You don't show the value of $csv_setup, but it needs to include binary => 1, also make sure that $/ = "\r\n"; before you open the feed. That should be enough.

Parse::CSV uses Text::CSV_XS as the underlying parser. So you can look at the documentation there for which parameters are available to you.

hi thanks for replying binary => 1 is included.. where do i exactly include. $/ = "\r\n"; i added it before if ($fh->open($feed)) { but it makes no difference

If you create an example file that contains one good record, and one problematic record, then you'll have a test case that is easier to Dump and share here. Below is a little example that works here and shows an embedded newline being handled correctly by Parse::CSV:

use Parse::CSV;
use Data::Dump 'pp';

my $fh = new IO::File('failing.csv', 'r');

my $fail = do { local $/; <$fh> };
pp $fail;    # Print input
$fh->seek(0,0);

$/ = "\r\n";
my  $parser = Parse::CSV->new(
      handle => $fh,
      csv_attr => { binary => 1 },
);
print pp $_  while $_ = $parser->fetch; # Print output
[download]

Which prints:

"a,b,c,d,e,f,g\r\na,b,c,\"kkkk\n\",d,e,f,g\r\n"
["a" .. "g"]["a", "b", "c", "kkkk\n", "d" .. "g"]
[download]

I have managed to diagnose there is LF in the middle of line and the normal CRLF in the end. how can i remove this LF.

You can use a negative look-behind assertion to remove each LF that is not immediately preceded by a CR:

#! perl
use strict;
use warnings;

my $string = qq{"Tap Rackmount for 3 units\x0A","USR4500-RMK         "
+\x0D\x0A};

print "Before:\n>>>$string<<<\n";

#    Dec Hex  Oct
# LF  10  0A  012
# CR  13  0D  015

$string =~ s/ (?<!\x0D) \x0A //gx;

print "After:\n>>>$string<<<\n";
[download]

Output:

12:49 >perl 679_SoPW.pl
Before:
>>>"Tap Rackmount for 3 units
","USR4500-RMK         "
<<<
After:
>>>"Tap Rackmount for 3 units","USR4500-RMK         "
<<<

12:49 >
[download]

See the section “Look-Around Assertions” in perlre#Extended-Patterns.

Hope that helps,

Or you can remove all line separators from the line (and then adding one at the end again if needed):

$line =~ s/[\r\n]//g;
[download]

raw dump of the file..first line is okay.Second line has the issue

"Networking Hardware \",\"MDUSR226    \",\"10 Gigabit LR Networking TA
+P            \",\"USR4516             \",\"USRobotics                
+    \",363.33,454.17,0\r\n\"Networking Hardware \",\"MDUSR227    \",\
+"Tap Rackmount for 3 units\n              \",\"USR4500-RMK         \"
+,\"USRobotics                    \",43.33,54.17,0\r\n\
[download]

There is \n within the quotes.I thought setting binary to 1 should take care of this..I have tried the other solutions but none of them seem to work..

Okay, the data looks just as you described it. I placed it into a file named "failing.csv" and ran:

use Parse::CSV;
use Data::Dump 'pp';

my $fh = new IO::File('failing.csv', 'r');

my $fail = do { local $/; <$fh> };
pp $fail;    # Print input
$fh->seek(0,0);

$/ = "\r\n";
my  $parser = Parse::CSV->new(
      handle => $fh,
      csv_attr => { binary => 1 },
);
pp $_  while $_ = $parser->fetch; # Print output
[download]

Which produced...

"\"Networking Hardware \",\"MDUSR226    \",\"10 Gigabit LR Networking 
+TAP            \",\"USR4516             \",\"USRobotics              
+      \",363.33,454.17,0\r\n\"Networking Hardware \",\"MDUSR227    \"
+,\"Tap Rackmount for 3 units\n              \",\"USR4500-RMK         
+\",\"USRobotics                    \",43.33,54.17,0\r\n"
[
  "Networking Hardware ",
  "MDUSR226    ",
  "10 Gigabit LR Networking TAP            ",
  "USR4516             ",
  "USRobotics                    ",
  363.33,
  454.17,
  0,
]
[
  "Networking Hardware ",
  "MDUSR227    ",
  "Tap Rackmount for 3 units\n              ",
  "USR4500-RMK         ",
  "USRobotics                    ",
  43.33,
  54.17,
  0,
]
[download]

If this doesn't work for you there is some local configuration issue. Odd.

sample data from the csv.. these should be two lines but they are split into three. the first line is okay.. but 2 and 3 should be one line

1)  "Networking Hardware ","MDUSR226    ","10 Gigabit LR Networking TAP            ","USR4516             ","USRobotics                    ",363.33,454.17,0

2) "Networking Hardware ","MDUSR227    ","Tap Rackmount for 3 units
3)              ","USR4500-RMK         ","USRobotics                  
+  ",43.33,54.17,0
[download]

when i print the dumper i am only geeting line before the error

$VAR1 = {
          'manufacturer_url' => '',
          'eancode' => '',
          'model' => 'MDUSR226',
          'availability' => 1,
          'longsummary' => '',
          'manufacturer_model' => 'USR4516',
          'category_id' => '198',
          'icecat_prodid' => 0,
          'manufacturer_id' => '49',
          'weight' => 0,
          'category' => 'Networking Hardware',
          'vendor_id' => 7,
          'quantity' => '0',
          'description' => '10 Gigabit LR Networking TAP',
          'image' => 'product_noimage.gif',
          'shortdesc' => '',
          'manufacturer' => 'US Robotics',
          'price' => '377.8632',
          'title' => '10 Gigabit LR Networking TAP',
          'shortsummary' => '',
          'product_url' => ''
        };
[download]

sample data from ...

That is not a Dumper-ing of a raw file -- that doesn't preserve the magic bytes

Try  perl -MData::Dump -MFile::Slurp -e " dd scalar read_file shift, { qw/ binmode :raw / }; "  tenlinefile > tenlinefileasperl.pl