removing duplicate letters

eduardo has asked for the wisdom of the Perl Monks concerning the following question:

ok, a friend of mine is taking a perl/cgi class in school and she had a problem posed to her. using perl, how do you make duplicate letters collapse to a single letter? example: let us say you have the word 'letters' it should colapse to 'leters' and 'foo' should collapse to 'fo' but 'fooo' should be 'foo' because the 3rd o isn't a duplicate... am i making myself clear? i am sure there is some neato tr// or s// that i could have used, however i am pitiful with regexp's so the best i could come up with was:

#!/usr/bin/perl -w

use strict;

while (<>) {
    my $temp = '';
    print join ('',
        map { 
            if ($_ eq $temp) {
                $temp = '';
            } else {
                $temp = $_;
            }
        } split('', $_)
    );
}
[download]

anyone have any better ideas? i would love to see some other cool perlish ways of doing this...

Comment on removing duplicate letters Download Code

Replies are listed 'Best First'.
Re: removing duplicate letters by plaid (Chaplain) on Jun 18, 2000 at 23:50 UTC
This seems to work: `s/(.)\1/$1/g;` [download] The tr operator does what you want, except it squashes all duplicates, not just two. For reference though: `tr/a-zA-Z//s;` [download]	[reply] [d/l] [select]
RE: Re: removing duplicate letters by Ovid (Cardinal) on Jun 19, 2000 at 00:50 UTC
Actually, the dot operator doesn't fit his definition. eduardo specified letters so we should change it to `s/(\w)(\1)/$1/g;` [download] Update: Oops! takshaka exposed me for the fake I am! (see reply below)	[reply] [d/l]
RE: RE: Re: removing duplicate letters by takshaka (Friar) on Jun 19, 2000 at 01:53 UTC
If we're going to be pedantic, \w matches nonletters as well. Use a character class with the appropriate definition of "letters". `s/([a-zA-Z])\1/$1/g;`	[reply] [d/l]
RE: RE: RE: Re: removing duplicate letters by mdillon (Priest) on Jun 19, 2000 at 05:39 UTC
RE: Re: removing duplicate letters by eduardo (Curate) on Jun 19, 2000 at 00:01 UTC
damn it ;) that's it, that's what i wanted... `s/(.)\1/$1/g;` [download] oh well, what i kept on trying was: `s/(.)\$1/$1/g;` [download] and that sure as hell didn't work ;) thank you very much!	[reply] [d/l] [select]
Re: removing duplicate letters by chromatic (Archbishop) on Jun 19, 2000 at 01:45 UTC
Here's a different approach. I may benchmark this later: `$newword = join '', split(/(\w+)\1/, $word);` [download] Whoops, missed the + the first time! Let's say the first was a warm up, and the second actually works: `$newword = join '', split(/((\w)+)\1(\2{0,1})/, $word);` [download] There. Appropriately hard to read. Update: Yes, I had the backreferences backward, which breaks things, as the Anonymous Monk points out below. That'll teach me to be crafty.	[reply] [d/l] [select]
RE: Re: removing duplicate letters by Anonymous Monk on Jun 19, 2000 at 02:30 UTC
Trying the latest suggestion... $s = "foooobaaaarrr"; print "Before: $s\n"; $newword = join '', split(/((\w)+)\2(\1{0,1})/, $s); print "After: $newword\n"; Result: no change. Not hereabouts, anyway.	[reply]
(jcwren) RE: removing duplicate letters by jcwren (Prior) on Jun 18, 2000 at 23:53 UTC
It seems like your rule needs some clarification before this problem could be solved. What dictatates that 'fooo' should be collapsed to 'foo', but not 'fo'? What does 'foooo' collapse to? 'foo' or 'fooo'? --Chris	[reply]