Embellishments

• Keep lists sorted:

  • Average insert cost: 1+N/(2 M)
  • Average search cost (hit): 1+N/(2 M)
  • Average search cost (miss): 1+N/(2 M)

• Move-to-front / transpose

  • Last item accessed in a list becomes the first or moves one closer (Self adjusting hashing)

• Store lists as a binary search tree:

  • Improves expected worst case

• Assume key can be reconstructed (in part) from hash value -- don’t store full key!

Previous slide

Next slide

Back to first slide

View graphic version