Maybe you want to remove the empty fields before you output them?

my @output= grep {/\S/} @file_line_contents;
print OUTFILE1 "\"$coid\"\,\"".join ('","', @output);
[download]

But for writing CSV, you really want to look at Text::CSV.

we are using a lot of CSV format at our job. Most of the time, we can control both the generation of the CSV file in one application and its later use in another application. So we can use a very simple CSV format (ASCII characters, no quotes, no escape char, almost always the same separator, etc.), and we are happily using join and split to generate and decypher our CSVs. It can really make coding very simple. For example, just today, I wrote a line of code looking like this:

my $err_code = (split /;/, $line)[8];
[download]

But some of our partners sometimes send us more complicated CSV files, with quote and escape characters, and so on, then, really, using the Text::CSV module is handy and is really, IMHO, the right way to go.

From your description of your data I imply that there are two consecutive commas in each line, either after field 3 or before field 4. You could just s/,,/,/; on each line.

Update: While you are at it you could also add quotes around the remaining commas and print the line:

while(<INFILE1>) {
  chomp;
  s/,,/,/;
  s/,/","/g;
  print OUTFILE1 "\"$coid\",\"$_\"\n";
}
[download]

Use Text::CSV_XS' csv function with the on_in callback:

$ cat test.pl
use 5.20.0;
use warnings;

use Text::CSV_XS "csv";

csv (in => "test.csv", blank_is_undef => 1, on_in => sub {
    splice @{$_[1]}, 2, 2, $_[1][2] // $_[1][3]; });
$ cat test.csv
1,2,,,5,6
1,2,3,,5,6
1,2,,4,5,6
1,2,3,4,5,6
$ perl test.pl
1,2,,5,6
1,2,3,5,6
1,2,4,5,6
1,2,3,5,6
[download]

update: That also surfaced a small buglet in the code, as I wanted to show that that all fields on output could be forced quoted in the same call, but somewhere that option currently gets lost. The workaround however is easy:

$ cat test.pl
use 5.20.0;
use warnings;

use Text::CSV_XS "csv";

csv (quote_always => 1, in => csv (in => "test.csv", blank_is_undef =>
+ 1,
     on_in => sub { splice @{$_[1]}, 2, 2, $_[1][2] // $_[1][3]; }));
$ perl test.pl
"1","2",,"5","6"
"1","2","3","5","6"
"1","2","4","5","6"
"1","2","3","5","6"
[download]

Working on a fix for 1.18 ...

update 2: fixed. I'll release after going through the long test procedure.

$ perl -MText::CSV_XS=csv -e'csv (in => "test.csv", blank_is_undef => 
+1, quote_always => 1, on_in => sub { splice @{$_[1]}, 2, 2, $_[1][2] 
+// $_[1][3]; });'
"1","2",,"5","6"
"1","2","3","5","6"
"1","2","4","5","6"
"1","2","3","5","6"
[download]

This might be gross, but here is the brute force approach that I see:

while(<INFILE1>) {
    @file_line_contents = split (/\,/,$_);

    # this assumes that $file_line_contents[3] never contains the valu
+e '0'
    $file_line_contents[3] ||= $file_line_contents[4];
    # now get rid of the unneeded element (assumes this is the last el
+ement in the array)
    pop @file_line_contents;

    print OUTFILE1 "\"$coid\"\,\"".join ('","', @file_line_contents);
}
[download]