1 # DB_File.pm -- Perl 5 interface to Berkeley DB
3 # written by Paul Marquess (Paul.Marquess@btinternet.com)
4 # last modified 30th July 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 {$_, 1}
136 qw( flags cachesize maxkeypage minkeypage psize
137 compare prefix lorder )
148 use vars qw($VERSION @ISA @EXPORT $AUTOLOAD $DB_BTREE $DB_HASH $DB_RECNO
149 $db_version $use_XSLoader
156 #typedef enum { DB_BTREE, DB_HASH, DB_RECNO } DBTYPE;
157 $DB_BTREE = new DB_File::BTREEINFO ;
158 $DB_HASH = new DB_File::HASHINFO ;
159 $DB_RECNO = new DB_File::RECNOINFO ;
166 eval { require XSLoader } ;
171 @ISA = qw(DynaLoader);
175 push @ISA, qw(Tie::Hash Exporter);
177 $DB_BTREE $DB_HASH $DB_RECNO
212 ($constname = $AUTOLOAD) =~ s/.*:://;
213 my $val = constant($constname, @_ ? $_[0] : 0);
215 if ($! =~ /Invalid/ || $!{EINVAL}) {
216 $AutoLoader::AUTOLOAD = $AUTOLOAD;
217 goto &AutoLoader::AUTOLOAD;
220 my($pack,$file,$line) = caller;
221 croak "Your vendor has not defined DB macro $constname, used at $file line $line.
225 eval "sub $AUTOLOAD { $val }";
231 # Make all Fcntl O_XXX constants available for importing
233 my @O = grep /^O_/, @Fcntl::EXPORT;
234 Fcntl->import(@O); # first we import what we want to export
239 { XSLoader::load("DB_File", $VERSION)}
241 { bootstrap DB_File $VERSION }
243 # Preloaded methods go here. Autoload methods go after __END__, and are
244 # processed by the autosplit program.
246 sub tie_hash_or_array
249 my $tieHASH = ( (caller(1))[3] =~ /TIEHASH/ ) ;
251 $arg[4] = tied %{ $arg[4] }
252 if @arg >= 5 && ref $arg[4] && $arg[4] =~ /=HASH/ && tied %{ $arg[4] } ;
254 # make recno in Berkeley DB version 2 work like recno in version 1.
255 if ($db_version > 1 and defined $arg[4] and $arg[4] =~ /RECNO/ and
256 $arg[1] and ! -e $arg[1]) {
257 open(FH, ">$arg[1]") or return undef ;
259 chmod $arg[3] ? $arg[3] : 0666 , $arg[1] ;
262 DoTie_($tieHASH, @arg) ;
267 tie_hash_or_array(@_) ;
272 tie_hash_or_array(@_) ;
280 my $status = $self->seq($key, $value, R_FIRST());
283 while ($status == 0) {
285 $status = $self->seq($key, $value, R_NEXT());
287 foreach $key (reverse @keys) {
288 my $s = $self->del($key);
298 my $current_length = $self->length() ;
300 if ($length < $current_length) {
302 for ($key = $current_length - 1 ; $key >= $length ; -- $key)
305 elsif ($length > $current_length) {
306 $self->put($length-1, "") ;
315 if (not defined $offset) {
316 carp 'Use of uninitialized value in splice';
320 my $length = @_ ? shift : 0;
321 # Carping about definedness comes _after_ the OFFSET sanity check.
322 # This is so we get the same error messages as Perl's splice().
327 my $size = $self->FETCHSIZE();
329 # 'If OFFSET is negative then it start that far from the end of
333 my $new_offset = $size + $offset;
334 if ($new_offset < 0) {
335 die "Modification of non-creatable array value attempted, "
336 . "subscript $offset";
338 $offset = $new_offset;
341 if ($offset > $size) {
345 if (not defined $length) {
346 carp 'Use of uninitialized value in splice';
350 # 'If LENGTH is omitted, removes everything from OFFSET onward.'
351 if (not defined $length) {
352 $length = $size - $offset;
355 # 'If LENGTH is negative, leave that many elements off the end of
359 $length = $size - $offset + $length;
362 # The user must have specified a length bigger than the
363 # length of the array passed in. But perl's splice()
364 # doesn't catch this, it just behaves as for length=0.
370 if ($length > $size - $offset) {
371 $length = $size - $offset;
374 # $num_elems holds the current number of elements in the database.
375 my $num_elems = $size;
377 # 'Removes the elements designated by OFFSET and LENGTH from an
381 foreach (0 .. $length - 1) {
383 my $status = $self->get($offset, $old);
385 my $msg = "error from Berkeley DB on get($offset, \$old)";
387 $msg .= ' (no such element?)';
390 $msg .= ": error status $status";
391 if (defined $! and $! ne '') {
392 $msg .= ", message $!";
399 $status = $self->del($offset);
401 my $msg = "error from Berkeley DB on del($offset)";
403 $msg .= ' (no such element?)';
406 $msg .= ": error status $status";
407 if (defined $! and $! ne '') {
408 $msg .= ", message $!";
417 # ...'and replaces them with the elements of LIST, if any.'
419 while (defined (my $elem = shift @list)) {
422 if ($pos >= $num_elems) {
423 $status = $self->put($pos, $elem);
426 $status = $self->put($pos, $elem, $self->R_IBEFORE);
430 my $msg = "error from Berkeley DB on put($pos, $elem, ...)";
432 $msg .= ' (no such element?)';
435 $msg .= ", error status $status";
436 if (defined $! and $! ne '') {
437 $msg .= ", message $!";
443 die "pos unexpectedly changed from $old_pos to $pos with R_IBEFORE"
451 # 'In list context, returns the elements removed from the
456 elsif (defined wantarray and not wantarray) {
457 # 'In scalar context, returns the last element removed, or
458 # undef if no elements are removed.'
461 my $last = pop @removed;
468 elsif (not defined wantarray) {
473 sub ::DB_File::splice { &SPLICE }
477 croak "Usage: \$db->find_dup(key,value)\n"
481 my ($origkey, $value_wanted) = @_ ;
482 my ($key, $value) = ($origkey, 0);
485 for ($status = $db->seq($key, $value, R_CURSOR() ) ;
487 $status = $db->seq($key, $value, R_NEXT() ) ) {
489 return 0 if $key eq $origkey and $value eq $value_wanted ;
497 croak "Usage: \$db->del_dup(key,value)\n"
501 my ($key, $value) = @_ ;
502 my ($status) = $db->find_dup($key, $value) ;
503 return $status if $status != 0 ;
505 $status = $db->del($key, R_CURSOR() ) ;
511 croak "Usage: \$db->get_dup(key [,flag])\n"
512 unless @_ == 2 or @_ == 3 ;
519 my $wantarray = wantarray ;
525 # iterate through the database until either EOF ($status == 0)
526 # or a different key is encountered ($key ne $origkey).
