my @headings = split /(?<=")\s*,\s*(?=")/, $line;
[download]

But that makes no attempt at tokenizing and parsing the line for balanced quotes, escaped quotes in a field, or quoted commas. If you need to deal with those issues, you've got more work ahead of you, the complexity of which is why people tend to recommend and use Text::CSV.

1. comma between two words like this - "Entries","Entry Time","Visit Length","Browser"

So this would be complicated by the fact that commas inside quoted text should not count, right? In other words, "Name, First","Name, Last","Birthday" should be split into "Name, First", then "Name, Last", and then "Birthday"?

I think that's starting to become a bit complicated for a single regex, though you could do it using match-time code evaluation and the \G assertion ("picking up where you left off"), perhaps, keeping track of whether you're in a quoted string or not as you go along.

A parser based on a formal grammar might be a better idea; take a look at Parse::RecDescent if you'd like to go that route.

Yet another option would be to use a loop and process the line in chunks: read from the beginning of (the current remainder of) the line up to the first comma or quote, keep track of whether you're in a quote right now etc., and use that information to either add the newly-read chunk to your current extracted column name, or increase your column counter and start a new column name with that chunk.

Since you want to learn how to do this on your own, I won't provide any code. :)

Now, all that said...

If commas are guaranteed to not appear in your quoted strings, everything becomes much easier, of course, \G alone should be enough to get the job done, though you could save yourself the trouble and simply split the line.

Hmm. I wonder if you could use split and then pull some clever tricks to glue the right bits together again in the general case, too. Well, you'll find out, if you decide to go down this route!

2. Match is multiple times a comma I tried many different regexes with no luck, please help!

I'm sorry, I don't know what you mean there.

$line =~ s/"//g;
my @fields = split /,/, $line;
[download]

"New York, NY","Entries","Entry Time","Visit Length","Browser"
[download]

On files coming from an outside organization on which you have no control, don't do it. In your case, the simple fact that there are both separators and delimiters is a sign that there is a clear danger that your process is probably going to break on the first special case coming up. The Text::CSV module is really a much more robust alternative in any case where you don't control everything yourself.

I know about the CSV module, but I wanna write my own program to learn ... So will you please tell me how to match a:

Super Search , search for CSV and for Text::CSV and you'll find all the hundreds of nodes on this exact thing

Then pick through them :)

Here is something similar to what you are looking for. Perhaps this will help you work out your particular solution. (I have a similar situation, dealing with csv files and not having the very useful modules available to me.) This code replaces commas inside of quotation marks with spaces so I could subsequently use split.

$olds='one,two,"three,four,five","six",seven,"eight,nine"';
$i = 0;
$news = join '', grep { $i++ % 2 ? {s/,/ /g} : 1 } split /"/,$olds;
print $news . "\n";
[download]

Output:

one,two,three four five,six,seven,eight nine
[download]

Update

Original version would delete single word inside quotation marks.