Rabin Karp Pattern Matching in C++

Algorithms: String Algorithms November 15, 2026 C++

Rabin-Karp uses a rolling hash to filter candidate matches, then compares tokens to guarantee correctness despite hash collisions..

cpp
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
#include <stdio.h>
#include <vector>
#include <algorithm>
#include <stdexcept>
#include <cstdint>
#include <limits>
// Rabin Karp Pattern Matching in C++
//
// PURPOSE
// Rabin-Karp uses a rolling hash to filter candidate matches, then compares tokens to
// guarantee correctness despite hash collisions.
//
// METHOD
// Hash the pattern and first window modulo a small prime. At each position verify matching
// hashes, then remove the old token and append the new one.
//
// COMPLEXITY
// Expected O(n+m) time on suitable data; O(nm) worst case due to verification. O(1)
// auxiliary space.
//
// CONTRACT
// Tokens are integers in [0,255]. Empty pattern matches at zero. Returns first index or -1;
// not a cryptographic hash.
// Public APIs infer array lengths except in C, which accepts pointer/length pairs. Vertex
// counts and matrix dimensions remain explicit. Arrays must contain signed 32-bit integers.
// Sorted order, correct tree topology, and caller-owned data-structure invariants are
// preconditions; checking them on every query would change query complexity.
// Scalar and shape checks are included. Value-range checks are included where needed to
// prevent invalid indexing or nontermination. Allocation sizes are capped at 4,000,000
// workspace entries. Arithmetic outside Python raises an error on 32-bit overflow; Python
// uses arbitrary precision. Native allocation failures propagate. In-place outputs may be
// partly updated after arithmetic failure. These are single-threaded, in-memory
// implementations.
//
// LANGUAGE DETAILS
// C++17: const vector references for inputs, mutable references for outputs, RAII workspace,
// standard lower_bound for LIS. Invalid inputs throw invalid_argument; overflow throws
// overflow_error; allocation failure propagates bad_alloc.
//
// EXAMPLE
// Expected return: 2.
// Run: g++ -std=c++17 -O2 example.cpp -o example && ./example
// Tests for this exact source artifact run before publication. The accompanying source
// archive includes unit cases, runtime results, and source hashes.

static int alg_min(int a, int b) { return a < b ? a : b; }
static int alg_max(int a, int b) { return a > b ? a : b; }
static void alg_require(bool valid) { if (!valid) throw std::invalid_argument("invalid algorithm input"); }
static int alg_checked(int64_t v) { if (v < INT32_MIN || v > INT32_MAX) throw std::overflow_error("32-bit result overflow"); return (int)v; }
static int alg_add(int a,int b) { return alg_checked((int64_t)a+b); }
static int alg_sub(int a,int b) { return alg_checked((int64_t)a-b); }
static int alg_mul(int a,int b) { return alg_checked((int64_t)a*b); }
static size_t alg_size(int n) { alg_require(n>=0 && n<=4000000); return (size_t)n; }

int rabin_karp(const std::vector<int>& text, const std::vector<int>& pattern) {
    alg_require(text.size() <= 1000000);
    alg_require(pattern.size() <= 1000000);
    int n = (int)text.size();
    int m = (int)pattern.size();
    alg_require(((n >= 0) && (n <= 1000000)));
    alg_require(((m >= 0) && (m <= 1000000)));
    alg_require(((int)pattern.size() == m));
    alg_require(((int)text.size() == n));
    for (int alg_i = 0; alg_i < (int)text.size(); alg_i++) {
        int alg_value = text[alg_i];
        alg_require(((alg_value >= 0) && (alg_value <= 255)));
    }
    for (int alg_i = 0; alg_i < (int)pattern.size(); alg_i++) {
        int alg_value = pattern[alg_i];
        alg_require(((alg_value >= 0) && (alg_value <= 255)));
    }
    if ((m == 0)) {
        return 0;
    }
    if ((m > n)) {
        return alg_sub(0, 1);
    }
    int power = 1;
    int target_hash = 0;
    int window_hash = 0;
    int i = 0;
    int j = 0;
    while ((i < m)) {
        target_hash = (alg_add(alg_mul(target_hash, 256), pattern[i]) % 10007);
        window_hash = (alg_add(alg_mul(window_hash, 256), text[i]) % 10007);
        if ((i < alg_sub(m, 1))) {
            power = (alg_mul(power, 256) % 10007);
        }
        i = alg_add(i, 1);
    }
    i = 0;
    while ((i <= alg_sub(n, m))) {
        if ((target_hash == window_hash)) {
            j = 0;
            while (((j < m) && (text[alg_add(i, j)] == pattern[j]))) {
                j = alg_add(j, 1);
            }
            if ((j == m)) {
                return i;
            }
        }
        if ((i < alg_sub(n, m))) {
            window_hash = (alg_add(alg_sub(window_hash, (alg_mul(text[i], power) % 10007)), 10007) % 10007);
            window_hash = (alg_add(alg_mul(window_hash, 256), text[alg_add(i, m)]) % 10007);
        }
        i = alg_add(i, 1);
    }
    return alg_sub(0, 1);
}

int main(void) {
    std::vector<int> text = {1, 2, 1, 2, 3};
    std::vector<int> pattern = {1, 2, 3};
    int result = rabin_karp(text, pattern);
    printf("%d\n", result);
    return 0;
}