difficulty in multi sorting

[texuser74]

Dear Monks

I want to do two types of sorting in the follwoing input data, e.g.

Input:<test.in>

Xu, Guo Qin B-5-OR9

Zhang, Yong Ping B-5-OR10

Xu, Guo Qin B-5-OR10

Tsuneyuki, Shinji B-6-IN11

Xu, Guo Qin B-6-IN12

1st sort: Sort the names, and then combine them:

open(IN, '<', "test.in");
open(OUT, '>', "test.out");
while (<IN>) {
  chomp;
  next unless /\S/;

  my ($name, $num) = /(.*)\s+(B-.*)/;

  $data{$name}{$num} = 1;
}

foreach my $name (sort keys %data) {
  print OUT "$name ";
  print OUT join ', ', sort { $a cmp $b } keys %{$data{$name}};
  print OUT "\n\n";
}

output:

Tsuneyuki, Shinji B-6-IN11

Xu, Guo Qin B-5-OR10, B-5-OR9, B-6-IN12

Zhang, Yong Ping B-5-OR10

2nd Sort: This is where i am having problem, now i want to sort the codes numerically inside each name, i.e. the required output is

Tsuneyuki, Shinji B-6-IN11

Xu, Guo Qin B-5-OR9, B-5-OR10, B-6-IN12

Zhang, Yong Ping B-5-OR10

please help me in doing this.

best regards,
raj

[jdporter]

So the essence of your problem is how to sort these funky codes. Well, it would help if you could give a precise definition of their format, rather than making us guess. I'm guessing that the first number (the 5 and 6 in your examples) could be more than one digit wide, and that there are always exactly two capital letters in the third part.

sub codes_sorted
{
    # do the GRT dance
    map { substr $_, 10 }
    sort
    map { /-(\d+)-([A-Z]{2})(\d+)/; pack('N',$1).$2.pack('N',$3).$_ }
}

Then you can do

  print OUT join ', ', codes_sorted( keys %{$data{$name}} );

[pelagic]

texuser74,
you probably want to get the data into some structure before processing it. So why don't you read through your file and 'feed' a Hash of Array in a way like:

...
my %hash;
while (<IN>) {
    my $name = ...
    my $code = ...
    push @{ $hash{$name} }, $code;
}

Then you can loop through the sorted hash keys and print the name together with all it's (sorted) codes.

foreach my $name (sort keys %hash) {
    print join (', ', $name, sort @{ $hash{$name} }), "\n";
}

That's some sort of pseudo code, so it will not work as is. .. ;)