[dvko]

I tried the linear search by station name in my first naive approach. Using a hashmap was at least 2-3x as fast with the ~415 distinct keys in the 1BRC dataset.