1 package List::MoreUtils;
8 use vars qw{ $VERSION @ISA @EXPORT_OK %EXPORT_TAGS };
11 @ISA = qw{ Exporter DynaLoader };
13 any all none notall true false
14 firstidx first_index lastidx last_index
15 insert_after insert_after_string
17 after after_incl before before_incl
18 firstval first_value lastval last_value
19 each_array each_arrayref
21 mesh zip uniq distinct
28 # Load the XS at compile-time so that redefinition warnings will be
29 # thrown correctly if the XS versions of part or indexes loaded
31 # PERL_DL_NONLAZY must be false, or any errors in loading will just
32 # cause the perl code to be tested
33 local $ENV{PERL_DL_NONLAZY} = 0 if $ENV{PERL_DL_NONLAZY};
35 bootstrap List::MoreUtils $VERSION;
38 } unless $ENV{LIST_MOREUTILS_PP};
41 # Always use Perl apply() until memory leaks are resolved.
44 &$action foreach my @values = @_;
45 wantarray ? @values : $values[-1];
48 # Always use Perl part() until memory leaks are resolved.
50 my ($code, @list) = @_;
52 push @{ $parts[ $code->($_) ] }, $_ foreach @list;
56 # Always use Perl indexes() until memory leaks are resolved.
65 # Load the pure-Perl versions of the other functions if needed
66 eval <<'END_PERL' unless defined &any;
68 # Use pure scalar boolean return values for compatibility with XS
69 use constant YES => ! 0;
70 use constant NO => ! 1;
83 return NO unless $f->();
99 return YES unless $f->();
117 $count++ unless $f->();
124 foreach my $i ( 0 .. $#_ ) {
133 foreach my $i ( reverse 0 .. $#_ ) {
140 sub insert_after (&$\@) {
141 my ($f, $val, $list) = @_;
144 foreach my $i ( 0 .. $#$list ) {
146 $c = $i, last if $f->();
151 @{$list}[ $c + 1 .. $#$list ],
152 ) and return 1 if $c != -1;
156 sub insert_after_string ($$\@) {
157 my ($string, $val, $list) = @_;
159 foreach my $i ( 0 .. $#$list ) {
161 $c = $i, last if $string eq $list->[$i];
166 @{$list}[ $c + 1 .. $#$list ],
167 ) and return 1 if $c != -1;
175 grep $started ||= do {
182 sub after_incl (&@) {
185 grep $started ||= $test->(), @_;
191 grep $more &&= ! $test->(), @_;
194 sub before_incl (&@) {
208 for ( $ix = $#_; $ix >= 0; $ix-- ) {
210 my $testval = $test->();
212 # Simulate $_ as alias
214 return $_ if $testval;
222 return $_ if $test->();
227 sub pairwise (&\@\@) {
230 # Symbols for caller's input arrays
231 use vars qw{ @A @B };
232 local ( *A, *B ) = @_;
235 my ( $caller_a, $caller_b ) = do {
238 \*{$pkg.'::a'}, \*{$pkg.'::b'};
241 # Loop iteration limit
242 my $limit = $#A > $#B? $#A : $#B;
244 # This map expression is also the return value
245 local( *$caller_a, *$caller_b );
247 # Assign to $a, $b as refs to caller's array elements
248 ( *$caller_a, *$caller_b ) = \( $A[$_], $B[$_] );
250 # Perform the transformation
255 sub each_array (\@;\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@) {
256 return each_arrayref(@_);
260 my @list = @_; # The list of references to the arrays
261 my $index = 0; # Which one the caller will get next
262 my $max = 0; # Number of elements in longest array
264 # Get the length of the longest input array
266 unless ( ref $_ eq 'ARRAY' ) {
268 Carp::croak("each_arrayref: argument is not an array reference\n");
270 $max = @$_ if @$_ > $max;
273 # Return the iterator as a closure wrt the above variables.