527 for ($status = $db->seq($key, $value, R_CURSOR()) ;
528 $status == 0 and $key eq $origkey ;
529 $status = $db->seq($key, $value, R_NEXT()) ) {
531 # save the value or count number of matches
534 { ++ $values{$value} }
536 { push (@values, $value) }
543 return ($wantarray ? ($flag ? %values : @values) : $counter) ;
552 DB_File - Perl5 access to Berkeley DB version 1.x
558 [$X =] tie %hash, 'DB_File', [$filename, $flags, $mode, $DB_HASH] ;
559 [$X =] tie %hash, 'DB_File', $filename, $flags, $mode, $DB_BTREE ;
560 [$X =] tie @array, 'DB_File', $filename, $flags, $mode, $DB_RECNO ;
562 $status = $X->del($key [, $flags]) ;
563 $status = $X->put($key, $value [, $flags]) ;
564 $status = $X->get($key, $value [, $flags]) ;
565 $status = $X->seq($key, $value, $flags) ;
566 $status = $X->sync([$flags]) ;
570 $count = $X->get_dup($key) ;
571 @list = $X->get_dup($key) ;
572 %list = $X->get_dup($key, 1) ;
573 $status = $X->find_dup($key, $value) ;
574 $status = $X->del_dup($key, $value) ;
582 @r = $X->splice(offset, length, elements);
585 $old_filter = $db->filter_store_key ( sub { ... } ) ;
586 $old_filter = $db->filter_store_value( sub { ... } ) ;
587 $old_filter = $db->filter_fetch_key ( sub { ... } ) ;
588 $old_filter = $db->filter_fetch_value( sub { ... } ) ;
595 B<DB_File> is a module which allows Perl programs to make use of the
596 facilities provided by Berkeley DB version 1.x (if you have a newer
597 version of DB, see L<Using DB_File with Berkeley DB version 2 or 3>).
598 It is assumed that you have a copy of the Berkeley DB manual pages at
599 hand when reading this documentation. The interface defined here
600 mirrors the Berkeley DB interface closely.
602 Berkeley DB is a C library which provides a consistent interface to a
603 number of database formats. B<DB_File> provides an interface to all
604 three of the database types currently supported by Berkeley DB.
612 This database type allows arbitrary key/value pairs to be stored in data
613 files. This is equivalent to the functionality provided by other
614 hashing packages like DBM, NDBM, ODBM, GDBM, and SDBM. Remember though,
615 the files created using DB_HASH are not compatible with any of the
616 other packages mentioned.
618 A default hashing algorithm, which will be adequate for most
619 applications, is built into Berkeley DB. If you do need to use your own
620 hashing algorithm it is possible to write your own in Perl and have
621 B<DB_File> use it instead.
625 The btree format allows arbitrary key/value pairs to be stored in a
626 sorted, balanced binary tree.
628 As with the DB_HASH format, it is possible to provide a user defined
629 Perl routine to perform the comparison of keys. By default, though, the
630 keys are stored in lexical order.
634 DB_RECNO allows both fixed-length and variable-length flat text files
635 to be manipulated using the same key/value pair interface as in DB_HASH
636 and DB_BTREE. In this case the key will consist of a record (line)
641 =head2 Using DB_File with Berkeley DB version 2 or 3
643 Although B<DB_File> is intended to be used with Berkeley DB version 1,
644 it can also be used with version 2 or 3. In this case the interface is
645 limited to the functionality provided by Berkeley DB 1.x. Anywhere the
646 version 2 or 3 interface differs, B<DB_File> arranges for it to work
647 like version 1. This feature allows B<DB_File> scripts that were built
648 with version 1 to be migrated to version 2 or 3 without any changes.
650 If you want to make use of the new features available in Berkeley DB
651 2.x or greater, use the Perl module B<BerkeleyDB> instead.
653 B<Note:> The database file format has changed in both Berkeley DB
654 version 2 and 3. If you cannot recreate your databases, you must dump
655 any existing databases with either the C<db_dump> or the C<db_dump185>
656 utility that comes with Berkeley DB.
657 Once you have rebuilt DB_File to use Berkeley DB version 2 or 3, your
658 databases can be recreated using C<db_load>. Refer to the Berkeley DB
659 documentation for further details.
661 Please read L<"COPYRIGHT"> before using version 2.x or 3.x of Berkeley
664 =head2 Interface to Berkeley DB
666 B<DB_File> allows access to Berkeley DB files using the tie() mechanism
667 in Perl 5 (for full details, see L<perlfunc/tie()>). This facility
668 allows B<DB_File> to access Berkeley DB files using either an
669 associative array (for DB_HASH & DB_BTREE file types) or an ordinary
670 array (for the DB_RECNO file type).
672 In addition to the tie() interface, it is also possible to access most
673 of the functions provided in the Berkeley DB API directly.
674 See L<THE API INTERFACE>.
676 =head2 Opening a Berkeley DB Database File
678 Berkeley DB uses the function dbopen() to open or create a database.
679 Here is the C prototype for dbopen():
682 dbopen (const char * file, int flags, int mode,
683 DBTYPE type, const void * openinfo)
685 The parameter C<type> is an enumeration which specifies which of the 3
686 interface methods (DB_HASH, DB_BTREE or DB_RECNO) is to be used.
687 Depending on which of these is actually chosen, the final parameter,
688 I<openinfo> points to a data structure which allows tailoring of the
689 specific interface method.
691 This interface is handled slightly differently in B<DB_File>. Here is
692 an equivalent call using B<DB_File>:
694 tie %array, 'DB_File', $filename, $flags, $mode, $DB_HASH ;
696 The C<filename>, C<flags> and C<mode> parameters are the direct
697 equivalent of their dbopen() counterparts. The final parameter $DB_HASH
698 performs the function of both the C<type> and C<openinfo> parameters in
701 In the example above $DB_HASH is actually a pre-defined reference to a
702 hash object. B<DB_File> has three of these pre-defined references.
703 Apart from $DB_HASH, there is also $DB_BTREE and $DB_RECNO.
705 The keys allowed in each of these pre-defined references is limited to
706 the names used in the equivalent C structure. So, for example, the
707 $DB_HASH reference will only allow keys called C<bsize>, C<cachesize>,
708 C<ffactor>, C<hash>, C<lorder> and C<nelem>.
710 To change one of these elements, just assign to it like this:
712 $DB_HASH->{'cachesize'} = 10000 ;
714 The three predefined variables $DB_HASH, $DB_BTREE and $DB_RECNO are
715 usually adequate for most applications. If you do need to create extra
716 instances of these objects, constructors are available for each file
719 Here are examples of the constructors and the valid options available
720 for DB_HASH, DB_BTREE and DB_RECNO respectively.
