consider the below csv data

  A,B,,D,E,F,E,G
      In the third place there is a null value instead of (A,B,C,D..).
      how to capture and handle that null value.

You have to be very careful with your terminology. It is not a null value, it is a zero-length string.

Most of the standard approaches to tokenizing (for example the tokenize() function in XPath 2.0, or head-tail recursion in 1.0) handle zero-length strings with no problems. If your code isn't handling this case then it is wrong, but to see where it is wrong, we first need to see it.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference