dictionary - Python `dict` indexed by tuple: Getting a slice of the pie -

let's have

my_dict = {   ("airport", "london"): "heathrow",   ("airport", "tokyo"): "narita",   ("hipsters", "london"): "soho" } 

what efficient (no scanning of keys), yet elegant way airports out of dictionary, i.e. expected output ["heathrow", "narita"]. in databases can index tuples, it's possible

airports = my_dict.get(("airport",*)) 

(but 'stars' sitting @ rightmost places in tuple since index stored in 1 order).

since imagine python index dictionary tuple keys in similar way (using keys's inherent order), imagine there might method use slice index way?

edit1: added expected output

edit2: removed last phrase. added '(no scanning of keys)' conditions make clearer.

the way data organized doesn't allow efficient lookup - have scan all keys.

dictionaries hash tables behind scenes, , way access value hash of key - , that, need whole key.

use nested hierarchy this, can direct o(1) lookup:

my_dict = {   "airport": {      "london": "heathrow",      "tokyo": "narita",   },   "hipsters": {      "london": "soho"   } }