array,
1316 -2 the second last, and so on. Attempting to access an element before
1317 the start of the array will raise a fatal run-time error.
1319 =head2 The 'bval' Option
1321 The operation of the bval option warrants some discussion. Here is the
1322 definition of bval from the Berkeley DB 1.85 recno manual page:
1324 The delimiting byte to be used to mark the end of a
1325 record for variable-length records, and the pad charac-
1326 ter for fixed-length records. If no value is speci-
1327 fied, newlines (``\n'') are used to mark the end of
1328 variable-length records and fixed-length records are
1331 The second sentence is wrong. In actual fact bval will only default to
1332 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1333 openinfo parameter is used at all, the value that happens to be in bval
1334 will be used. That means you always have to specify bval when making
1335 use of any of the options in the openinfo parameter. This documentation
1336 error will be fixed in the next release of Berkeley DB.
1338 That clarifies the situation with regards Berkeley DB itself. What
1339 about B<DB_File>? Well, the behavior defined in the quote above is
1340 quite useful, so B<DB_File> conforms to it.
1342 That means that you can specify other options (e.g. cachesize) and
1343 still have bval default to C<"\n"> for variable length records, and
1344 space for fixed length records.
1346 Also note that the bval option only allows you to specify a single byte
1349 =head2 A Simple Example
1351 Here is a simple example that uses RECNO (if you are using a version
1352 of Perl earlier than 5.004_57 this example won't work -- see
1353 L<Extra RECNO Methods> for a workaround).
1359 my $filename = "text" ;
1363 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1364 or die "Cannot open file 'text': $!\n" ;
1366 # Add a few key/value pairs to the file
1371 push @h, "green", "black" ;
1373 my $elements = scalar @h ;
1374 print "The array contains $elements entries\n" ;
1377 print "popped $last\n" ;
1379 unshift @h, "white" ;
1380 my $first = shift @h ;
1381 print "shifted $first\n" ;
1383 # Check for existence of a key
1384 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1386 # use a negative index
1387 print "The last element is $h[-1]\n" ;
1388 print "The 2nd last element is $h[-2]\n" ;
1392 Here is the output from the script:
1394 The array contains 5 entries
1397 Element 1 Exists with value blue
1398 The last element is green
1399 The 2nd last element is yellow
1401 =head2 Extra RECNO Methods
1403 If you are using a version of Perl earlier than 5.004_57, the tied
1404 array interface is quite limited. In the example script above
1405 C<push>, C<pop>, C<shift>, C<unshift>
1406 or determining the array length will not work with a tied array.
1408 To make the interface more useful for older versions of Perl, a number
1409 of methods are supplied with B<DB_File> to simulate the missing array
1410 operations. All these methods are accessed via the object returned from
1413 Here are the methods:
1417 =item B<$X-E<gt>push(list) ;>
1419 Pushes the elements of C<list> to the end of the array.
1421 =item B<$value = $X-E<gt>pop ;>
1423 Removes and returns the last element of the array.
1425 =item B<$X-E<gt>shift>
1427 Removes and returns the first element of the array.
1429 =item B<$X-E<gt>unshift(list) ;>
1431 Pushes the elements of C<list> to the start of the array.
1433 =item B<$X-E<gt>length>
1435 Returns the number of elements in the array.
1437 =item B<$X-E<gt>splice(offset, length, elements);>
1439 Returns a splice of the the array.
1443 =head2 Another Example
1445 Here is a more complete example that makes use of some of the methods
1446 described above. It also makes use of the API interface directly (see
1447 L<THE API INTERFACE>).
1451 my (@h, $H, $file, $i) ;
1459 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1460 or die "Cannot open file $file: $!\n" ;
1462 # first create a text file to play with
1470 # Print the records in order.
1472 # The length method is needed here because evaluating a tied
1473 # array in a scalar context does not return the number of
1474 # elements in the array.
1476 print "\nORIGINAL\n" ;
1477 foreach $i (0 .. $H->length - 1) {
1478 print "$i: $h[$i]\n" ;
1481 # use the push & pop methods
1484 print "\nThe last record was [$a]\n" ;
1486 # and the shift & unshift methods
1488 $H->unshift("first") ;
1489 print "The first record was [$a]\n" ;
1491 # Use the API to add a new record after record 2.