722 $a = new DB_File::HASHINFO ;
730 $b = new DB_File::BTREEINFO ;
740 $c = new DB_File::RECNOINFO ;
749 The values stored in the hashes above are mostly the direct equivalent
750 of their C counterpart. Like their C counterparts, all are set to a
751 default values - that means you don't have to set I<all> of the
752 values when you only want to change one. Here is an example:
754 $a = new DB_File::HASHINFO ;
755 $a->{'cachesize'} = 12345 ;
756 tie %y, 'DB_File', "filename", $flags, 0777, $a ;
758 A few of the options need extra discussion here. When used, the C
759 equivalent of the keys C<hash>, C<compare> and C<prefix> store pointers
760 to C functions. In B<DB_File> these keys are used to store references
761 to Perl subs. Below are templates for each of the subs:
767 # return the hash value for $data
773 my ($key, $key2) = @_ ;
775 # return 0 if $key1 eq $key2
776 # -1 if $key1 lt $key2
777 # 1 if $key1 gt $key2
778 return (-1 , 0 or 1) ;
783 my ($key, $key2) = @_ ;
785 # return number of bytes of $key2 which are
786 # necessary to determine that it is greater than $key1
790 See L<Changing the BTREE sort order> for an example of using the
793 If you are using the DB_RECNO interface and you intend making use of
794 C<bval>, you should check out L<The 'bval' Option>.
796 =head2 Default Parameters
798 It is possible to omit some or all of the final 4 parameters in the
799 call to C<tie> and let them take default values. As DB_HASH is the most
800 common file format used, the call:
802 tie %A, "DB_File", "filename" ;
806 tie %A, "DB_File", "filename", O_CREAT|O_RDWR, 0666, $DB_HASH ;
808 It is also possible to omit the filename parameter as well, so the
815 tie %A, "DB_File", undef, O_CREAT|O_RDWR, 0666, $DB_HASH ;
817 See L<In Memory Databases> for a discussion on the use of C<undef>
818 in place of a filename.
820 =head2 In Memory Databases
822 Berkeley DB allows the creation of in-memory databases by using NULL
823 (that is, a C<(char *)0> in C) in place of the filename. B<DB_File>
824 uses C<undef> instead of NULL to provide this functionality.
828 The DB_HASH file format is probably the most commonly used of the three
829 file formats that B<DB_File> supports. It is also very straightforward
832 =head2 A Simple Example
834 This example shows how to create a database, add key/value pairs to the
835 database, delete keys/value pairs and finally how to enumerate the
836 contents of the database.
841 use vars qw( %h $k $v ) ;
844 tie %h, "DB_File", "fruit", O_RDWR|O_CREAT, 0666, $DB_HASH
845 or die "Cannot open file 'fruit': $!\n";
847 # Add a few key/value pairs to the file
848 $h{"apple"} = "red" ;
849 $h{"orange"} = "orange" ;
850 $h{"banana"} = "yellow" ;
851 $h{"tomato"} = "red" ;
853 # Check for existence of a key
854 print "Banana Exists\n\n" if $h{"banana"} ;
856 # Delete a key/value pair.
859 # print the contents of the file
860 while (($k, $v) = each %h)
861 { print "$k -> $v\n" }
873 Note that the like ordinary associative arrays, the order of the keys
874 retrieved is in an apparently random order.
878 The DB_BTREE format is useful when you want to store data in a given
879 order. By default the keys will be stored in lexical order, but as you
880 will see from the example shown in the next section, it is very easy to
881 define your own sorting function.
883 =head2 Changing the BTREE sort order
885 This script shows how to override the default sorting algorithm that
886 BTREE uses. Instead of using the normal lexical ordering, a case
887 insensitive compare function will be used.
897 my ($key1, $key2) = @_ ;
898 "\L$key1" cmp "\L$key2" ;
901 # specify the Perl sub that will do the comparison
902 $DB_BTREE->{'compare'} = \&Compare ;
905 tie %h, "DB_File", "tree", O_RDWR|O_CREAT, 0666, $DB_BTREE
906 or die "Cannot open file 'tree': $!\n" ;
908 # Add a key/value pair to the file
909 $h{'Wall'} = 'Larry' ;
910 $h{'Smith'} = 'John' ;
911 $h{'mouse'} = 'mickey' ;
912 $h{'duck'} = 'donald' ;
917 # Cycle through the keys printing them in order.
918 # Note it is not necessary to sort the keys as
919 # the btree will have kept them in order automatically.
925 Here is the output from the code above.
931 There are a few point to bear in mind if you want to change the
932 ordering in a BTREE database:
938 The new compare function must be specified when you create the database.
942 You cannot change the ordering once the database has been created. Thus
943 you must use the same compare function every time you access the
948 Duplicate keys are entirely defined by the comparison function.
949 In the case-insensitive example above, the keys: 'KEY' and 'key'
950 would be considered duplicates, and assigning to the second one
951 would overwirte the first. If duplicates are allowed for (with the
952 R_DUPS flag discussed below), only a single copy of duplicate keys
953 is stored in the database --- so (again with example above) assigning
954 three values to the keys: 'KEY', 'Key', and 'key' would leave just
955 the first key: 'KEY' in the database with three values. For some
956 situations this results in information loss, so care should be taken
957 to provide fully qualified comparison functions when necessary.
958 For example, the above comparison routine could be modified to
959 additionally compare case-sensitively if two keys are equal in the
960 case insensitive comparison:
963 my($key1, $key2) = @_;
964 lc $key1 cmp lc $key2 ||
968 And now you will only have duplicates when the keys themselves
969 are truly the same. (note: in versions of the db library prior to
970 about November 1996, such duplicate keys were retained so it was
971 possible to recover the original keys in sets of keys that
977 =head2 Handling Duplicate Keys
979 The BTREE file type optionally allows a single key to be associated
980 with an arbitrary number of values. This option is enabled by setting
981 the flags element of C<$DB_BTREE> to R_DUP when creating the database.
983 There are some difficulties in using the tied hash interface if you
984 want to manipulate a BTREE database with duplicate keys. Consider this
991 use vars qw($filename %h ) ;
996 # Enable duplicate records
997 $DB_BTREE->{'flags'} = R_DUP ;
999 tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1000 or die "Cannot open $filename: $!\n";
1002 # Add some key/value pairs to the file
1003 $h{'Wall'} = 'Larry' ;
1004 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1005 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1006 $h{'Smith'} = 'John' ;
1007 $h{'mouse'} = 'mickey' ;
1009 # iterate through the associative array
1010 # and print each key/value pair.
1011 foreach (sort keys %h)
1012 { print "$_ -> $h{$_}\n" }
1024 As you can see 3 records have been successfully created with key C<Wall>
1025 - the only thing is, when they are retrieved from the database they
1026 I<seem> to have the same value, namely C<Larry>. The problem is caused
1027 by the way that the associative array interface works. Basically, when
1028 the associative array interface is used to fetch the value associated
1029 with a given key, it will only ever retrieve the first value.
