Depends on what you're using to normalize I suppose. Maybe that's Awk too! Maybe Awk is easier for exploration when the data is already clean but you have to write some parsing layer in Java or something and that's not conducive to one-liner exploration.