Search code examples
algorithmsuffix-treesuffix-arraysuffix

Suffix array labeling internal nodes


Knowing the internal nodes is helpful in a suffix tree, since they can help you solve problems like finding the longest repeating substring.

These are hard to construct on the spot (think a whiteboard interview). So people have told me to look into suffix arrays.

I have a two part question:

1. Can you create a suffix array without building a suffix tree first? From what I have seen, most implementations build the trie and then traverse it to create a suffix array.

2. Given a suffix array, how can you identify the internal nodes?


Solution

  • (In my opinion this would be an exceptionally hard question for a whiteboard interview...)

    To answer part 1, yes it is possible (and usual) to construct the suffix array directly.

    This link to stanford.edu gives a short O(nlog^2n) algorithm that is simple to implement:

    #include <cstdio>
    #include <cstring>
    #include <algorithm> using namespace std;
    #define MAXN 65536
    #define MAXLG 17
    char A[MAXN];
    struct entry { int nr[2], p;
    } L[MAXN];
    int P[MAXLG][MAXN], N, i, stp, cnt;
    int cmp(struct entry a, struct entry b)
    {
      return a.nr[0] == b.nr[0] ? (a.nr[1] < b.nr[1] ? 1 : 0) : (a.nr[0] < b.nr[0] ? 1 : 0);
    }
    int main(void)
    {
      gets(A); for (N = strlen(A), i = 0; i < N; i ++)
      P[0][i] = A[i] - 'a';
      for (stp = 1, cnt = 1; cnt >> 1 < N; stp ++, cnt <<= 1) {
        for (i = 0; i < N; i ++)
          { L[i].nr[0] = P[stp - 1][i];
            L[i].nr[1] = i + cnt < N ? P[stp - 1][i + cnt] : -1;
            L[i].p = i; }
        sort(L, L + N, cmp);
        for (i = 0; i < N; i ++) P[stp][L[i].p] = i > 0 && L[i].nr[0] == L[i - 1].nr[0] && L[i].nr[1] == L[i - 1].nr[1] ?
        P[stp][L[i - 1].p] : i;
      } return 0;
    } 
    

    This PDF also discusses how to use suffix arrays in practical examples.

    Alternatively, this 2005 paper "Linear Work Suffix Array Construction" gives a O(n) approach for constructing suffix arrays with 50 lines of code.

    In my experiments on a string of length 100k, I found a suffix tree (using Ukkonen's O(n) algorithm) to take 16 seconds, the O(nlog^2n) suffix array to take 2.4 seconds, and the O(n) suffix array to take 0.5 seconds.