average of three columns by specific number in a row using perl - perl

Question

i have an input file with following 5 columns and i want to average the column numbers 3, 4 ,5 individually 3, 4, 5 till its 2nd column value is 5 and similarly for 2nd column value 7 and 2.
PHE 5 2 4 6
PHE 5 4 6 4
PHE 5 4 2 8
TRP 7 5 5 9
TRP 7 5 7 1
TRP 7 5 7 3
TYR 2 4 4 4
TYR 2 4 4 0
TYR 2 4 5 3
and i want an output like this
PHE 5 3.3 4 6
TRP 7 5 6.3 4.3
TYR 2 4 4.3 2.3

Answer 1

perl -lane'
$k = join "\t", splice(#F, 0, 2);
$h{$k}{c}++ or push(#r, $k);
$h{$k}{t}[$_] += $F[$_] for 0 .. $#F;
END {
$, ="\t";
for (#r) {
($t, $c) = #{$h{$_}}{"t", "c"};
print $_, map sprintf("%.1f", $_/$c)*1, #$t;
}
}
' file
output
PHE 5 3.3 4 6
TRP 7 5 6.3 4.3
TYR 2 4 4.3 2.3

Answer 2

Nice solution mpapec.
I started the the following solution as an experiment to see if I could code something that would only require a single for loop, with no END block needed. It devolved into 6 for loops instead and a perfect example of how never to code unless your goal is obfuscation.
Yep, it uses an external module. Yes, it's the stupidist code I'll ever post (I hope). But at the very least it might get a chuckle out of someone. And yep, it works! :)
use Array::Transpose;
use List::Util qw(sum max);
use strict;
use warnings;
my $g;
my $l;
print "$_\n" for map {
join ' ', map {sprintf "%-$_->[0]s", $_->[1]} transpose [$l, $_]
} grep {
$l = [map {max #$_} transpose [[map {length $_} #$_], $l || ()]]
} [qw(Txt Num Ave Ave Ave)], map {
my #c = transpose $_;
[$c[0][0], $c[1][0], map {map {/\./ ? sprintf("%.1f", $_) : $_} sum(#$_) / #$_} #c[2..$#c]]
} map {
$g && $g->[0][0] eq $_->[0] ? (push #$g, $_) && () : ($g = [$_])
} map {[split]} (<DATA>);
__DATA__
PHE 5 2 4 6
PHE 5 4 6 4
PHE 5 4 2 8
TRP 7 5 5 9
TRP 7 5 7 1
TRP 7 5 7 3
TYR 2 4 4 4
TYR 2 4 4 0
TYR 2 4 5 3
Outputs
Txt Num Ave Ave Ave
PHE 5 3.3 4 6
TRP 7 5 6.3 4.3
TYR 2 4 4.3 2.3

Answer 3

Script without using modules.
Try this....
#!/usr/bin/env perl
open(DATA, "<input.txt") or die "Couldn't open file file.txt, $!";
my %h=();
my %c=();
print "\n";
while(<DATA>){
my $temp=$_;
if($temp=~m/^([A-Z]{3})\s+([\d]+)\s+([\d]+)\s+([\d]+)\s+([\d]+)/is)
{
my $key=$1;
$h{$key}{1} +=$2;
$h{$key}{2} +=$3;
$h{$key}{3} +=$4;
$h{$key}{4} +=$5;
if($c{$key})
{
$c{$key}++;
}
else
{
$c{$key}=1;
}
}
}
foreach $key (sort(keys %h)) {
#print $key.'='.$h{$key}{1}/$c{$key}." ".$h{$key}{2}/$c{$key}." ".$h{$key}{3}/$c{$key}." ".$h{$key}{4}/$c{$key};
printf("%s %d %.1f %.1f %.1f", $key, $h{$key}{1}/$c{$key},$h{$key}{2}/$c{$key},$h{$key}{3}/$c{$key},$h{$key}{4}/$c{$key});
print "\n";
}
print "\n";
close(DATA);
______OUTPUT________
PHE 5 3.3 4.0 6.0
TRP 7 5.0 6.3 4.3
TYR 2 4.0 4.3 2.3