Ignore consecutive whitespaces with CSV.read(...)

I think it is a normal behaviour for counting double space as the one with missing column. Otherwise, there is an ambiguity on how to define possible NAs