1031 Although it may not be immediately obvious from the code above, the
1032 associative array interface can be used to write values with duplicate
1033 keys, but it cannot be used to read them back from the database.
1035 The way to get around this problem is to use the Berkeley DB API method
1036 called C<seq>. This method allows sequential access to key/value
1037 pairs. See L<THE API INTERFACE> for details of both the C<seq> method
1038 and the API in general.
1040 Here is the script above rewritten using the C<seq> API method.
1046 use vars qw($filename $x %h $status $key $value) ;
1048 $filename = "tree" ;
1051 # Enable duplicate records
1052 $DB_BTREE->{'flags'} = R_DUP ;
1054 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1055 or die "Cannot open $filename: $!\n";
1057 # Add some key/value pairs to the file
1058 $h{'Wall'} = 'Larry' ;
1059 $h{'Wall'} = 'Brick' ; # Note the duplicate key
1060 $h{'Wall'} = 'Brick' ; # Note the duplicate key and value
1061 $h{'Smith'} = 'John' ;
1062 $h{'mouse'} = 'mickey' ;
1064 # iterate through the btree using seq
1065 # and print each key/value pair.
1067 for ($status = $x->seq($key, $value, R_FIRST) ;
1069 $status = $x->seq($key, $value, R_NEXT) )
1070 { print "$key -> $value\n" }
1083 This time we have got all the key/value pairs, including the multiple
1084 values associated with the key C<Wall>.
1086 To make life easier when dealing with duplicate keys, B<DB_File> comes with
1087 a few utility methods.
1089 =head2 The get_dup() Method
1091 The C<get_dup> method assists in
1092 reading duplicate values from BTREE databases. The method can take the
1095 $count = $x->get_dup($key) ;
1096 @list = $x->get_dup($key) ;
1097 %list = $x->get_dup($key, 1) ;
1099 In a scalar context the method returns the number of values associated
1100 with the key, C<$key>.
1102 In list context, it returns all the values which match C<$key>. Note
1103 that the values will be returned in an apparently random order.
1105 In list context, if the second parameter is present and evaluates
1106 TRUE, the method returns an associative array. The keys of the
1107 associative array correspond to the values that matched in the BTREE
1108 and the values of the array are a count of the number of times that
1109 particular value occurred in the BTREE.
1111 So assuming the database created above, we can use C<get_dup> like
1118 use vars qw($filename $x %h ) ;
1120 $filename = "tree" ;
1122 # Enable duplicate records
1123 $DB_BTREE->{'flags'} = R_DUP ;
1125 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1126 or die "Cannot open $filename: $!\n";
1128 my $cnt = $x->get_dup("Wall") ;
1129 print "Wall occurred $cnt times\n" ;
1131 my %hash = $x->get_dup("Wall", 1) ;
1132 print "Larry is there\n" if $hash{'Larry'} ;
1133 print "There are $hash{'Brick'} Brick Walls\n" ;
1135 my @list = sort $x->get_dup("Wall") ;
1136 print "Wall => [@list]\n" ;
1138 @list = $x->get_dup("Smith") ;
1139 print "Smith => [@list]\n" ;
1141 @list = $x->get_dup("Dog") ;
1142 print "Dog => [@list]\n" ;
1147 Wall occurred 3 times
1149 There are 2 Brick Walls
1150 Wall => [Brick Brick Larry]
1154 =head2 The find_dup() Method
1156 $status = $X->find_dup($key, $value) ;
1158 This method checks for the existence of a specific key/value pair. If the
1159 pair exists, the cursor is left pointing to the pair and the method
1160 returns 0. Otherwise the method returns a non-zero value.
1162 Assuming the database from the previous example:
1168 use vars qw($filename $x %h $found) ;
1170 my $filename = "tree" ;
1172 # Enable duplicate records
1173 $DB_BTREE->{'flags'} = R_DUP ;
1175 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1176 or die "Cannot open $filename: $!\n";
1178 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1179 print "Larry Wall is $found there\n" ;
1181 $found = ( $x->find_dup("Wall", "Harry") == 0 ? "" : "not") ;
1182 print "Harry Wall is $found there\n" ;
1190 Harry Wall is not there
1193 =head2 The del_dup() Method
1195 $status = $X->del_dup($key, $value) ;
1197 This method deletes a specific key/value pair. It returns
1198 0 if they exist and have been deleted successfully.
1199 Otherwise the method returns a non-zero value.
1201 Again assuming the existence of the C<tree> database
1207 use vars qw($filename $x %h $found) ;
1209 my $filename = "tree" ;
1211 # Enable duplicate records
1212 $DB_BTREE->{'flags'} = R_DUP ;
1214 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1215 or die "Cannot open $filename: $!\n";
1217 $x->del_dup("Wall", "Larry") ;
1219 $found = ( $x->find_dup("Wall", "Larry") == 0 ? "" : "not") ;
1220 print "Larry Wall is $found there\n" ;
1227 Larry Wall is not there
1229 =head2 Matching Partial Keys
1231 The BTREE interface has a feature which allows partial keys to be
1232 matched. This functionality is I<only> available when the C<seq> method
1233 is used along with the R_CURSOR flag.
1235 $x->seq($key, $value, R_CURSOR) ;
1237 Here is the relevant quote from the dbopen man page where it defines
1238 the use of the R_CURSOR flag with seq:
1240 Note, for the DB_BTREE access method, the returned key is not
1241 necessarily an exact match for the specified key. The returned key
1242 is the smallest key greater than or equal to the specified key,
1243 permitting partial key matches and range searches.
1245 In the example script below, the C<match> sub uses this feature to find
1246 and print the first matching key/value pair given a partial key.
1253 use vars qw($filename $x %h $st $key $value) ;
1259 my $orig_key = $key ;
1260 $x->seq($key, $value, R_CURSOR) ;
1261 print "$orig_key\t-> $key\t-> $value\n" ;
1264 $filename = "tree" ;
1267 $x = tie %h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_BTREE
1268 or die "Cannot open $filename: $!\n";
1270 # Add some key/value pairs to the file
1271 $h{'mouse'} = 'mickey' ;
1272 $h{'Wall'} = 'Larry' ;
1273 $h{'Walls'} = 'Brick' ;
1274 $h{'Smith'} = 'John' ;
1278 print "IN ORDER\n" ;
1279 for ($st = $x->seq($key, $value, R_FIRST) ;
1281 $st = $x->seq($key, $value, R_NEXT) )
1283 { print "$key -> $value\n" }
1285 print "\nPARTIAL MATCH\n" ;
1305 a -> mouse -> mickey
1309 DB_RECNO provides an interface to flat text files. Both variable and
1310 fixed length records are supported.
1312 In order to make RECNO more compatible with Perl, the array offset for
1313 all RECNO arrays begins at 0 rather than 1 as in Berkeley DB.
1315 As with normal Perl arrays, a RECNO array can be accessed using
1316 negative indexes. The index -1 refers to the last element of the array,
1317 -2 the second last, and so on. Attempting to access an element before
1318 the start of the array will raise a fatal run-time error.
