There is nothing sneaky about it. Excel imports things it thinks are dates as dates. It's not a silent modification. It's a loading issue. Properly loading data into R is not particularly easy either.