in reply to Re^2: how to sum over rows based on column headings in perl
in thread how to sum over rows based on column headings in perl

That error is pretty clear! You are missing the right curly bracket closing the while loop you opened on line 26. The end of line 36 is given as the location of the error because that's where Perl runs out of code in which to find the closing bracket.

You have a stray opening bracket under the line you commented out at line 27. This is a good example of why you should have consistent syntax: your commented-out while statement has the bracket on the next line, while the new while statement has the bracket on the same line. Pick one and stick with it. In my view, putting the brackets on the same line reduces the chances of them going astray.

Update: explain cause of error

The way forward always starts with a minimal test.

Replies are listed 'Best First'.
Re^4: how to sum over rows based on column headings in perl
by angerusso (Novice) on Jul 31, 2015 at 19:07 UTC

    Actually, I think I did reply too fast! Obviously i don't understand map function the way you used it.

    Could you explain your usage of map? E.g. I changed "M" to "MUTS" and the results are screwed up. Wherever you used "M", I changed it to "MUTS" and changed "M" in datafile to "MUTS" and I don't get the expected result. 

    map { $_ eq $label ? 'MUTS' : '_' } @headers[1..$#headers];

    [download]

    My modified code gives wrong results 

    #!/usr/bin/perl

use strict;
use warnings;
use autodie;

if (@ARGV != 1){
    print "USAGE: ./parse-counts.pl file\n";
    exit(1);
}

my $mutfile = $ARGV[0];
open(INPUTR,"<$mutfile") or die "Can't open \$mutfile for reading. \n"
+;

my (%counts, %unique, %masks);
my ($headname, @unique) = grep !$unique{$_}++, my @headers = split /\t
+/, <INPUTR>;

# the basic syntax is @out = map { CODE } @in;

for my $label ( @unique )
{
    $masks{$label} = join "\t",
    map { $_ eq $label ? 'MUTS' : '_' } @headers[1..$#headers];
}

my $line;

while($line=<INPUTR>)
#while(<DATA>)
{
    chomp $line;
    $line =~ s/\t/\t/g; # for uniform spacing
    my ($name, $letters) = split /\t/, $line, 2;
    $counts{$name}{$_} += ($masks{$_} | $letters) =~ /MUTS/ for @uniqu
+e;
    print $name."\n";
    print $letters."\n";
}

print "$headname @unique\n";
print "$_ @{ $counts{$_} }{@unique}\n" for sort keys %counts;


The output produced is:
Gname G1 G2 G3

A 3 2 0
B 2 1 1
C 2 0 0

The modified datafiles is
Gname   G1      G1      G1      G1      G2      G2      G3
A       W       W       MUTS    W       W       W       MUTS
A       W       W       W       W       W       W       W
A       W       W       W       W       W       W       W
B       W       W       W       W       W       MUTS    MUTS
B       MUTS    W       W       W       W       MUTS    MUTS
C       MUTS    MUTS    MUTS    W       W       W       W
C       MUTS    W       W       MUTS    MUTS    W       W

    [download]
[reply]
[d/l]
[select]

      No, the heart of the algorithm is the or operator, and changing M to MUTS breaks alignment that allows the or operator to work properly.

      Print out both sides of the or operator to see the misalignment.

[reply]

        Hmmm ... I am not very comfortable with interpreting $_. I don't know how to print $masks{$_} to understand what it is. That is my problem. I know I am dumb! 

            print $name."\n";
    print $letters."\n";
    print $masks{$_}."\n";

        [download]
        which gives obvious error:
A
W       W       M       W       W       W       M
Use of uninitialized value $_ in hash element at ./parse-counts.pl lin
+e 37, <INPUTR> line 2.
Use of uninitialized value within %masks in concatenation (.) or strin
+g at ./parse-counts.pl line 37, <INPUTR> line 2.

        [download]
[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom