Use of uninitialized value $data

[ImJustAFriend]

Hello monks, I am in need of assistance please. I have a file I am parsing. At the end of parsing, I am left with lines that look like this:

label1=data1,label2=data2, ...

There can be a variable number of label/data pairs on each line. My code for parsing these lines:

my @data = split(',', $line);
foreach my $datapair (@data) {
   my ($label, $data) = split('=', $datapair);
   print "label: $label; data: $data\n";
   print OUT1 "$data;";
}

When I run the code, I get on my stderr

Use of uninitialized value $data in concatenation (.) or string at ./t
+est.pl line 124, <IN4> chunk 2147.
Use of uninitialized value $data in concatenation (.) or string at ./t
+est.pl line 125, <IN4> chunk 2147.

I can see in my debug logs (further up, code not shown) that I am getting @data loaded appropriately with lines like my sample lines above. So why is $data uninitialized? This has been driving me nuts for a couple of hours now. Thanks all!

[Corion]

My hint is to not trust your logging further up and to move the logging to where the problem is. If $data is undef, then $datapair does not contain a =.

So, investigate $datapair and the contents of @data just before your loop.

Maybe/most likely, $line is not what you think it is. Maybe it ends with a comma and a newline, and split helpfully keeps that newline as its own element in @data?

[ImJustAFriend]

Thanks for the input! I just ran it again. Here is what $line (one example) looks like in my debug log (obscured for privacy):

"line: label1=N,label2=N,label3=N,label4=NNNNNNNNNNN,label5=N,label6=W
+WWN"

So it looks as though $line is populating properly... hmmmm....

[Corion]

Then please show us the relevant code+data that produces the output.

I can't reproduce the problem using the following, very short program:

#!perl
use strict;
use warnings;

my $line = "line: label1=N,label2=N,label3=N,label4=NNNNNNNNNNN,label5
+=N,label6=WWWN";

my @data = split(',', $line);
foreach my $datapair (@data) {
   my ($label, $data) = split('=', $datapair);
   print "label: $label; data: $data\n";
   #print OUT1 "$data;";
}

__END__
label: line: label1; data: N
label: label2; data: N
label: label3; data: N
label: label4; data: NNNNNNNNNNN
label: label5; data: N
label: label6; data: WWWN

... to me, that means, you are either doing something differently in your code that you haven't shown, or your data is something else.

[Marshall]

Perhaps use a hash?

#!/usr/bin/perl
use strict;
use warnings;


my $line ="line: label1=N,label2=N,label3=N,label4=NNNNNNNNNNN,label5=
+N,label6=WWWN";

my %values = $line =~ /(\w+)=(\w+)/g;

foreach my $label (sort keys %values)
{
    print "$label $values{$label}\n";
}

__END__
label1 N
label2 N
label3 N
label4 NNNNNNNNNNN
label5 N
label6 WWWN

[LanX]

Did you have a look at chunk 2147?

Why don't you print $line if not defined $data; ?

Even better use Data::Dumper to check for unexpected characters.

[Marshall]

Look at line 2147 of the data.
This could be a blank line?

Perhaps a regex is better?

#!/usr/bin/perl
use strict;
use warnings;
$|=1; #turn of buffering for STDERR


while (my $line = <DATA>)
{
   next unless $line =~ /\S/; #skip blank lines
   
   my ($label, $data) = $line =~ /\s*(\w+)\s*=\s*(\w+)/;
   die "bad line: $line" unless defined $data;
   
   print "label $label is: $data\n";
   
}

=prints
label a is: 3
label b is: 4
label c is: 55
label d is: 99
bad line: x=

Process completed with exit code 255

=cut


__DATA__
a = 3

b=4
  c = 55       
  d=99
x=

[haj]

I was about to make the same suggestions as Corion did, only I am typing slower. Here's one more possible input: If you have an empty input line, i.e. a line which contains only a newline "\n", you get the same warnings.

And, of course, you should closely inspect chunk 2147 of your input file.

[hippo]

As pretty much everyone else has said now the problem is in your data - the stuff you are not showing us. See How to ask better questions using Test::More and sample data for rationale and instructions to produce an SSCCE like this one.

use strict;
use warnings;

use Test::More tests => 1;
use Test::NoWarnings;

my $line = 'foo=a,bar=3,baz=Fizzlegrapplebaum';
my @data = split(',', $line);
foreach my $datapair (@data) {
    my ($label, $data) = split('=', $datapair);
    my $out = "label: $label; data: $data\n";
}

Of course, here the test succeeds because the initial data is good. Your task is to make this test (or one very similar to it) fail by supplying the duff value for $line.

[ImJustAFriend]

Thanks for all the thoughts on the chunk. Unfortunately I'm getting this pair of errors for every input chunk in the input file... :(

[Marshall]

Run my code and see what it does.
Try this as an alternative:

foreach my $datapair (@data) {
   my ($label, $data) = split('=', $datapair);
######
   die "Bad Line: $datapair" if !defined ($data);
#####
   print "label: $label; data: $data\n";
   print OUT1 "$data;";
}

[Laurent_R]

Then please show the content of the @data array using Data::Dumper.

[Marshall]

please show a few lines of the input file that demonstrates the problem.

[ImJustAFriend]

Thank you everyone. I found out today they had me looking at the wrong log file, so this code is no longer being used...