Christopher Yee Mon wrote:
> I have an array of strings whose members consist of a number followed by
> a comma followed by a text string
>
> e.g.
> 1,fresh
> 2,testurl
>
> I want to sort by descending numerical order according to the number
> part so I made this sort subroutine
>
> sub by_counter_field {
> my($a, $b) = @_;
> $a =~ s/^(.*?),.*/$1/g;
> $b =~ s/^(.*?),.*/$1/g;
>
> if ($a > $b) { -1 } elsif ($a < $b) { 1 } else { 0 }
> }
>
> and sorted it with
> sort by_counter_field @array;
> and it didn't work.
>
> I've also tried replacing
> if ($a > $b) { -1 } elsif ($a < $b) { 1 } else { 0 }
> with
> $b <=> $a
> and it still didn't work.
>
> I also tried this
> sort { ($b =~ /(.*?),.*/)[0] <=> ($a =~ /(.*?),.*/)[0] } @array;
> and it didn't work either.
>
> I can't figure it out for the life of me
>
> Any assistance would be deeply appreciated.
You misunderstand the way $a and $b work. It is an ugly mechanism, but when
'sort' calls your comparison subroutine the global variables $a and $b are
already set to the two values to be compared. What is even more hazardous is
that if you modify them as you have tried to do with s/// you will change the
values in the list you are sorting.
If the derivation is non-trivial then I usually write code that evaluates
lexical variables $aa and $bb from $a and $b, and then simply compare them with
<=> or 'cmp' as appropriate.
The code below is a working rewrite of your program. I hope it helps.
Rob
use strict;
use warnings;
no warnings 'qw';
my @data = qw/
99,tomato
12,pear
1,fresh
2,testurl
16,summary
0,zero
76,trombones
88,fatso
/;
sub by_counter_field {
my $aa = $a =~ /^(\d+)/ ? $1 : 0;
my $bb = $b =~ /^(\d+)/ ? $1 : 0;
$bb <=> $aa;
}
my @sorted = sort by_counter_field @data;
print map "$_\n", @sorted;
--
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
http://learn.perl.org/