1320 =head2 The 'bval' Option
1322 The operation of the bval option warrants some discussion. Here is the
1323 definition of bval from the Berkeley DB 1.85 recno manual page:
1325 The delimiting byte to be used to mark the end of a
1326 record for variable-length records, and the pad charac-
1327 ter for fixed-length records. If no value is speci-
1328 fied, newlines (``\n'') are used to mark the end of
1329 variable-length records and fixed-length records are
1332 The second sentence is wrong. In actual fact bval will only default to
1333 C<"\n"> when the openinfo parameter in dbopen is NULL. If a non-NULL
1334 openinfo parameter is used at all, the value that happens to be in bval
1335 will be used. That means you always have to specify bval when making
1336 use of any of the options in the openinfo parameter. This documentation
1337 error will be fixed in the next release of Berkeley DB.
1339 That clarifies the situation with regards Berkeley DB itself. What
1340 about B<DB_File>? Well, the behavior defined in the quote above is
1341 quite useful, so B<DB_File> conforms to it.
1343 That means that you can specify other options (e.g. cachesize) and
1344 still have bval default to C<"\n"> for variable length records, and
1345 space for fixed length records.
1347 Also note that the bval option only allows you to specify a single byte
1350 =head2 A Simple Example
1352 Here is a simple example that uses RECNO (if you are using a version
1353 of Perl earlier than 5.004_57 this example won't work -- see
1354 L<Extra RECNO Methods> for a workaround).
1360 my $filename = "text" ;
1364 tie @h, "DB_File", $filename, O_RDWR|O_CREAT, 0666, $DB_RECNO
1365 or die "Cannot open file 'text': $!\n" ;
1367 # Add a few key/value pairs to the file
1372 push @h, "green", "black" ;
1374 my $elements = scalar @h ;
1375 print "The array contains $elements entries\n" ;
1378 print "popped $last\n" ;
1380 unshift @h, "white" ;
1381 my $first = shift @h ;
1382 print "shifted $first\n" ;
1384 # Check for existence of a key
1385 print "Element 1 Exists with value $h[1]\n" if $h[1] ;
1387 # use a negative index
1388 print "The last element is $h[-1]\n" ;
1389 print "The 2nd last element is $h[-2]\n" ;
1393 Here is the output from the script:
1395 The array contains 5 entries
1398 Element 1 Exists with value blue
1399 The last element is green
1400 The 2nd last element is yellow
1402 =head2 Extra RECNO Methods
1404 If you are using a version of Perl earlier than 5.004_57, the tied
1405 array interface is quite limited. In the example script above
1406 C<push>, C<pop>, C<shift>, C<unshift>
1407 or determining the array length will not work with a tied array.
1409 To make the interface more useful for older versions of Perl, a number
1410 of methods are supplied with B<DB_File> to simulate the missing array
1411 operations. All these methods are accessed via the object returned from
1414 Here are the methods:
1418 =item B<$X-E<gt>push(list) ;>
1420 Pushes the elements of C<list> to the end of the array.
1422 =item B<$value = $X-E<gt>pop ;>
1424 Removes and returns the last element of the array.
1426 =item B<$X-E<gt>shift>
1428 Removes and returns the first element of the array.
1430 =item B<$X-E<gt>unshift(list) ;>
1432 Pushes the elements of C<list> to the start of the array.
1434 =item B<$X-E<gt>length>
1436 Returns the number of elements in the array.
1438 =item B<$X-E<gt>splice(offset, length, elements);>
1440 Returns a splice of the the array.
1444 =head2 Another Example
1446 Here is a more complete example that makes use of some of the methods
1447 described above. It also makes use of the API interface directly (see
1448 L<THE API INTERFACE>).
1452 use vars qw(@h $H $file $i) ;
1460 $H = tie @h, "DB_File", $file, O_RDWR|O_CREAT, 0666, $DB_RECNO
1461 or die "Cannot open file $file: $!\n" ;
1463 # first create a text file to play with
1471 # Print the records in order.
1473 # The length method is needed here because evaluating a tied
1474 # array in a scalar context does not return the number of
1475 # elements in the array.
1477 print "\nORIGINAL\n" ;
1478 foreach $i (0 .. $H->length - 1) {
1479 print "$i: $h[$i]\n" ;
1482 # use the push & pop methods
1485 print "\nThe last record was [$a]\n" ;
1487 # and the shift & unshift methods
1489 $H->unshift("first") ;
1490 print "The first record was [$a]\n" ;
1492 # Use the API to add a new record after record 2.
1494 $H->put($i, "Newbie", R_IAFTER) ;
1496 # and a new record before record 1.
1498 $H->put($i, "New One", R_IBEFORE) ;
1503 # now print the records in reverse order
1504 print "\nREVERSE\n" ;
1505 for ($i = $H->length - 1 ; $i >= 0 ; -- $i)
1506 { print "$i: $h[$i]\n" }
1508 # same again, but use the API functions instead
1509 print "\nREVERSE again\n" ;
1510 my ($s, $k, $v) = (0, 0, 0) ;
1511 for ($s = $H->seq($k, $v, R_LAST) ;
1513 $s = $H->seq($k, $v, R_PREV))
1514 { print "$k: $v\n" }
1519 and this is what it outputs:
1528 The last record was [four]
1529 The first record was [zero]
1553 Rather than iterating through the array, C<@h> like this:
1557 it is necessary to use either this:
1559 foreach $i (0 .. $H->length - 1)
1563 for ($a = $H->get($k, $v, R_FIRST) ;
1565 $a = $H->get($k, $v, R_NEXT) )
1569 Notice that both times the C<put> method was used the record index was
1570 specified using a variable, C<$i>, rather than the literal value
1571 itself. This is because C<put> will return the record number of the
1572 inserted line via that parameter.
1576 =head1 THE API INTERFACE
1578 As well as accessing Berkeley DB using a tied hash or array, it is also
1579 possible to make direct use of most of the API functions defined in the
1580 Berkeley DB documentation.
1582 To do this you need to store a copy of the object returned from the tie.
1584 $db = tie %hash, "DB_File", "filename" ;
1586 Once you have done that, you can access the Berkeley DB API functions
1587 as B<DB_File> methods directly like this:
1589 $db->put($key, $value, R_NOOVERWRITE) ;
1591 B<Important:> If you have saved a copy of the object returned from
1592 C<tie>, the underlying database file will I<not> be closed until both
1593 the tied variable is untied and all copies of the saved object are
1597 $db = tie %hash, "DB_File", "filename"
1598 or die "Cannot tie filename: $!" ;
1603 See L<The untie() Gotcha> for more details.
1605 All the functions defined in L<dbopen> are available except for
1606 close() and dbopen() itself. The B<DB_File> method interface to the
1607 supported functions have been implemented to mirror the way Berkeley DB
1608 works whenever possible. In particular note that:
1614 The methods return a status value. All return 0 on success.
1615 All return -1 to signify an error and set C<$!> to the exact
1616 error code. The return code 1 generally (but not always) means that the
1617 key specified did not exist in the database.
