华南理工大学数据结构chapter9.pptVIP

* * * * * * * * * * * * * * * Hashing is not good for range queries. * * * * * * * * * * * * * * For (a): prob(3) = 4/11, prob(4) = 1/11, prob(5) = 1/11, prob(6) = 1/11, prob(10) = 4/11. For(b): prob(3) = 8/11, prob(4,5,6) = 1/11 each. * If M=10 and c=2, then we effectively have created 2 hash tables (evens vs. odds). * * Probe sequence for k1: 30, 31, 34, 39. Probe sequence for k2: 29, 30, 33, 38. * * Probe sequence for k1: 30, 32, 34, 36. Probe sequence for k2: 28, 33, 38,43. Probe sequence for k3: 30, 35, 40, 45. * * * * * * * * * Quadratic Probing Set the i’th value in the probe sequence as for some choice of constants c1, c2, and c3. The simplest variation is p(K; i) = i2 (i.e., c1 = 1, c2 = 0, and c3 = 0. Then the ith value in the probe sequence would be (h(K) + i2) mod M. Under quadratic probing, two keys with different home positions will have diverging probe sequences. Example: M=101 h(k1)=30, h(k2) = 29. Probe sequence for k1 :30，31，34，39 Probe sequence for k2 :29，30，33，38 Disadvantage: not all hash table slots will necessarily be on the probe sequence. * Secondary Clustering Both pseudo-random probing and quadratic probing eliminate primary clustering, the problem of keys sharing substantial segments of a probe sequence. But if two keys hash to the same home position, they follow the same probe sequence. This is called secondary clustering. To avoid secondary clustering, need probe sequence to be a function of the original key value, not just the home position. * Double Hashing p(K, i) = i * h2(K) Be sure that all probe sequence constants (h2(K)) are relatively prime to M. This will be true if M is prime, 1=h2(k)=M-1; or if M=2m and h2(k) is odd belonging to [1, 2m]. Probe sequences is (h(k)+i* h2(K))%M (i=0) Example: Hash table of size M=101 h(k1)=30, h(k2)=28, h(k3)=30. h2(k1)=2, h2(k2)=5, h2(k3)=5. Probe sequence for k1 is: 30, 32, 34, 36. Probe sequence for k2 is: 28, 33, 38, 43. Probe sequence for k3 is: 30, 35, 40, 45. * Analysis of Clo