CPMagical CP Library

Problem solution · C++

Longest Balanced Subarray II

Longest Balanced Subarray II: a C++ solution using segment tree or range structure. Learn the idea, check the complexity, and read the full code, with credit to Kamyu LeetCode Solutions.

Technique
Segment tree or range structure
Source
Kamyu LeetCode Solutions
Length
110 lines
Start with the idea.

Try the problem first. If you get stuck, read the approach below, then write your own solution. The full code is at the bottom.

Approach

Segment tree or range structure

For Longest Balanced Subarray II, the implementation stores interval information in a range-query data structure so updates and queries avoid rescanning the full input.

  1. Choose the aggregate stored for each interval or prefix.
  2. Build or initialize the structure from the input.
  3. Apply updates and combine the affected nodes to answer each query.

Code notes

  • 110 lines of C++ from the credited upstream file longest-balanced-subarray-ii.cpp.
  • The implementation visibly relies on sequence storage, hash lookup.
  • 5 loop blocks detected.

Complexity

Count the build once, then multiply the logarithmic update or query path by the number of operations.

Check the problem constraints before deciding whether this complexity will pass.

Source

Code and credit

This code comes from Kamyu LeetCode Solutions by kamyu104 and is used under the MIT licence.

Open problem ↗Original source ↗
Full codeLongest Balanced Subarray II · C++C++
Use this to learn the idea, then write your own version.
// Time:  O(nlogn)// Space: O(n) // segment tree, binary search, prefix sumclass Solution {public:    int longestBalanced(vector<int>& nums) {        const int n = size(nums) + 1;        SegmentTree st(n);        int result = 0;        unordered_map<int, int> lookup;        for (int i = 1, curr = 0; i <= size(nums); ++i) {            const auto& d = nums[i - 1] & 1 ? +1 : -1;            if (lookup.count(nums[i - 1])) {                st.update(lookup[nums[i - 1]], n - 1, -d);                curr -= d;            }            curr += d;            lookup[nums[i - 1]] = i;            st.update(lookup[nums[i - 1]], n - 1, +d);            const auto& l = i - st.binary_search(curr);            if (l > result) {                result = l;            }        }        return result;    } private:    class SegmentTree {    public:        explicit SegmentTree(int N)         :  base_(N > 1 ? 1 << (__lg(N - 1) + 1) : 1),            lazy_(base_),            min(N > 1 ? 1 << (__lg(N - 1) + 2) : 2),            max(N > 1 ? 1 << (__lg(N - 1) + 2) : 2) {         }         void update(int L, int R, const int val) {            L += base_;            R += base_;            int L0 = L, R0 = R;            for (; L <= R; L >>= 1, R >>= 1) {                if ((L & 1) == 1) {                    apply(L++, val);                }                if ((R & 1) == 0) {                    apply(R--, val);                }            }            pull(L0);            pull(R0);        }         int binary_search(int x) {            int i = 1;            while (!(i >= base_)) {                if (lazy_[i]) {                    apply(i << 1, lazy_[i]);                    apply((i << 1) | 1, lazy_[i]);                    lazy_[i] = 0;                }                i <<= 1;                if (!(min[i] <= x && x <= max[i])) {                    i |= 1;                }            }            return i - base_;        }         vector<int> min, max;     private:        void apply(int x, const int val) {            min[x] += val;            max[x] += val;            if (x < base_) {                lazy_[x] += val;            }        }         void pull(int x) {            while (x > 1) {                x >>= 1;                min[x] = min[x << 1] < min[(x << 1) | 1] ? min[x << 1] : min[(x << 1) | 1];                max[x] = max[x << 1] > max[(x << 1) | 1] ? max[x << 1] : max[(x << 1) | 1];                if (lazy_[x]) {                    min[x] += lazy_[x];                    max[x] += lazy_[x];                }            }        }         void push(int x) {            for (int h = __lg(x) - 1; h > 0; --h) {                int y = x >> h;                if (lazy_[y]) {                    apply(y << 1, lazy_[y]);                    apply((y << 1) | 1, lazy_[y]);                    lazy_[y] = 0;                }            }        }         int base_;        vector<int> lazy_;    };};

Keep practising

Did this explanation save you time? I'm a Grade 11 student building this free library to make difficult algorithms easier to understand.

Buy me a coffee ↗