1493 $H->put($i, "Newbie", R_IAFTER) ;
1495 # and a new record before record 1.
1497 $H->put($i, "New One", R_IBEFORE) ;
1502 # now print the records in reverse order
1503 print "\nREVERSE\n" ;
1504 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1505 { print "$i: $h[$i]\n" }
1507 # same again, but use the API functions instead
1508 print "\nREVERSE again\n" ;
1509 my ($s, $k, $v) = (0, 0, 0) ;
1510 for ($s = $H->seq($k, $v, R_LAST) ;
1512 $s = $H->seq($k, $v, R_PREV))
1513 { print "$k: $v\n" }
1518 and this is what it outputs:
1527 The last record was [four]
1528 The first record was [zero]
1552 Rather than iterating through the array, C<@h> like this:
1556 it is necessary to use either this:
1558 foreach $i (0 .. $H->length - 1)
1562 for ($a = $H->get($k, $v, R_FIRST) ;
1564 $a = $H->get($k, $v, R_NEXT) )
1568 Notice that both times the C<put> method was used the record index was
1569 specified using a variable, C<$i>, rather than the literal value
1570 itself. This is because C<put> will return the record number of the
1571 inserted line via that parameter.
1575 =head1 THE API INTERFACE
1577 As well as accessing Berkeley DB using a tied hash or array, it is also
1578 possible to make direct use of most of the API functions defined in the
1579 Berkeley DB documentation.
1581 To do this you need to store a copy of the object returned from the tie.
1583 $db = tie %hash, "DB_File", "filename" ;
1585 Once you have done that, you can access the Berkeley DB API functions
1586 as B<DB_File> methods directly like this:
1588 $db->put($key, $value, R_NOOVERWRITE) ;
1590 B<Important:> If you have saved a copy of the object returned from
1591 C<tie>, the underlying database file will I<not> be closed until both
1592 the tied variable is untied and all copies of the saved object are
1596 $db = tie %hash, "DB_File", "filename"
1597 or die "Cannot tie filename: $!" ;
1602 See L<The untie() Gotcha> for more details.
1604 All the functions defined in L<dbopen> are available except for
1605 close() and dbopen() itself. The B<DB_File> method interface to the
1606 supported functions have been implemented to mirror the way Berkeley DB
1607 works whenever possible. In particular note that:
1613 The methods return a status value. All return 0 on success.
1614 All return -1 to signify an error and set C<$!> to the exact
1615 error code. The return code 1 generally (but not always) means that the
1616 key specified did not exist in the database.
1618 Other return codes are defined. See below and in the Berkeley DB
1619 documentation for details. The Berkeley DB documentation should be used
1620 as the definitive source.
1624 Whenever a Berkeley DB function returns data via one of its parameters,
1625 the equivalent B<DB_File> method does exactly the same.
1629 If you are careful, it is possible to mix API calls with the tied
1630 hash/array interface in the same piece of code. Although only a few of
1631 the methods used to implement the tied interface currently make use of
1632 the cursor, you should always assume that the cursor has been changed
1633 any time the tied hash/array interface is used. As an example, this
1634 code will probably not do what you expect:
1636 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1637 or die "Cannot tie $filename: $!" ;
1639 # Get the first key/value pair and set the cursor
1640 $X->seq($key, $value, R_FIRST) ;
1642 # this line will modify the cursor
1643 $count = scalar keys %x ;
1645 # Get the second key/value pair.
1646 # oops, it didn't, it got the last key/value pair!
1647 $X->seq($key, $value, R_NEXT) ;
1649 The code above can be rearranged to get around the problem, like this:
1651 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1652 or die "Cannot tie $filename: $!" ;
1654 # this line will modify the cursor
1655 $count = scalar keys %x ;
1657 # Get the first key/value pair and set the cursor
1658 $X->seq($key, $value, R_FIRST) ;
1660 # Get the second key/value pair.
