split on comma unless within quotes...

[bcarroll]

I am trying to parse a CSV file into an HTML table, but I am running into an issue where one of the fields contains commas.

I am thinking I may need to use a different regular expression in the split statement, but wanted to check to see if anyone had a better approach.

Sample CSV:

Source,Destination,User,State
192.168.0.2,192.168.0.6,"cn=user1,ou=infrastructure,ou=accounts,o=ORG,
+c=US",Allowed
192.168.0.3,192.168.0.6,"cn=user2,ou=infrastructure,ou=accounts,o=ORG,
+c=US",Denied

This is what I have put together so far...

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

my $lineNum=1;

print "<table>\n";
if ( -f $ARGV[0] ){
  #$ARGV[0] is a file
  open(CSV,'<',$ARGV[0]);
  while (<CSV>){
    csvLine2Html($_);
  }
  close(CSV);
} else {
  #TODO...
  #$ARGV[0] is not a file
}
print "</table>\n";

sub csvLine2Html{
    my $line = shift;    
    chomp($line);
    if ($lineNum == 1){
        #first line contains header information
        print "\t<tr>\n";
        print map{ "\t\t<th>$_</th>\n" } split /,/, $line;
        print "\t</tr>\n";
    } else {
        print "\t<tr>\n";
        print map{ "\t\t<td>$_</td>\n" } split /,/, $line;
        print "\t</tr>\n";      
    }
    $lineNum++;
}

[2teez]

Hi bcarroll,

I am trying to parse a CSV file into an HTML table, but I am running into an issue where one of the fields contains commas. I am thinking I may need to use a different regular expression in the split statement, but wanted to check to see if anyone had a better approach.

To parse CSV file, you will want to use Text::CSV instead of parsing by hand using split function. Though that "may" do on some occasions but it's alot better using a tested module.

Though you didn't show how you want the output look like and which field that you are still having comma in. Using your dataset and going by the title of your post something like this works:

use warnings;
use strict;
use Text::CSV;

my $csv = Text::CSV->new( { binary => 1 } )
  or die Text::CSV->error_diag();

print
"<Table border = 1 cellpadding = 1 callspacing = 1 width = 80% align =
+ 'center'>";
while ( my $row = $csv->getline( \*DATA ) ) {
    print "<tr>";
    print "<td>$_</td>" for @$row;
    print "</tr>";
}
print "</Table>";

__DATA__
Source,Destination,User,State
192.168.0.2,192.168.0.6,"cn=user1,ou=infrastructure,ou=accounts,o=ORG,
+c=US",Allowed
192.168.0.3,192.168.0.6,"cn=user2,ou=infrastructure,ou=accounts,o=ORG,
+c=US",Denied

OUTPUT


Source	Destination	User	State
192.168.0.2	192.168.0.6	cn=user1,ou=infrastructure,ou=accounts,o=ORG,c=US	Allowed
192.168.0.3	192.168.0.6	cn=user2,ou=infrastructure,ou=accounts,o=ORG,c=US	Denied

[frozenwithjoy]

Probably best (and safest!) to use something like Text::CSV.

[Anonymous Monk]

The better approach is to use Text::CSV_XS (or Text::CSV, if you like).

[kcott]

G'day bcarroll,

I'd strongly recommend not spending any time coding your own solution to this problem.

It's already been done in Text::CSV.

[Your Mother]

How many perlmonks does it take to screw in a light bulb help with CSV parsing? :P

Three. The answer is three.

Four. The answer is four.

[bcarroll]

Perhaps I left out the most important part...

The goal is to accomplish this using a module that I don't have to install on every server I intend to run this script on.

Maybe I just need to learn how to use something like PAR

[Anonymous Monk]

Consider that it is likely that:

1. the time you spend writing, testing and debugging your own implementation will be greater than the time you spend installing the module on the servers
2. your implementation will have less features than the existing modules
3. the module you need is already installed on some of those servers

If you still think you want to roll your own, then Text::Balanced is a core module that can help you deal with quoted strings.

[LanX]

I'm sure this was discussed before, searching the archives should be the best strategy.

Some of the answers I gave in the following might help (and illustrate the complexity) --> In place replace, ignoring between quotes

update

especially this post: Re^2: In place replace, ignoring between quotes (escapes)