Perl: Inserting values into specific columns of CSV file - arrays

Question

I have CSV data of the form:
S.No,Label,Customer1,Customer2,Customer3...
1,label1,Y,N,Y
2,label2,N,Y,N
...
I need to reproduce the "label" to the left of "customer" columns marked with Y - and have nothing ("") to the left of columns marked with N.
Expected output:
S.No,Label,Customer1,Customer1,Customer2,Customer2,Customer3,Customer3...
1,label1,label1,Y,"",N,label1,Y
2,label2,"",N,label2,Y,"",N
When opened using Excel, it would look like this:
S.No Label Customer1 Customer1 Customer2 Customer2 Customer3 Customer3...
1 label1 label1 Y N label1 Y
2 label2 N label2 Y N
The two leftmost columns, referring to S.No and the original "Label" column, are constant.
What is the simplest way to do this? I tried the following code:
use strict;
use warnings;
my $nonIncludesFile = "nonIncludes.csv";
open(my $xfh, "+>", $nonIncludesFile) or warn "Unable to open $nonIncludesFile, $!";
chomp( my $header = <$xfh> );
my #names = split ",", $header;
my #names1;
my #fields;
my #fields1;
for(my $j=0; $j< scalar(#names); $j++)
{
$names1[$j] = $names[$j];
}
while(<$xfh>)
{
my $nonIncLine = $_;
$nonIncLine = chomp($nonIncLine);
#fields = split ",", $nonIncLine;
next if $. == 1; #skip the first line
for(my $i = 0; $i < scalar(#fields) -2; $i++) #Number of "customers" = scalar(#fields) -2
{
$fields1[0] = $fields[0];
$fields1[1] = $fields[1];
if('Y' eq $fields[ $i + 2 ])
{
$fields1[$i+2] = 'Y';
substr(#fields1, $i + 1, 0, $fields[1]); #insert the label to the left - HERE
}
else
{
$fields1[$i+2] = 'N';
substr(#fields1, $i + 1, 0, "");
}
}
}
print $xfh #names1;
print $xfh #fields1;
close($xfh);
This however complains of "substr outside of string" at the line marked by "HERE".
What am I doing wrong? And is there any simpler (and better) way to do this?

Answer 1

Something like this maybe?
#!/usr/bin/perl
use strict;
use warnings;
#read the header row
chomp( my ( $sn, $label, #customers ) = split( /,/, <DATA> ) );
#double the 'customers' column headings (one is suffixed "_label")
print join( ",", $sn, $label, map { $_ . "_label", $_ } #customers ), "\n";
#iterate data
while (<DATA>) {
#strip trailing linefeed
chomp;
#extract fields with split - note breaks if you've quoted commas inline.
my ( $sn, $label, #row ) = split /,/;
print "$sn,$label,";
#iterate Y/N values, and either prints "Y" + label, or anything else + blank.
foreach my $value (#row) {
print join( ",", $value eq "Y" ? $label : "", $value ),",";
}
print "\n";
}
__DATA__
S.No,Label,Customer1,Customer2,Customer3
1,label1,Y,N,Y
2,label2,N,Y,N
Assumes you don't have any fruity special characters (e.g. commas) in the fields, because it'll break if you do, and you might want to consider Text::CSV instead.

Answer 2

It is always much better to post some usable test data than write a something like this question
However, it looks like your data has no quoted fields or escaped characters, so it looks like you can just use split and join to process the CSV data
Here's a sample Perl program that fulfils your requirement. The example output uses your data as it is. Each line of data has to be processed backwards so that the insertions don't affect the indices of elements that are yet to be processed
use strict;
use warnings 'all';
use feature 'say';
while ( <DATA> ) {
chomp;
my #fields = split /,/;
for ( my $i = $#fields; $i > 1; --$i ) {
my $newval =
$. == 1 ? $fields[$i] :
lc $fields[$i] eq 'y' ? $fields[1] :
'';
splice #fields, $i, 0, $newval;
}
say join ',', #fields;
}
__DATA__
S.No,Label,Customer1,Customer2,Customer3...
1,label1,Y,N,Y
2,label2,N,Y,N
output
S.No,Label,Customer1,Customer1,Customer2,Customer2,Customer3...,Customer3...
1,label1,label1,Y,,N,label1,Y
2,label2,,N,label2,Y,,N