That breaks on embedded newlines.

For a solution that solves that, check out Text::xSV (which I intend to put on CPAN this weekend, should have done it ages ago).

use Text::ParseWords;

sub parse_csv {
    return quoteword(",",0, $_[0]);
}
[download]

• "," the separator
• 0 the elements can be quoted
• $_[0] the string to process (the arg to the sub>

traveler

if you still wanna do this manually... try to export your database as a text special-combination-of-characters-delimited file.. try some combination that cannot be found in a file... I tried this once using ::$:: for delimiter, and I'm sure your text won't include these in sequence

hope this helps

If you want to manipulate the data using the DBI (to do SQL queries on it for example) you can also use DBD::RAM or DBD::CSV. This has the added bonus that if one day you want to access the data straight from the DB all you will have to change is the connection instruction.

Easy, split by " :)

Ok, this all assumes we don't have embedded parentheses in the quoted fields (If you do, it gets more complicated). But if you have a variable, $row, let's say, with the following value:

1,"Hugo,Inc.",4,"This is a test",34.5, 17.9, "Marc, Cindy and Rob"

Now, if you split this variable on ", and store it in @splitArray, then you will get the following:

$splitArray[0] = 1,
$splitArray[1] = Hugo, Inc.
$splitArray[2] = ,4,
$splitArray[3] = This is a test
$splitArray[4] = ,34.5,17.9,
$splitArray[5] = Marc, Cindy and Rob
[download]

Now, notice that all of the even subscripts contain the entries that were not in quotes, while the odd subscripts contain the quoted strings. Therefore...

use strict;

my $row = "1, \"Hugo, Inc.\", 4, \"This is a test\", 34.5, 17.9, \"Mar
+c, Cindy and Rob\"";

print "ROW: $row\n";

my @splitArray = split '\"', $row;

my @answerArray;
for ( my $i = 0 ; $i <= $#splitArray ; $i++ )
{
    my @tempArray;
  if ( $i % 2 == 0 ) # if $i is even
  {
    @tempArray = split ',', $splitArray[$i];
  }
  else # if $i is odd
  {
    @tempArray = ( $splitArray[$i] );
  }
  # Shove all of @tempArray into @answerArray
  while ( @tempArray ) {
      if ( $i % 2 == 0 ) {
          $tempArray[0] =~ s/ //g;
      }
       if ( $tempArray[0] =~ /\S/ )
      {    
           push @answerArray, shift @tempArray;
      }
      else {
          shift @tempArray;
      }
  }
}

foreach my $answer ( @answerArray ) {
    print "$answer\n";
}
[download]

There is some extra little cleanup garbage in there which you might not want, but that has the basic idea. Hope it helps.

Some people drink from the fountain of knowledge, others just gargle.

AHA!

I can't always depend on the odd/even thing, because I'm looking for somewhat of a universal way to do this. But, if I first split by ", into a temp array, then go through the entries one by one, if they don't have commas either at the beginning or end (or both), push them into the real array, and if they do, strip off beginning and/or ending comma, and then split by , pushing the results into the real array. That should do it. I'll try it.

Thanks!!