If you think about how much data is passed between subroutines you should be able to see why...

Yes, I realize that. This came out of my toying with functional programming in perl. Most of the other solutions I've seen here have side effects that I was trying (for intellectual, rather than practical reasons) to avoid.

Well a random tip then. Perl does particularly poorly when faced with deep recursion. However it certainly supports functional techniques, and I refer to several examples on my home node.

BTW I already gave a solution to a generalization of the original problem which is O(n) and has no side-effects at Re (tilly) 1: interleaving lists.

my @interleave = mapcar { @_ } \@list1, \@list2, \@list3;
[download]

sub interleave {
  my $y;
  while (@{ $_[0] } or @{ $_[1] }) {
      push @$y, shift @{ $_[0] } if @{ $_[0] };
      push @$y, shift @{ $_[1] } if @{ $_[1] };
  }
  return $y;
}
[download]

Ciao,
Gryn

Update: Eek, this doesn't work. Please ignore -- I'm at work and will be a while to fix.

Update: Ok that seems better. Thank you for your patience :)

Shockingly (for me to see this question here), I just solved this problem two days ago.

Try this:

   @foo = ( 1 , 3 , 5 , 7 );
   @bar = ( 2 , 4 , 6 , 8 );

   @baz = map { ( $_ , shift @bar ) } @foo;
[download]

It suffers from losing the end of @bar if length( @bar ) > length ( @foo ) though.

Scott, I tried your code on a similar problem that I had
print STDOUT my @foobar = map { ( $_, shift @foo ) } @bar;
and it worked, but when I turn on use warnings; I get the following error:
Use of uninitialized value in print at line 3
I've done some research and it seems that some people just complain about using warning and turn them off (I don't want to do that), and other say to re-work the code into a while loop. Anyone have any advice as to how I can keep the code while getting rid of the error?

Hmmm ... this works for me with no comments from perl:

use strict;
use warnings;

my @foo = ( 1 , 3 , 5 , 7 );
my @bar = ( 2 , 4 , 6 , 8 );

print STDOUT my @baz = map { ( $_ , shift @bar ) } @foo;
[download]

I originally had a problem with `uninitialized' warnings when trying to reproduce your problem but it turned out to be because I had

[ ... stuff ... ]
my @baz = map { ( $_ , shift @bar ) } @foo;

print STDOUT my @baz = map { ( $_ , shift @bar ) } @foo;
[download]

This, of course, eats up @bar the first time so there's nothing left to shift the second time. :(

What's on your lines 1 and 2?

As for the general problem of warnings, I only turn warnings off for short one-shot scripts that I know will work even though I wrote bad code. Or, rather, I ignore them unless they get in the way of the output.

I've never encountered a warning (from perl) that wasn't justified so if I'm doing anything at all important I eliminate them.

This can also be used to interleave arrays, though it is more suited to iterating through them rather than just building the interleaved list (also, take note of chipmunk's correction to handle parallel undefs).

This uses a hash slice to achieve the same outcome:

#!/usr/bin/perl -w

use strict;

my @keys   = qw(a b c d);
my @values = qw(1 2 3 4);

print interleave(\@keys, \@values);

sub interleave {
  my %interleaved;
  @interleaved{ @{$_[0]} } = @{$_[1]};
  return %interleaved;
}
[download]

That only works if you consider "interleaving" to mean "throw away any pairs where the first array has duplicate values" and "sort the merged pairs in a pseudo-random order". (:

        - tye (but my friends call me "Tye")

Thank you for pointing my mistake in using a hash slice to interleave two lists, tye. Here's my second attempt at interleaving two lists. This time I use an array slice, and (hope this makes sense) calculates the offset position for the two arrays in the final interleaved array:

#!/usr/bin/perl -w

use strict;
use Data::Dumper qw(Dumper);

my @keys   = qw(a b c d e f g h i j); #is longer than
my @values = qw(1 2 3 4 5);           #this..

my @array = interleave(\@keys, \@values);
print Dumper(\@array); #let's see what happens

sub interleave {
  my $array1 = shift;
  my $array2 = shift;

  my(@even, @odd);
  push @even, $_ * 2     for 0..$#$array1;
  push @odd,  $_ * 2 + 1 for 0..$#$array2;

  my @interleaved;
  @interleaved[@even, @odd] = (@$array1, @$array2);

  return wantarray ? @interleaved : \@interleaved;
}

__END__
[download]

If one of the arrays is longer than the other, it will still interleave them as if they are the same length. This behaviour is essential if the interleaved array ever gets copied to a hash, or you need the array elements to be in pairs.

The method that I used to calculate the offsets inside interleave() is the fastest I could find through benchmarking. If there is a faster/cleaner/better way, I would love to hear it.

Kind of a cop-out but the mesh function in List::MoreUtils is a turnkey solution:

http://stackoverflow.com/questions/38345/is-there-an-elegant-zip-to-interleave-two-lists-in-perl-5

my @outies = mesh (@lines1,@lines2, @lines3);
[download]

#the following gives: a 1 b 2 c 3

@arr1 = qw(a b c);
@arr2 = qw(1 2 3);

while (@arr1) {push @arr3, shift @arr1,shift @arr2}

print "@arr3";



#BUT why does the following give an infinite loop?


@arr1 = qw(a b c);
@arr2 = qw(1 2 3);

while ($push @arr3, shift @arr1,shift @arr2){}
[download]

See the documentation of push:

Returns the number of elements in the array following the completed "push"

Also, remove the dollar sign before "push".

لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ

thanks. how do i edit an item i've already posted?