This is tougher to do in PX than in server - I would use a server job, where missing columns at the end of a line can be handled automatically - a functionality not present in PX jobs. The Server job could read & reformat the file and then a PX job could do the real processing.
About the only way I can think of (apart from using cascading reject links on Column Import stages) is to read the entire line as a single VarChar field and parse it using Field() functions in a Transformer stage. Missing fields will return "" from the Field() function; you can do with these what ever you wish. You can also count the delimiters (in a stage variable) to check whether the field is truly missing or validly empty.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.