MongoDB more optimized approach than multiple distinct(“key1.key2”) or aggregate $group + $addToSet

Question

I have follow collection:

User{
  _id:
  data: {
    languages: ["English", "Chinese", "Spanish"..],
    books: ["foo"