1619 Other return codes are defined. See below and in the Berkeley DB
1620 documentation for details. The Berkeley DB documentation should be used
1621 as the definitive source.
1625 Whenever a Berkeley DB function returns data via one of its parameters,
1626 the equivalent B<DB_File> method does exactly the same.
1630 If you are careful, it is possible to mix API calls with the tied
1631 hash/array interface in the same piece of code. Although only a few of
1632 the methods used to implement the tied interface currently make use of
1633 the cursor, you should always assume that the cursor has been changed
1634 any time the tied hash/array interface is used. As an example, this
1635 code will probably not do what you expect:
1637 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1638 or die "Cannot tie $filename: $!" ;
1640 # Get the first key/value pair and set the cursor
1641 $X->seq($key, $value, R_FIRST) ;
1643 # this line will modify the cursor
1644 $count = scalar keys %x ;
1646 # Get the second key/value pair.
1647 # oops, it didn't, it got the last key/value pair!
1648 $X->seq($key, $value, R_NEXT) ;
1650 The code above can be rearranged to get around the problem, like this:
1652 $X = tie %x, 'DB_File', $filename, O_RDWR|O_CREAT, 0777, $DB_BTREE
1653 or die "Cannot tie $filename: $!" ;
1655 # this line will modify the cursor
1656 $count = scalar keys %x ;
1658 # Get the first key/value pair and set the cursor
1659 $X->seq($key, $value, R_FIRST) ;
1661 # Get the second key/value pair.
1663 $X->seq($key, $value, R_NEXT) ;
1667 All the constants defined in L<dbopen> for use in the flags parameters
1668 in the methods defined below are also available. Refer to the Berkeley
1669 DB documentation for the precise meaning of the flags values.
1671 Below is a list of the methods available.
1675 =item B<$status = $X-E<gt>get($key, $value [, $flags]) ;>
1677 Given a key (C<$key>) this method reads the value associated with it
1678 from the database. The value read from the database is returned in the
1679 C<$value> parameter.
1681 If the key does not exist the method returns 1.
1683 No flags are currently defined for this method.
1685 =item B<$status = $X-E<gt>put($key, $value [, $flags]) ;>
1687 Stores the key/value pair in the database.
1689 If you use either the R_IAFTER or R_IBEFORE flags, the C<$key> parameter
1690 will have the record number of the inserted key/value pair set.
1692 Valid flags are R_CURSOR, R_IAFTER, R_IBEFORE, R_NOOVERWRITE and
1695 =item B<$status = $X-E<gt>del($key [, $flags]) ;>
1697 Removes all key/value pairs with key C<$key> from the database.
1699 A return code of 1 means that the requested key was not in the
1702 R_CURSOR is the only valid flag at present.
1704 =item B<$status = $X-E<gt>fd ;>
1706 Returns the file descriptor for the underlying database.
1708 See L<Locking: The Trouble with fd> for an explanation for why you should
1709 not use C<fd> to lock your database.
1711 =item B<$status = $X-E<gt>seq($key, $value, $flags) ;>
1713 This interface allows sequential retrieval from the database. See
1714 L<dbopen> for full details.
1716 Both the C<$key> and C<$value> parameters will be set to the key/value
1717 pair read from the database.
1719 The flags parameter is mandatory. The valid flag values are R_CURSOR,
1720 R_FIRST, R_LAST, R_NEXT and R_PREV.
1722 =item B<$status = $X-E<gt>sync([$flags]) ;>
1724 Flushes any cached buffers to disk.
1726 R_RECNOSYNC is the only valid flag at present.
1732 A DBM Filter is a piece of code that is be used when you I<always>
1733 want to make the same transformation to all keys and/or values in a
1736 There are four methods associated with DBM Filters. All work identically,
1737 and each is used to install (or uninstall) a single DBM Filter. Each
1738 expects a single parameter, namely a reference to a sub. The only
1739 difference between them is the place that the filter is installed.
1745 =item B<filter_store_key>
1747 If a filter has been installed with this method, it will be invoked
1748 every time you write a key to a DBM database.
1750 =item B<filter_store_value>
1752 If a filter has been installed with this method, it will be invoked
1753 every time you write a value to a DBM database.
1756 =item B<filter_fetch_key>
1758 If a filter has been installed with this method, it will be invoked
1759 every time you read a key from a DBM database.
1761 =item B<filter_fetch_value>
1763 If a filter has been installed with this method, it will be invoked
1764 every time you read a value from a DBM database.
1768 You can use any combination of the methods, from none, to all four.
1770 All filter methods return the existing filter, if present, or C<undef>
1773 To delete a filter pass C<undef> to it.
1777 When each filter is called by Perl, a local copy of C<$_> will contain
1778 the key or value to be filtered. Filtering is achieved by modifying
1779 the contents of C<$_>. The return code from the filter is ignored.
1781 =head2 An Example -- the NULL termination problem.
1783 Consider the following scenario. You have a DBM database
1784 that you need to share with a third-party C application. The C application
1785 assumes that I<all> keys and values are NULL terminated. Unfortunately
1786 when Perl writes to DBM databases it doesn't use NULL termination, so
1787 your Perl application will have to manage NULL termination itself. When
1788 you write to the database you will have to use something like this:
1790 $hash{"$key\0"} = "$value\0" ;
1792 Similarly the NULL needs to be taken into account when you are considering
1793 the length of existing keys/values.
1795 It would be much better if you could ignore the NULL terminations issue
1796 in the main application code and have a mechanism that automatically
1797 added the terminating NULL to all keys and values whenever you write to
1798 the database and have them removed when you read from the database. As I'm
1799 sure you have already guessed, this is a problem that DBM Filters can
1807 my $filename = "/tmp/filt" ;
1810 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1811 or die "Cannot open $filename: $!\n" ;
1813 # Install DBM Filters
1814 $db->filter_fetch_key ( sub { s/\0$// } ) ;
1815 $db->filter_store_key ( sub { $_ .= "\0" } ) ;
1816 $db->filter_fetch_value( sub { s/\0$// } ) ;
1817 $db->filter_store_value( sub { $_ .= "\0" } ) ;
1819 $hash{"abc"} = "def" ;
1820 my $a = $hash{"ABC"} ;
1825 Hopefully the contents of each of the filters should be
1826 self-explanatory. Both "fetch" filters remove the terminating NULL,
1827 and both "store" filters add a terminating NULL.
1830 =head2 Another Example -- Key is a C int.