277 unless ( $method eq 'index' ) {
279 Carp::croak("each_array: unknown argument '$method' passed to iterator.");
282 # Return current (last fetched) index
283 return undef if $index == 0 || $index > $max;
287 # No more elements to return
288 return if $index >= $max;
291 # Return ith elements
292 return map $_->[$i], @list;
300 return splice @list, 0, $n;
304 sub mesh (\@\@;\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@) {
306 $max < $#$_ && ( $max = $#$_ ) foreach @_;
315 grep { not $seen{$_}++ } @_;
320 my $min = my $max = $_[0];
322 for ( my $i = 1; $i < @_; $i += 2 ) {
323 if ( $_[$i-1] <= $_[$i] ) {
324 $min = $_[$i-1] if $min > $_[$i-1];
325 $max = $_[$i] if $max < $_[$i];
327 $min = $_[$i] if $min > $_[$i];
328 $max = $_[$i-1] if $max < $_[$i-1];
334 if ($_[$i-1] <= $_[$i]) {
335 $min = $_[$i-1] if $min > $_[$i-1];
336 $max = $_[$i] if $max < $_[$i];
338 $min = $_[$i] if $min > $_[$i];
339 $max = $_[$i-1] if $max < $_[$i-1];
354 *first_index = \&firstidx;
355 *last_index = \&lastidx;
356 *first_value = \&firstval;
357 *last_value = \&lastval;
369 List::MoreUtils - Provide the stuff missing in List::Util
373 use List::MoreUtils qw{
374 any all none notall true false
375 firstidx first_index lastidx last_index
376 insert_after insert_after_string
378 after after_incl before before_incl
379 firstval first_value lastval last_value
380 each_array each_arrayref
382 mesh zip uniq distinct minmax part
387 B<List::MoreUtils> provides some trivial but commonly needed functionality on
388 lists which is not going to go into L<List::Util>.
390 All of the below functions are implementable in only a couple of lines of Perl
391 code. Using the functions from this module however should give slightly better
392 performance as everything is implemented in C. The pure-Perl implementation of
393 these functions only serves as a fallback in case the C portions of this module
394 couldn't be compiled on this machine.
400 Returns a true value if any item in LIST meets the criterion given through
401 BLOCK. Sets C<$_> for each item in LIST in turn:
403 print "At least one value undefined"
404 if any { ! defined($_) } @list;
406 Returns false otherwise, or if LIST is empty.
410 Returns a true value if all items in LIST meet the criterion given through
411 BLOCK. Sets C<$_> for each item in LIST in turn:
413 print "All items defined"
414 if all { defined($_) } @list;
416 Returns false otherwise, or if LIST is empty.
418 =item none BLOCK LIST
420 Logically the negation of C<any>. Returns a true value if no item in LIST meets
421 the criterion given through BLOCK. Sets C<$_> for each item in LIST in turn:
423 print "No value defined"
424 if none { defined($_) } @list;
426 Returns false otherwise, or if LIST is empty.
428 =item notall BLOCK LIST
430 Logically the negation of C<all>. Returns a true value if not all items in LIST
431 meet the criterion given through BLOCK. Sets C<$_> for each item in LIST in
434 print "Not all values defined"
435 if notall { defined($_) } @list;
437 Returns false otherwise, or if LIST is empty.
439 =item true BLOCK LIST
441 Counts the number of elements in LIST for which the criterion in BLOCK is true.
442 Sets C<$_> for each item in LIST in turn:
444 printf "%i item(s) are defined", true { defined($_) } @list;
446 =item false BLOCK LIST
448 Counts the number of elements in LIST for which the criterion in BLOCK is false.
449 Sets C<$_> for each item in LIST in turn:
451 printf "%i item(s) are not defined", false { defined($_) } @list;
453 =item firstidx BLOCK LIST
455 =item first_index BLOCK LIST
457 Returns the index of the first element in LIST for which the criterion in BLOCK
458 is true. Sets C<$_> for each item in LIST in turn:
460 my @list = (1, 4, 3, 2, 4, 6);
461 printf "item with index %i in list is 4", firstidx { $_ == 4 } @list;
463 item with index 1 in list is 4
465 Returns C<-1> if no such item could be found.
467 C<first_index> is an alias for C<firstidx>.
469 =item lastidx BLOCK LIST
471 =item last_index BLOCK LIST
473 Returns the index of the last element in LIST for which the criterion in BLOCK
474 is true. Sets C<$_> for each item in LIST in turn:
476 my @list = (1, 4, 3, 2, 4, 6);
477 printf "item with index %i in list is 4", lastidx { $_ == 4 } @list;
479 item with index 4 in list is 4
481 Returns C<-1> if no such item could be found.
483 C<last_index> is an alias for C<lastidx>.
485 =item insert_after BLOCK VALUE LIST
487 Inserts VALUE after the first item in LIST for which the criterion in BLOCK is
488 true. Sets C<$_> for each item in LIST in turn.
490 my @list = qw/This is a list/;
491 insert_after { $_ eq "a" } "longer" => @list;
494 This is a longer list
496 =item insert_after_string STRING VALUE LIST
498 Inserts VALUE after the first item in LIST which is equal to STRING.
500 my @list = qw/This is a list/;
501 insert_after_string "a", "longer" => @list;
504 This is a longer list
506 =item apply BLOCK LIST
508 Applies BLOCK to each item in LIST and returns a list of the values after BLOCK
509 has been applied. In scalar context, the last element is returned. This
510 function is similar to C<map> but will not modify the elements of the input
514 my @mult = apply { $_ *= 2 } @list;
515 print "\@list = @list\n";
516 print "\@mult = @mult\n";
521 Think of it as syntactic sugar for
523 for (my @mult = @list) { $_ *= 2 }
525 =item before BLOCK LIST
527 Returns a list of values of LIST upto (and not including) the point where BLOCK
528 returns a true value. Sets C<$_> for each element in LIST in turn.
