• I agree with the UDF approach. We use 2 UDFs, one for splitting character data, one for splitting int. Comes in very handy and only done once.