1832 Here is another real-life example. By default, whenever Perl writes to
1833 a DBM database it always writes the key and value as strings. So when
1836 $hash{12345} = "soemthing" ;
1838 the key 12345 will get stored in the DBM database as the 5 byte string
1839 "12345". If you actually want the key to be stored in the DBM database
1840 as a C int, you will have to use C<pack> when writing, and C<unpack>
1843 Here is a DBM Filter that does it:
1849 my $filename = "/tmp/filt" ;
1853 my $db = tie %hash, 'DB_File', $filename, O_CREAT|O_RDWR, 0666, $DB_HASH
1854 or die "Cannot open $filename: $!\n" ;
1856 $db->filter_fetch_key ( sub { $_ = unpack("i", $_) } ) ;
1857 $db->filter_store_key ( sub { $_ = pack ("i", $_) } ) ;
1858 $hash{123} = "def" ;
1863 This time only two filters have been used -- we only need to manipulate
1864 the contents of the key, so it wasn't necessary to install any value
1867 =head1 HINTS AND TIPS
1870 =head2 Locking: The Trouble with fd
1872 Until version 1.72 of this module, the recommended technique for locking
1873 B<DB_File> databases was to flock the filehandle returned from the "fd"
1874 function. Unfortunately this technique has been shown to be fundamentally
1875 flawed (Kudos to David Harris for tracking this down). Use it at your own
1878 The locking technique went like this.
1880 $db = tie(%db, 'DB_File', '/tmp/foo.db', O_CREAT|O_RDWR, 0666)
1881 || die "dbcreat /tmp/foo.db $!";
1883 open(DB_FH, "+<&=$fd") || die "dup $!";
1884 flock (DB_FH, LOCK_EX) || die "flock: $!";
1886 $db{"Tom"} = "Jerry" ;
1888 flock(DB_FH, LOCK_UN);
1893 In simple terms, this is what happens:
1899 Use "tie" to open the database.
1903 Lock the database with fd & flock.
1907 Read & Write to the database.
1911 Unlock and close the database.
1915 Here is the crux of the problem. A side-effect of opening the B<DB_File>
1916 database in step 2 is that an initial block from the database will get
1917 read from disk and cached in memory.
1919 To see why this is a problem, consider what can happen when two processes,
1920 say "A" and "B", both want to update the same B<DB_File> database
1921 using the locking steps outlined above. Assume process "A" has already
1922 opened the database and has a write lock, but it hasn't actually updated
1923 the database yet (it has finished step 2, but not started step 3 yet). Now
1924 process "B" tries to open the same database - step 1 will succeed,
1925 but it will block on step 2 until process "A" releases the lock. The
1926 important thing to notice here is that at this point in time both
1927 processes will have cached identical initial blocks from the database.
1929 Now process "A" updates the database and happens to change some of the
1930 data held in the initial buffer. Process "A" terminates, flushing
1931 all cached data to disk and releasing the database lock. At this point
1932 the database on disk will correctly reflect the changes made by process
1935 With the lock released, process "B" can now continue. It also updates the
1936 database and unfortunately it too modifies the data that was in its
1937 initial buffer. Once that data gets flushed to disk it will overwrite
1938 some/all of the changes process "A" made to the database.
1940 The result of this scenario is at best a database that doesn't contain
1941 what you expect. At worst the database will corrupt.
1943 The above won't happen every time competing process update the same
1944 B<DB_File> database, but it does illustrate why the technique should
1947 =head2 Safe ways to lock a database
1949 Starting with version 2.x, Berkeley DB has internal support for locking.
1950 The companion module to this one, B<BerkeleyDB>, provides an interface
1951 to this locking functionality. If you are serious about locking
1952 Berkeley DB databases, I strongly recommend using B<BerkeleyDB>.
1954 If using B<BerkeleyDB> isn't an option, there are a number of modules
1955 available on CPAN that can be used to implement locking. Each one
1956 implements locking differently and has different goals in mind. It is
1957 therefore worth knowing the difference, so that you can pick the right
1958 one for your application. Here are the three locking wrappers:
1962 =item B<Tie::DB_Lock>
1964 A B<DB_File> wrapper which creates copies of the database file for
1965 read access, so that you have a kind of a multiversioning concurrent read
1966 system. However, updates are still serial. Use for databases where reads
1967 may be lengthy and consistency problems may occur.
1969 =item B<Tie::DB_LockFile>
1971 A B<DB_File> wrapper that has the ability to lock and unlock the database
1972 while it is being used. Avoids the tie-before-flock problem by simply
1973 re-tie-ing the database when you get or drop a lock. Because of the
1974 flexibility in dropping and re-acquiring the lock in the middle of a
1975 session, this can be massaged into a system that will work with long
1976 updates and/or reads if the application follows the hints in the POD
1979 =item B<DB_File::Lock>
1981 An extremely lightweight B<DB_File> wrapper that simply flocks a lockfile
1982 before tie-ing the database and drops the lock after the untie. Allows
1983 one to use the same lockfile for multiple databases to avoid deadlock
1984 problems, if desired. Use for databases where updates are reads are
1985 quick and simple flock locking semantics are enough.
1989 =head2 Sharing Databases With C Applications
1991 There is no technical reason why a Berkeley DB database cannot be
1992 shared by both a Perl and a C application.
1994 The vast majority of problems that are reported in this area boil down
1995 to the fact that C strings are NULL terminated, whilst Perl strings are
1996 not. See L<DBM FILTERS> for a generic way to work around this problem.
1998 Here is a real example. Netscape 2.0 keeps a record of the locations you
1999 visit along with the time you last visited them in a DB_HASH database.
2000 This is usually stored in the file F<~/.netscape/history.db>. The key
2001 field in the database is the location string and the value field is the
2002 time the location was last visited stored as a 4 byte binary value.
2004 If you haven't already guessed, the location string is stored with a
2005 terminating NULL. This means you need to be careful when accessing the
2008 Here is a snippet of code that is loosely based on Tom Christiansen's
2009 I<ggh> script (available from your nearest CPAN archive in
2010 F<authors/id/TOMC/scripts/nshist.gz>).
2017 use vars qw( $dotdir $HISTORY %hist_db $href $binary_time $date ) ;
2018 $dotdir = $ENV{HOME} || $ENV{LOGNAME};
2020 $HISTORY = "$dotdir/.netscape/history.db";
2022 tie %hist_db, 'DB_File', $HISTORY
2023 or die "Cannot open $HISTORY: $!\n" ;;
2025 # Dump the complete database
2026 while ( ($href, $binary_time) = each %hist_db ) {
2028 # remove the terminating NULL
2029 $href =~ s/\x00$// ;
2031 # convert the binary time into a user friendly string
2032 $date = localtime unpack("V", $binary_time);
2033 print "$date $href\n" ;
2036 # check for the existence of a specific key
2037 # remember to add the NULL
2038 if ( $binary_time = $hist_db{"http://mox.perl.com/\x00"} ) {
2039 $date = localtime unpack("V", $binary_time) ;
2040 print "Last visited mox.perl.com on $date\n" ;
2043 print "Never visited mox.perl.com\n"
2048 =head2 The untie() Gotcha
2050 If you make use of the Berkeley DB API, it is I<very> strongly
2051 recommended that you read L<perltie/The untie Gotcha>.
