Hi, are you thinking about duplicate rows or duplicate PI values?
If you're thinking about duplicate rows, then you'll probably have a multiset table with a NUPI on it. 1- if you don't want to have duplicate rows in a table convert your table to set and/or put a UPI (or also a USI) on it. 2- fastload automatically discards perfectly duplicated rows. if you want to load them in a table you have to use another utility, for example multiload.
If you're thinking about duplicate values for the PI: 1- if you don't want to have duplicate values put a UPI on your table 2- using fastload it will automatically discard duplicate PI values and put them in errortable 1. have a look to the manuals.