What have you tried? Does it have to be the same file (not sure if that's a requirement or not)?

Personally, I'd use DBD::CSV. And then the whole thing becomes a trivial SQL query from your source file and a simple SQL insert to your output file. It is a bit more to set up, but since I set up DBD::CSV from time to time (I use it in a few projects), it becomes a bit more mundane (normal or ordinary). And I avoid a whole lot of work by letting SQL::Statement and friends do all my work for me. And then it's also an easier transition to a real database (MySQL, PostgreSQL, DB2, Oracle, SQLServer, or a more-real database like SQLite) as your needs scale.

#!/usr/bin/perl

 use Text::ParseWords;                                                
+   
use Data::Dumper;                                                     
+ 

                                                                      

use IO::File;                                                         

#my $file = test.csv;                                                 
+ 
my $file  = $ARGV[0];                                                 
my ($output_file) = ($file =~ /(.*)\./);                              
+ 
                                                            
open(my $fh, '<', $file) or die "Can't read csv file '$file' [$!]\n"; 
+ 

my $countlines = 0;   

my $countlines1 = 0;                                                  

while (my $line = <$fh>)
{while (my $line = <$fh>) 
{  chomp $line;
    my @fields = Text::ParseWords::parse_line(',', 0,$line);          
+ 
 if($countlines >= 0  && @fields[0]==201120)
{if ( @fields[1]==1){ $countlines1 = $countlines1 + 1;} 


$countlines++; 
    
}                                                     
                                                        
}

print "$countlines\n";
print "$countlines1\n"; 


$yield = ($countlines1 / $countlines)*100 ;

print $yield;                                 

 }
[download]

Thank for your reply,it does not have to be the same csv file. Currently , my code able read one week only and print it volume and yield. I need some guides from you all.

First, use Text::CSV_XS instead of Text::ParseWords - it reads more like the spec that way (having your code match your spec is always good in my books). Second, I don't see any commas in your original data, so there is an obvious mismatch.

Next, you are hardcoding your week number in your code. Don't do that. Accumulate it in a hash. I would even go so far as use a hash of hashes with separate fields for the sum and count.

Also, learn to use proper indentation. Your code is quite simply a mess. It's very difficult to follow nesting, and would get impossible to follow fairly quickly.

I also see a while inside a while, both reading from the same file. That's nonsensical. Instead, just read and discard a single line: scalar <$fh> should do it.

my %data;
my $csv = Text::CSV_XS->new();
scalar <$fh>;
while (my $row = $csv->getline($fh))
{
  $data{$row->[0]}{sum} += $row->[1];
  $data{$row->[0]}{count}++;
}

for my $weeknum (sort keys %data)
{
  printf "%s %d %.1f\n", $weeknum,
                         $data{$weeknum}{sum} / $data{$weeknum}{count}
+,
                         $data{$weeknum}{count};
}
[download]

That should give you enough to go on.

Update: s/comments/comma/ in the first paragraph

$ echo " 201120 1 05-10-2011_16:47
201120 0 05-10-2011_16:48
201120 1 05-10-2011_16:49
201121 0 05-10-2011_15:47
201121 1 05-10-2011_15:48
201121 1 05-10-2011_15:49
201121 1 05-10-2011_15:50 " | perl -ane'$c{$F[0]}{T}++;$c{$F[0]}{C}+=$
+F[1]}{printf"%s %d %.1f%%\n",$_,$c{$_}{T},$c{$_}{C}/$c{$_}{T}*100 for
+ sort keys%c'
201120 3 66.7%
201121 4 75.0%
[download]

hi thank for your reply, actually I am not really understand your code, I am just a beginner in learning perl. can i know what if i have a csv file like below?

 week    power   test     Test_end_date_time
201120   5        1       05-10-2011_16:47
201120   699      17      05-10-2011_16:48
201120   454      1       05-10-2011_16:49
201121   5687     26      12-10-2011_15:47
201121   211      1       12-10-2011_15:48
201121   1231     36      12-10-2011_15:49
201121   171      1       12-10-2011_15:50
201121   677      57      12-10-2011_15:51
201121   178      1       12-10-2011_15:52
 
if test = 1 mean pass, other than 1 is fail

I wish the output in another csv is like below, any guides for me?

week   volume    yield
201120 3         66.7%
201121 6         50.0%
[download]

$ echo " week    power   test     Test_end_date_time
201120   5        1       05-10-2011_16:47
201120   699      17      05-10-2011_16:48
201120   454      1       05-10-2011_16:49
201121   5687     26      12-10-2011_15:47
201121   211      1       12-10-2011_15:48
201121   1231     36      12-10-2011_15:49
201121   171      1       12-10-2011_15:50
201121   677      57      12-10-2011_15:51
201121   178      1       12-10-2011_15:52
" | perl -ane'/^\D/&&next;$c{$F[0]}{T}++;$c{$F[0]}{C}+=$F[2]==1}{print
+f"%s %d %.1f%%\n",$_,$c{$_}{T},$c{$_}{C}/$c{$_}{T}*100 for sort keys%
+c'
201120 3 66.7%
201121 6 50.0%
[download]

PerlMonks FAQ, How do I post a question effectively?, Is PerlMonks a good place to get answers for homework?