1662 $X->seq($key, $value, R_NEXT) ;
1666 All the constants defined in L<dbopen> for use in the flags parameters
1667 in the methods defined below are also available. Refer to the Berkeley
1668 DB documentation for the precise meaning of the flags values.
1670 Below is a list of the methods available.
1674 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1676 Given a key (C<$key>) this method reads the value associated with it
1677 from the database. The value read from the database is returned in the
1678 C<$value> parameter.
1680 If the key does not exist the method returns 1.
1682 No flags are currently defined for this method.
1684 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1686 Stores the key/value pair in the database.
1688 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1689 will have the record number of the inserted key/value pair set.
1691 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1694 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1696 Removes all key/value pairs with key C<$key> from the database.
1698 A return code of 1 means that the requested key was not in the
1701 R_CURSOR is the only valid flag at present.
1703 =item B<$status = $X-E<gt>fd ;>
1705 Returns the file descriptor for the underlying database.
1707 See L<Locking: The Trouble with fd> for an explanation for why you should
1708 not use C<fd> to lock your database.
1710 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1712 This interface allows sequential retrieval from the database. See
1713 L<dbopen> for full details.
1715 Both the C<$key> and C<$value> parameters will be set to the key/value
1716 pair read from the database.
1718 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1719 R_FIRST, R_LAST, R_NEXT and R_PREV.
1721 =item B<$status = $X-E<gt>sync([$flags]) ;>
1723 Flushes any cached buffers to disk.
1725 R_RECNOSYNC is the only valid flag at present.
1731 A DBM Filter is a piece of code that is be used when you I<always>
1732 want to make the same transformation to all keys and/or values in a
1735 There are four methods associated with DBM Filters. All work identically,
1736 and each is used to install (or uninstall) a single DBM Filter. Each
1737 expects a single parameter, namely a reference to a sub. The only
1738 difference between them is the place that the filter is installed.
1744 =item B<filter_store_key>
1746 If a filter has been installed with this method, it will be invoked
1747 every time you write a key to a DBM database.
1749 =item B<filter_store_value>
1751 If a filter has been installed with this method, it will be invoked
1752 every time you write a value to a DBM database.
1755 =item B<filter_fetch_key>
1757 If a filter has been installed with this method, it will be invoked
1758 every time you read a key from a DBM database.
1760 =item B<filter_fetch_value>
1762 If a filter has been installed with this method, it will be invoked
1763 every time you read a value from a DBM database.
1767 You can use any combination of the methods, from none, to all four.
1769 All filter methods return the existing filter, if present, or C<undef>
1772 To delete a filter pass C<undef> to it.
1776 When each filter is called by Perl, a local copy of C<$_> will contain
1777 the key or value to be filtered. Filtering is achieved by modifying
1778 the contents of C<$_>. The return code from the filter is ignored.
1780 =head2 An Example -- the NULL termination problem.
1782 Consider the following scenario. You have a DBM database
1783 that you need to share with a third-party C application. The C application
1784 assumes that I<all> keys and values are NULL terminated. Unfortunately
1785 when Perl writes to DBM databases it doesn't use NULL termination, so
1786 your Perl application will have to manage NULL termination itself. When
1787 you write to the database you will have to use something like this:
1789 $hash{"$key\0"} = "$value\0" ;
1791 Similarly the NULL needs to be taken into account when you are considering
1792 the length of existing keys/values.
1794 It would be much better if you could ignore the NULL terminations issue
1795 in the main application code and have a mechanism that automatically
1796 added the terminating NULL to all keys and values whenever you write to
1797 the database and have them removed when you read from the database. As I'm
1798 sure you have already guessed, this is a problem that DBM Filters can
1806 my $filename = "/tmp/filt" ;
1809 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1810 or die "Cannot open $filename: $!\n" ;
1812 # Install DBM Filters
1813 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1814 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1815 $db->filter_fetch_value( sub { s/\0$// } ) ;
1816 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1818 $hash{"abc"} = "def" ;
1819 my $a = $hash{"ABC"} ;
1824 Hopefully the contents of each of the filters should be
1825 self-explanatory. Both "fetch" filters remove the terminating NULL,
1826 and both "store" filters add a terminating NULL.
1829 =head2 Another Example -- Key is a C int.
