Re: regular expression (search and destroy)

You surely *can* split the record in one go -

Version 1 - '\' escaped quotes:

use strict;
use Data::Dumper;

while (<DATA>)
{
  chomp;
  my @rec;
  # was - foreach (split /"(.*,.*)"|,/) ...
  foreach (split /"((?:\\"|.)*?)"|,/) { push @rec, $_ if $_ }
  print Dumper(\@rec);
}

__DATA__
1,"Hello, world",This is good,2
121212,"Simpson, Bart",Springfield,"Roger"
121212,"2\" tape, \"white",springfield,"Roger"
121212,"Simpson \", Bart",Springfield,"Roger"
[download]

And the output is -

$VAR1 = [
          '1',
          'Hello, world',
          'This is good',
          '2'
        ];
$VAR1 = [
          '121212',
          'Simpson, Bart',
          'Springfield',
          'Roger'
        ];
$VAR1 = [
          '121212',
          '2\\" tape, \\"white',
          'springfield',
          'Roger'
        ];
$VAR1 = [
          '121212',
          'Simpson \\", Bart',
          'Springfield',
          'Roger'
        ];
[download]

Update: There was a minor flaw in the original solution, I did not search for escaped quotes inside the quote, here's the enhanced version.

Version 2 - '"' escaped quotes:

use strict;
use Data::Dumper;

while (<DATA>)
{
    chomp;
    my @rec;
    foreach (split /"(.*?)(?:(?<!")"(?!")|(?<="")"(?!"))|,/)
        { s/""/"/g, push @rec, $_ if $_ }
    print Dumper(\@rec);
}

__DATA__
1,"Hello, world",This is good,2
121212,"Simpson, Bart",Springfield,"Roger"
121212,"2"" tape, ""white",springfield,"Roger"
121212,"Simpson "", Bart",Springfield,"Roger"
121212,"2""",springfield,"Roger"
[download]

And output is -

$VAR1 = [
          '1',
          'Hello, world',
          'This is good',
          '2'
        ];
$VAR1 = [
          '121212',
          'Simpson, Bart',
          'Springfield',
          'Roger'
        ];
$VAR1 = [
          '121212',
          '2" tape, "white',
          'springfield',
          'Roger'
        ];
$VAR1 = [
          '121212',
          'Simpson ", Bart',
          'Springfield',
          'Roger'
        ];
$VAR1 = [
          '121212',
          '2"',
          'springfield',
          'Roger'
        ];
[download]

Update: Thanks to antirice to point out that the quotes are escaped by quote ("), not backslash (\) inside the quote.

Comment on Re: regular expression (search and destroy) Select or Download Code

Replies are listed 'Best First'.

Re: Re: regular expression (search and destroy)
by antirice (Priest) on Nov 12, 2003 at 23:24 UTC

If I recall correctly, you can have empty fields and the way you actually escape a double quote is as "" instead of \". In other words, these are allowed:

1,"Hello world",,2
2,"""Hello world""",,2
[download]

I also believe that if the field contains a double quote, then the entire field must be enclosed in double quotes so escaping rules could apply. There are more rules, but they're mostly allowed values and input record separator issues more than anything else, but that's just for saving and verification that the file you have is really CSV. I think you can read more about them at the end of the documentation here.

Of course, this is only if you want valid CSV or CSV produced by programs such as Excel and whatnot.

Hope this helps.

antirice
The first rule of Perl club is - use Perl
The ith rule of Perl club is - follow rule i - 1 for i > 1

[reply]
[d/l]


Syntactic Confectionery Delight
	PerlMonks