2053 Even if you don't currently make use of the API interface, it is still
2056 Here is an example which illustrates the problem from a B<DB_File>
2065 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_TRUNC
2066 or die "Cannot tie first time: $!" ;
2072 tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2073 or die "Cannot tie second time: $!" ;
2077 When run, the script will produce this error message:
2079 Cannot tie second time: Invalid argument at bad.file line 14.
2081 Although the error message above refers to the second tie() statement
2082 in the script, the source of the problem is really with the untie()
2083 statement that precedes it.
2085 Having read L<perltie> you will probably have already guessed that the
2086 error is caused by the extra copy of the tied object stored in C<$X>.
2087 If you haven't, then the problem boils down to the fact that the
2088 B<DB_File> destructor, DESTROY, will not be called until I<all>
2089 references to the tied object are destroyed. Both the tied variable,
2090 C<%x>, and C<$X> above hold a reference to the object. The call to
2091 untie() will destroy the first, but C<$X> still holds a valid
2092 reference, so the destructor will not get called and the database file
2093 F<tst.fil> will remain open. The fact that Berkeley DB then reports the
2094 attempt to open a database that is already open via the catch-all
2095 "Invalid argument" doesn't help.
2097 If you run the script with the C<-w> flag the error message becomes:
2099 untie attempted while 1 inner references still exist at bad.file line 12.
2100 Cannot tie second time: Invalid argument at bad.file line 14.
2102 which pinpoints the real problem. Finally the script can now be
2103 modified to fix the original problem by destroying the API object
2112 $X = tie %x, 'DB_File', 'tst.fil' , O_RDWR|O_CREAT
2116 =head1 COMMON QUESTIONS
2118 =head2 Why is there Perl source in my database?
2120 If you look at the contents of a database file created by DB_File,
2121 there can sometimes be part of a Perl script included in it.
2123 This happens because Berkeley DB uses dynamic memory to allocate
2124 buffers which will subsequently be written to the database file. Being
2125 dynamic, the memory could have been used for anything before DB
2126 malloced it. As Berkeley DB doesn't clear the memory once it has been
2127 allocated, the unused portions will contain random junk. In the case
2128 where a Perl script gets written to the database, the random junk will
2129 correspond to an area of dynamic memory that happened to be used during
2130 the compilation of the script.
2132 Unless you don't like the possibility of there being part of your Perl
2133 scripts embedded in a database file, this is nothing to worry about.
2135 =head2 How do I store complex data structures with DB_File?
2137 Although B<DB_File> cannot do this directly, there is a module which
2138 can layer transparently over B<DB_File> to accomplish this feat.
2140 Check out the MLDBM module, available on CPAN in the directory
2141 F<modules/by-module/MLDBM>.
2143 =head2 What does "Invalid Argument" mean?
2145 You will get this error message when one of the parameters in the
2146 C<tie> call is wrong. Unfortunately there are quite a few parameters to
2147 get wrong, so it can be difficult to figure out which one it is.
2149 Here are a couple of possibilities:
2155 Attempting to reopen a database without closing it.
2159 Using the O_WRONLY flag.
2163 =head2 What does "Bareword 'DB_File' not allowed" mean?
2165 You will encounter this particular error message when you have the
2166 C<strict 'subs'> pragma (or the full strict pragma) in your script.
2167 Consider this script:
2173 tie %x, DB_File, "filename" ;
2175 Running it produces the error in question:
2177 Bareword "DB_File" not allowed while "strict subs" in use
2179 To get around the error, place the word C<DB_File> in either single or
2180 double quotes, like this:
2182 tie %x, "DB_File", "filename" ;
2184 Although it might seem like a real pain, it is really worth the effort
2185 of having a C<use strict> in all your scripts.
2189 Articles that are either about B<DB_File> or make use of it.
2195 I<Full-Text Searching in Perl>, Tim Kientzle (tkientzle@ddj.com),
2196 Dr. Dobb's Journal, Issue 295, January 1999, pp 34-41
2202 Moved to the Changes file.
2206 Some older versions of Berkeley DB had problems with fixed length
2207 records using the RECNO file format. This problem has been fixed since
2208 version 1.85 of Berkeley DB.
2210 I am sure there are bugs in the code. If you do find any, or can
2211 suggest any enhancements, I would welcome your comments.
2215 B<DB_File> comes with the standard Perl source distribution. Look in
2216 the directory F<ext/DB_File>. Given the amount of time between releases
2217 of Perl the version that ships with Perl is quite likely to be out of
2218 date, so the most recent version can always be found on CPAN (see
2219 L<perlmod/CPAN> for details), in the directory
2220 F<modules/by-module/DB_File>.
2222 This version of B<DB_File> will work with either version 1.x, 2.x or
2223 3.x of Berkeley DB, but is limited to the functionality provided by
2226 The official web site for Berkeley DB is F<http://www.sleepycat.com>.
2227 All versions of Berkeley DB are available there.
2229 Alternatively, Berkeley DB version 1 is available at your nearest CPAN
2230 archive in F<src/misc/db.1.85.tar.gz>.
2232 If you are running IRIX, then get Berkeley DB version 1 from
2233 F<http://reality.sgi.com/ariel>. It has the patches necessary to
2234 compile properly on IRIX 5.3.
2238 Copyright (c) 1995-2001 Paul Marquess. All rights reserved. This program
2239 is free software; you can redistribute it and/or modify it under the
2240 same terms as Perl itself.
2242 Although B<DB_File> is covered by the Perl license, the library it
2243 makes use of, namely Berkeley DB, is not. Berkeley DB has its own
2244 copyright and its own license. Please take the time to read it.
2246 Here are are few words taken from the Berkeley DB FAQ (at
2247 F<http://www.sleepycat.com>) regarding the license:
2249 Do I have to license DB to use it in Perl scripts?
2251 No. The Berkeley DB license requires that software that uses
2252 Berkeley DB be freely redistributable. In the case of Perl, that
2253 software is Perl, and not your scripts. Any Perl scripts that you
2254 write are your property, including scripts that make use of
2255 Berkeley DB. Neither the Perl license nor the Berkeley DB license
2256 place any restriction on what you may do with them.
2258 If you are in any doubt about the license situation, contact either the
2259 Berkeley DB authors or the author of DB_File. See L<"AUTHOR"> for details.
2264 L<perl(1)>, L<dbopen(3)>, L<hash(3)>, L<recno(3)>, L<btree(3)>,
2269 The DB_File interface was written by Paul Marquess
2270 E<lt>Paul.Marquess@btinternet.comE<gt>.
2271 Questions about the DB system itself may be addressed to
2272 E<lt>db@sleepycat.com<gt>.