1831 Here is another real-life example. By default, whenever Perl writes to
1832 a DBM database it always writes the key and value as strings. So when
1835 $hash{12345} = "soemthing" ;
1837 the key 12345 will get stored in the DBM database as the 5 byte string
1838 "12345". If you actually want the key to be stored in the DBM database
1839 as a C int, you will have to use C<pack> when writing, and C<unpack>
1842 Here is a DBM Filter that does it:
1848 my $filename = "/tmp/filt" ;
1852 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1853 or die "Cannot open $filename: $!\n" ;
1855 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1856 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1857 $hash{123} = "def" ;
1862 This time only two filters have been used -- we only need to manipulate
1863 the contents of the key, so it wasn't necessary to install any value
1866 =head1 HINTS AND TIPS
1869 =head2 Locking: The Trouble with fd
1871 Until version 1.72 of this module, the recommended technique for locking
1872 B<DB_File> databases was to flock the filehandle returned from the "fd"
1873 function. Unfortunately this technique has been shown to be fundamentally
1874 flawed (Kudos to David Harris for tracking this down). Use it at your own
1877 The locking technique went like this.
1879 $db = tie(%db, 'DB_File', '/tmp/foo.db', O_CREAT|O_RDWR, 0666)
1880 || die "dbcreat /tmp/foo.db $!";
1882 open(DB_FH, "+<&=$fd") || die "dup $!";
1883 flock (DB_FH, LOCK_EX) || die "flock: $!";
1885 $db{"Tom"} = "Jerry" ;
1887 flock(DB_FH, LOCK_UN);
1892 In simple terms, this is what happens:
1898 Use "tie" to open the database.
1902 Lock the database with fd & flock.
1906 Read & Write to the database.
1910 Unlock and close the database.
1914 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1915 database in step 2 is that an initial block from the database will get
1916 read from disk and cached in memory.
1918 To see why this is a problem, consider what can happen when two processes,
1919 say "A" and "B", both want to update the same B<DB_File> database
1920 using the locking steps outlined above. Assume process "A" has already
1921 opened the database and has a write lock, but it hasn't actually updated
1922 the database yet (it has finished step 2, but not started step 3 yet). Now
1923 process "B" tries to open the same database - step 1 will succeed,
1924 but it will block on step 2 until process "A" releases the lock. The
1925 important thing to notice here is that at this point in time both
1926 processes will have cached identical initial blocks from the database.
1928 Now process "A" updates the database and happens to change some of the
1929 data held in the initial buffer. Process "A" terminates, flushing
1930 all cached data to disk and releasing the database lock. At this point
1931 the database on disk will correctly reflect the changes made by process
1934 With the lock released, process "B" can now continue. It also updates the
1935 database and unfortunately it too modifies the data that was in its
1936 initial buffer. Once that data gets flushed to disk it will overwrite
1937 some/all of the changes process "A" made to the database.
1939 The result of this scenario is at best a database that doesn't contain
1940 what you expect. At worst the database will corrupt.
1942 The above won't happen every time competing process update the same
1943 B<DB_File> database, but it does illustrate why the technique should
1946 =head2 Safe ways to lock a database
1948 Starting with version 2.x, Berkeley DB has internal support for locking.
1949 The companion module to this one, B<BerkeleyDB>, provides an interface
1950 to this locking functionality. If you are serious about locking
1951 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1953 If using B<BerkeleyDB> isn't an option, there are a number of modules
1954 available on CPAN that can be used to implement locking. Each one
1955 implements locking differently and has different goals in mind. It is
1956 therefore worth knowing the difference, so that you can pick the right
1957 one for your application. Here are the three locking wrappers:
1961 =item B<Tie::DB_Lock>
1963 A B<DB_File> wrapper which creates copies of the database file for
1964 read access, so that you have a kind of a multiversioning concurrent read
1965 system. However, updates are still serial. Use for databases where reads
1966 may be lengthy and consistency problems may occur.
