It all depends on what you are doing with the data.
Do you need to just scan through the file until you find the line(s) you want, or just process each line one at a time? Then there's no point in putting the whole file in any data structure. Just read one line at a time and process it without saving it.
Do you need to repeatedly look up which files to send by email address? Then a hash of arrays (with email address as the key) is probably best.
Do you need to repeatedly look up all addresses which send a certain file?