If you know that the identifier always begins with CWG and thereafter consists of zero-or-more alphanumerics, write a regex that looks for i"(comma?) .. CWG .. alphanumeric."/i Use the "/g" modifier to allow the regex to be used more than once in the same string. If you know that CWG never occurs at the start of the string, the comma is merely part of the leading string that you are looking for. Do not try to remove characters first since this disrupts the structure of the string and introduces the possibility for errors.