1968 =item B<Tie::DB_LockFile>
1970 A B<DB_File> wrapper that has the ability to lock and unlock the database
1971 while it is being used. Avoids the tie-before-flock problem by simply
1972 re-tie-ing the database when you get or drop a lock. Because of the
1973 flexibility in dropping and re-acquiring the lock in the middle of a
1974 session, this can be massaged into a system that will work with long
1975 updates and/or reads if the application follows the hints in the POD
1978 =item B<DB_File::Lock>
1980 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
1981 before tie-ing the database and drops the lock after the untie. Allows
1982 one to use the same lockfile for multiple databases to avoid deadlock
1983 problems, if desired. Use for databases where updates are reads are
1984 quick and simple flock locking semantics are enough.
1988 =head2 Sharing Databases With C Applications
1990 There is no technical reason why a Berkeley DB database cannot be
1991 shared by both a Perl and a C application.
1993 The vast majority of problems that are reported in this area boil down
1994 to the fact that C strings are NULL terminated, whilst Perl strings are
1995 not. See L<DBM FILTERS> for a generic way to work around this problem.
1997 Here is a real example. Netscape 2.0 keeps a record of the locations you
1998 visit along with the time you last visited them in a DB_HASH database.
1999 This is usually stored in the file F<~/.netscape/history.db>. The key
2000 field in the database is the location string and the value field is the
2001 time the location was last visited stored as a 4 byte binary value.
2003 If you haven't already guessed, the location string is stored with a
2004 terminating NULL. This means you need to be careful when accessing the
2007 Here is a snippet of code that is loosely based on Tom Christiansen's
2008 I<ggh> script (available from your nearest CPAN archive in
2009 F<authors/id/TOMC/scripts/nshist.gz>).
2016 my ($dotdir, $HISTORY, %hist_db, $href, $binary_time, $date) ;
2017 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2019 $HISTORY = "$dotdir/.netscape/history.db";
2021 tie %hist_db, 'DB_File', $HISTORY
2022 or die "Cannot open $HISTORY: $!\n" ;;
2024 # Dump the complete database
2025 while ( ($href, $binary_time) = each %hist_db ) {
2027 # remove the terminating NULL
2028 $href =~ s/\x00$// ;
2030 # convert the binary time into a user friendly string
2031 $date = localtime unpack("V", $binary_time);
2032 print "$date $href\n" ;
2035 # check for the existence of a specific key
2036 # remember to add the NULL
2037 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2038 $date = localtime unpack("V", $binary_time) ;
2039 print "Last visited mox.perl.com on $date\n" ;
2042 print "Never visited mox.perl.com\n"
2047 =head2 The untie() Gotcha
2049 If you make use of the Berkeley DB API, it is I<very> strongly
2050 recommended that you read L<perltie/The untie Gotcha>.
2052 Even if you don't currently make use of the API interface, it is still
2055 Here is an example which illustrates the problem from a B<DB_File>
2064 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2065 or die "Cannot tie first time: $!" ;
2071 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2072 or die "Cannot tie second time: $!" ;
2076 When run, the script will produce this error message:
2078 Cannot tie second time: Invalid argument at bad.file line 14.
2080 Although the error message above refers to the second tie() statement
2081 in the script, the source of the problem is really with the untie()
2082 statement that precedes it.
2084 Having read L<perltie> you will probably have already guessed that the
2085 error is caused by the extra copy of the tied object stored in C<$X>.
2086 If you haven't, then the problem boils down to the fact that the
2087 B<DB_File> destructor, DESTROY, will not be called until I<all>
2088 references to the tied object are destroyed. Both the tied variable,
2089 C<%x>, and C<$X> above hold a reference to the object. The call to
2090 untie() will destroy the first, but C<$X> still holds a valid
2091 reference, so the destructor will not get called and the database file
2092 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2093 attempt to open a database that is already open via the catch-all
2094 "Invalid argument" doesn't help.
2096 If you run the script with the C<-w> flag the error message becomes:
2098 untie attempted while 1 inner references still exist at bad.file line 12.
2099 Cannot tie second time: Invalid argument at bad.file line 14.
2101 which pinpoints the real problem. Finally the script can now be
2102 modified to fix the original problem by destroying the API object
2111 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2115 =head1 COMMON QUESTIONS
2117 =head2 Why is there Perl source in my database?
2119 If you look at the contents of a database file created by DB_File,
2120 there can sometimes be part of a Perl script included in it.