530 =item before_incl BLOCK LIST
532 Same as C<before> but also includes the element for which BLOCK is true.
534 =item after BLOCK LIST
536 Returns a list of the values of LIST after (and not including) the point
537 where BLOCK returns a true value. Sets C<$_> for each element in LIST in turn.
539 @x = after { $_ % 5 == 0 } (1..9); # returns 6, 7, 8, 9
541 =item after_incl BLOCK LIST
543 Same as C<after> but also inclues the element for which BLOCK is true.
545 =item indexes BLOCK LIST
547 Evaluates BLOCK for each element in LIST (assigned to C<$_>) and returns a list
548 of the indices of those elements for which BLOCK returned a true value. This is
549 just like C<grep> only that it returns indices instead of values:
551 @x = indexes { $_ % 2 == 0 } (1..10); # returns 1, 3, 5, 7, 9
553 =item firstval BLOCK LIST
555 =item first_value BLOCK LIST
557 Returns the first element in LIST for which BLOCK evaluates to true. Each
558 element of LIST is set to C<$_> in turn. Returns C<undef> if no such element
561 C<first_val> is an alias for C<firstval>.
563 =item lastval BLOCK LIST
565 =item last_value BLOCK LIST
567 Returns the last value in LIST for which BLOCK evaluates to true. Each element
568 of LIST is set to C<$_> in turn. Returns C<undef> if no such element has been
571 C<last_val> is an alias for C<lastval>.
573 =item pairwise BLOCK ARRAY1 ARRAY2
575 Evaluates BLOCK for each pair of elements in ARRAY1 and ARRAY2 and returns a
576 new list consisting of BLOCK's return values. The two elements are set to C<$a>
577 and C<$b>. Note that those two are aliases to the original value so changing
578 them will modify the input arrays.
582 @x = pairwise { $a + $b } @a, @b; # returns 12, 14, 16, 18, 20
587 @x = pairwise { ($a, $b) } @a, @b; # returns a, 1, b, 2, c, 3
589 =item each_array ARRAY1 ARRAY2 ...
591 Creates an array iterator to return the elements of the list of arrays ARRAY1,
592 ARRAY2 throughout ARRAYn in turn. That is, the first time it is called, it
593 returns the first element of each array. The next time, it returns the second
594 elements. And so on, until all elements are exhausted.
596 This is useful for looping over more than one array at once:
598 my $ea = each_array(@a, @b, @c);
599 while ( my ($a, $b, $c) = $ea->() ) { .... }
601 The iterator returns the empty list when it reached the end of all arrays.
603 If the iterator is passed an argument of 'C<index>', then it retuns
604 the index of the last fetched set of values, as a scalar.
606 =item each_arrayref LIST
608 Like each_array, but the arguments are references to arrays, not the
611 =item natatime BLOCK LIST
613 Creates an array iterator, for looping over an array in chunks of
614 C<$n> items at a time. (n at a time, get it?). An example is
615 probably a better explanation than I could give in words.
619 my @x = ('a' .. 'g');
620 my $it = natatime 3, @x;
621 while (my @vals = $it->())
632 =item mesh ARRAY1 ARRAY2 [ ARRAY3 ... ]
634 =item zip ARRAY1 ARRAY2 [ ARRAY3 ... ]
636 Returns a list consisting of the first elements of each array, then
637 the second, then the third, etc, until all arrays are exhausted.
643 @z = mesh @x, @y; # returns a, 1, b, 2, c, 3, d, 4
647 @c = qw/zip zap zot/;
648 @d = mesh @a, @b, @c; # x, 1, zip, undef, 2, zap, undef, undef, zot
650 C<zip> is an alias for C<mesh>.
656 Returns a new list by stripping duplicate values in LIST. The order of
657 elements in the returned list is the same as in LIST. In scalar context,
658 returns the number of unique elements in LIST.
660 my @x = uniq 1, 1, 2, 2, 3, 5, 3, 4; # returns 1 2 3 5 4
661 my $x = uniq 1, 1, 2, 2, 3, 5, 3, 4; # returns 5
665 Calculates the minimum and maximum of LIST and returns a two element list with
666 the first element being the minimum and the second the maximum. Returns the
667 empty list if LIST was empty.
669 The C<minmax> algorithm differs from a naive iteration over the list where each
670 element is compared to two values being the so far calculated min and max value
671 in that it only requires 3n/2 - 2 comparisons. Thus it is the most efficient
674 However, the Perl implementation of it has some overhead simply due to the fact
675 that there are more lines of Perl code involved. Therefore, LIST needs to be
676 fairly big in order for C<minmax> to win over a naive implementation. This
677 limitation does not apply to the XS version.
