thou are god
thank you

I can't get this regex to work if the CSV looks like this...

Title,Price,Description
HelloWorld,"10,00",Desc

I get an array lenght of 4 for the second line.

I changed it a little to be more 'up-to-date' for Java and to work in any case.

static String[] splitCSV( String line ) {
// java.util.ArrayList<String> elements = new java.util.ArrayList<String>(); // JAVA >=1.5
java.util.ArrayList elements = new java.util.ArrayList(); // JAVA <=1.4
java.util.regex.Matcher m = java.util.regex.Pattern.compile( "(?:^|,)(\"(?:[^\"]|\"\")*\"|[^,]*)" ).matcher( line );
while( m.find() ) {
elements.add( m.group()
.replaceAll( "^,", "" ) // remove first comma if any
.replaceAll( "^?\"(.*)\"$", "$1" ) // remove outer quotations if any
.replaceAll( "\"\"", "\"" ) ); // replace double inner quotations if any
}
return (String[])elements.toArray( new String[0] );
}

Doesn't work if you have a comma in your value.

Hi
I've find the some problem with the Regex used in my following code:


System.Text.RegularExpressions.RegexOptions options =
((System.Text.RegularExpressions.RegexOptions.IgnorePatternWhitespace |
System.Text.RegularExpressions.RegexOptions.Multiline) |
System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Regex reg =newRegex("(?:^|,)(\\\"(?:[^\\\"]+|\\\"\\\")*\\\"|[^,]*)",options);
MatchCollection coll = reg.Matches(line);
string[]items = new string[coll.Count];
int i = 0;
foreach(Match m in coll)
{
items[i++] = m.Groups[0].Value.Trim('"').Trim(',').Trim('"').Trim();
}

This code is for splitting CSV. Exactly, I'm getting stuck at the Regex method Matches which is taking so much time around 10 Mins when I've passed the following CSV:

string line = "C,\"123 PSST, MASSACHUSETTS,5245352,3343432";



Any Help will be appreciated.

Thank you.

For a personal project, I'm using this one which I thought up:

("(?:[^\\"]|\\.)*")\s*($|,)

It's not exactly a pure CSV file regexp, but it works for the purpose I designed it for. It matches everything between a pair of double quotes in a CSV file, including escaped double quotes. You can use commas in the values, and empty values are ignored. If it's fed incorrect data as such as [ "value 1","value 2" error here "value 3", "value 4" ], it will ignore the second value and return the last usable quoted string before the comma.

The quoted content is returned in the first (and only) matching group, stripped of remaining leading and trailing whitespace (and newlines).

There might be a few errors still lurking in this regexp though, but it works a 100% for me.

With regards to my previous reply;

Add a "?:" in the last matching group to correct the second matching group appearing in the resultset. I copied the wrong regexp from my text editor. Sorry for that--my bad.

This is the (full) correct regexp:

("(?:[^\\"]|\\.)*")\s*(?:$|,)

thou are go0d
thank you

Thanks you

Thank youS

Thanks for this and thanks @Mihi for the Java version!

Thank u!!!!

There's something broken in that regex. If you goto http://regexpal.com/ and test it with the string

,test,test2,test3,test4,test5

You'll see it skips matching the first test.

Thank you very much!!!

hi

you can add "" at first like this.

if (line.startsWith(",")) {
line = "\"\"" + line;
}

Thanks a lot dude!