2122 This happens because Berkeley DB uses dynamic memory to allocate
2123 buffers which will subsequently be written to the database file. Being
2124 dynamic, the memory could have been used for anything before DB
2125 malloced it. As Berkeley DB doesn't clear the memory once it has been
2126 allocated, the unused portions will contain random junk. In the case
2127 where a Perl script gets written to the database, the random junk will
2128 correspond to an area of dynamic memory that happened to be used during
2129 the compilation of the script.
2131 Unless you don't like the possibility of there being part of your Perl
2132 scripts embedded in a database file, this is nothing to worry about.
2134 =head2 How do I store complex data structures with DB_File?
2136 Although B<DB_File> cannot do this directly, there is a module which
2137 can layer transparently over B<DB_File> to accomplish this feat.
2139 Check out the MLDBM module, available on CPAN in the directory
2140 F<modules/by-module/MLDBM>.
2142 =head2 What does "Invalid Argument" mean?
2144 You will get this error message when one of the parameters in the
2145 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2146 get wrong, so it can be difficult to figure out which one it is.
2148 Here are a couple of possibilities:
2154 Attempting to reopen a database without closing it.
2158 Using the O_WRONLY flag.
2162 =head2 What does "Bareword 'DB_File' not allowed" mean?
2164 You will encounter this particular error message when you have the
2165 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2166 Consider this script:
2172 tie %x, DB_File, "filename" ;
2174 Running it produces the error in question:
2176 Bareword "DB_File" not allowed while "strict subs" in use
2178 To get around the error, place the word C<DB_File> in either single or
2179 double quotes, like this:
2181 tie %x, "DB_File", "filename" ;
2183 Although it might seem like a real pain, it is really worth the effort
2184 of having a C<use strict> in all your scripts.
2188 Articles that are either about B<DB_File> or make use of it.
2194 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2195 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2201 Moved to the Changes file.
2205 Some older versions of Berkeley DB had problems with fixed length
2206 records using the RECNO file format. This problem has been fixed since
2207 version 1.85 of Berkeley DB.
2209 I am sure there are bugs in the code. If you do find any, or can
2210 suggest any enhancements, I would welcome your comments.
2214 B<DB_File> comes with the standard Perl source distribution. Look in
2215 the directory F<ext/DB_File>. Given the amount of time between releases
2216 of Perl the version that ships with Perl is quite likely to be out of
2217 date, so the most recent version can always be found on CPAN (see
2218 L<perlmod/CPAN> for details), in the directory
2219 F<modules/by-module/DB_File>.
2221 This version of B<DB_File> will work with either version 1.x, 2.x or
2222 3.x of Berkeley DB, but is limited to the functionality provided by
2225 The official web site for Berkeley DB is F<http://www.sleepycat.com>.
2226 All versions of Berkeley DB are available there.
2228 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2229 archive in F<src/misc/db.1.85.tar.gz>.
2231 If you are running IRIX, then get Berkeley DB version 1 from
2232 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2233 compile properly on IRIX 5.3.
2237 Copyright (c) 1995-2002 Paul Marquess. All rights reserved. This program
2238 is free software; you can redistribute it and/or modify it under the
2239 same terms as Perl itself.
2241 Although B<DB_File> is covered by the Perl license, the library it
2242 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2243 copyright and its own license. Please take the time to read it.
2245 Here are are few words taken from the Berkeley DB FAQ (at
2246 F<http://www.sleepycat.com>) regarding the license:
2248 Do I have to license DB to use it in Perl scripts?
2250 No. The Berkeley DB license requires that software that uses
2251 Berkeley DB be freely redistributable. In the case of Perl, that
2252 software is Perl, and not your scripts. Any Perl scripts that you
2253 write are your property, including scripts that make use of
2254 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2255 place any restriction on what you may do with them.
2257 If you are in any doubt about the license situation, contact either the
2258 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2263 L<perl(1)>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2268 The DB_File interface was written by Paul Marquess
2269 E<lt>Paul.Marquess@btinternet.comE<gt>.
2270 Questions about the DB system itself may be addressed to
2271 E<lt>db@sleepycat.com<gt>.