679 =item part BLOCK LIST
681 Partitions LIST based on the return value of BLOCK which denotes into which
682 partition the current value is put.
684 Returns a list of the partitions thusly created. Each partition created is a
685 reference to an array.
688 my @part = part { $i++ % 2 } 1 .. 8; # returns [1, 3, 5, 7], [2, 4, 6, 8]
690 You can have a sparse list of partitions as well where non-set partitions will
693 my @part = part { 2 } 1 .. 10; # returns undef, undef, [ 1 .. 10 ]
695 Be careful with negative values, though:
697 my @part = part { -1 } 1 .. 10;
699 Modification of non-creatable array value attempted, subscript -1 ...
701 Negative values are only ok when they refer to a partition previously created:
703 my @idx = ( 0, 1, -1 );
705 my @part = part { $idx[$++ % 3] } 1 .. 8; # [1, 4, 7], [2, 3, 5, 6, 8]
711 Nothing by default. To import all of this module's symbols, do the conventional
713 use List::MoreUtils ':all';
715 It may make more sense though to only import the stuff your program actually
718 use List::MoreUtils qw{ any firstidx };
722 When C<LIST_MOREUTILS_PP> is set, the module will always use the pure-Perl
723 implementation and not the XS one. This environment variable is really just
724 there for the test-suite to force testing the Perl implementation, and possibly
725 for reporting of bugs. I don't see any reason to use it in a production
730 There is a problem with a bug in 5.6.x perls. It is a syntax error to write
733 my @x = apply { s/foo/bar/ } qw{ foo bar baz };
735 It has to be written as either
737 my @x = apply { s/foo/bar/ } 'foo', 'bar', 'baz';
741 my @x = apply { s/foo/bar/ } my @dummy = qw/foo bar baz/;
743 Perl 5.5.x and Perl 5.8.x don't suffer from this limitation.
745 If you have a functionality that you could imagine being in this module, please
746 drop me a line. This module's policy will be less strict than L<List::Util>'s
747 when it comes to additions as it isn't a core module.
749 When you report bugs, it would be nice if you could additionally give me the
750 output of your program with the environment variable C<LIST_MOREUTILS_PP> set
751 to a true value. That way I know where to look for the problem (in XS,
752 pure-Perl or possibly both).
756 Bugs should always be submitted via the CPAN bug tracker.
758 L<http://rt.cpan.org/NoAuth/ReportBug.html?Queue=List-MoreUtils>
762 Credits go to a number of people: Steve Purkis for giving me namespace advice
763 and James Keenan and Terrence Branno for their effort of keeping the CPAN
764 tidier by making L<List::Utils> obsolete.
766 Brian McCauley suggested the inclusion of apply() and provided the pure-Perl
767 implementation for it.
769 Eric J. Roode asked me to add all functions from his module C<List::MoreUtil>
770 into this one. With minor modifications, the pure-Perl implementations of those
773 The bunch of people who almost immediately pointed out the many problems with
774 the glitchy 0.07 release (Slaven Rezic, Ron Savage, CPAN testers).
776 A particularly nasty memory leak was spotted by Thomas A. Lowery.
778 Lars Thegler made me aware of problems with older Perl versions.
780 Anno Siegel de-orphaned each_arrayref().
782 David Filmer made me aware of a problem in each_arrayref that could ultimately
785 Ricardo Signes suggested the inclusion of part() and provided the
788 Robin Huston kindly fixed a bug in perl's MULTICALL API to make the
789 XS-implementation of part() work.
793 A pile of requests from other people is still pending further processing in
794 my mailbox. This includes:
798 =item * List::Util export pass-through
800 Allow B<List::MoreUtils> to pass-through the regular L<List::Util>
801 functions to end users only need to C<use> the one module.
805 Use code-reference to extract a key based on which the uniqueness is
806 determined. Suggested by Aaron Crane.
812 =item * random_item_delete_index
814 =item * list_diff_hash
816 =item * list_diff_inboth
818 =item * list_diff_infirst
820 =item * list_diff_insecond
822 These were all suggested by Dan Muey.
826 Always return a flat list when either a simple scalar value was passed or an
827 array-reference. Suggested by Mark Summersault.
837 Tassilo von Parseval E<lt>tassilo.von.parseval@rwth-aachen.deE<gt>
839 =head1 COPYRIGHT AND LICENSE
841 Copyright 2004 - 2010 by Tassilo von Parseval
843 This library is free software; you can redistribute it and/or modify
844 it under the same terms as Perl itself, either Perl version 5.8.4 or,
845 at your option, any later version of Perl 5 you may have available.