Posts tagged as “union find”

花花酱 LeetCode 2076. Process Restricted Friend Requests

By zxi on November 15, 2021

You are given an integer n indicating the number of people in a network. Each person is labeled from 0 to n - 1.

You are also given a 0-indexed 2D integer array restrictions, where restrictions[i] = [x_i, y_i] means that person x_i and person y_i cannot become friends,either directly or indirectly through other people.

Initially, no one is friends with each other. You are given a list of friend requests as a 0-indexed 2D integer array requests, where requests[j] = [u_j, v_j] is a friend request between person u_j and person v_j.

A friend request is successful if u_j and v_j can be friends. Each friend request is processed in the given order (i.e., requests[j] occurs before requests[j + 1]), and upon a successful request, u_j and v_j become direct friends for all future friend requests.

Return a boolean array result, where each result[j] is true if the j^th friend request is successful or false if it is not.

Note: If u_j and v_j are already direct friends, the request is still successful.

Example 1:

Input: n = 3, restrictions = [[0,1]], requests = [[0,2],[2,1]]
Output: [true,false]
Explanation:
Request 0: Person 0 and person 2 can be friends, so they become direct friends. 
Request 1: Person 2 and person 1 cannot be friends since person 0 and person 1 would be indirect friends (1--2--0).

Example 2:

Input: n = 3, restrictions = [[0,1]], requests = [[1,2],[0,2]]
Output: [true,false]
Explanation:
Request 0: Person 1 and person 2 can be friends, so they become direct friends.
Request 1: Person 0 and person 2 cannot be friends since person 0 and person 1 would be indirect friends (0--2--1).

Example 3:

Input: n = 5, restrictions = [[0,1],[1,2],[2,3]], requests = [[0,4],[1,2],[3,1],[3,4]]
Output: [true,false,true,false]
Explanation:
Request 0: Person 0 and person 4 can be friends, so they become direct friends.
Request 1: Person 1 and person 2 cannot be friends since they are directly restricted.
Request 2: Person 3 and person 1 can be friends, so they become direct friends.
Request 3: Person 3 and person 4 cannot be friends since person 0 and person 1 would be indirect friends (0--4--3--1).

Constraints:

2 <= n <= 1000
0 <= restrictions.length <= 1000
restrictions[i].length == 2
0 <= x_i, y_i <= n - 1
x_i != y_i
1 <= requests.length <= 1000
requests[j].length == 2
0 <= u_j, v_j <= n - 1
u_j != v_j

Solution: Union Find / Brute Force

For each request, check all restrictions.

Time complexity: O(req * res)
Space complexity: O(n)

C++

// Author: Huahua

class Solution {

public:

vector<bool> friendRequests(int n, vector<vector<int>>& restrictions, vector<vector<int>>& requests) {

vector<int> parents(n);

iota(begin(parents), end(parents), 0);

function<int(int)> find = [&](int x) {

if (parents[x] == x) return x;

return parents[x] = find(parents[x]);

};

auto check = [&](int u, int v) {

for (const auto& r : restrictions) {

int pu = find(r[0]);

int pv = find(r[1]);

if ((pu == u && pv == v) || (pu == v && pv == u))

return false;

}

return true;

};

vector<bool> ans;

for (const auto& r : requests) {

int pu = find(r[0]);

int pv = find(r[1]);

if (pu == pv || check(pu, pv)) {

parents[pu] = pv;

ans.push_back(true);

} else {

ans.push_back(false);

}

return ans;

}

};

花花酱 LeetCode 1697. Checking Existence of Edge Length Limited Paths

By zxi on December 20, 2020

An undirected graph of n nodes is defined by edgeList, where edgeList[i] = [u_i, v_i, dis_i] denotes an edge between nodes u_i and v_i with distance dis_i. Note that there may be multiple edges between two nodes.

Given an array queries, where queries[j] = [p_j, q_j, limit_j], your task is to determine for each queries[j] whether there is a path between p_j and q_jsuch that each edge on the path has a distance strictly less than limit_j .

Return a boolean array answer, where answer.length == queries.length and the j^th value of answer is true if there is a path for queries[j] is true, and false otherwise.

Example 1:

Input: n = 3, edgeList = [[0,1,2],[1,2,4],[2,0,8],[1,0,16]], queries = [[0,1,2],[0,2,5]]
Output: [false,true]
Explanation: The above figure shows the given graph. Note that there are two overlapping edges between 0 and 1 with distances 2 and 16.
For the first query, between 0 and 1 there is no path where each distance is less than 2, thus we return false for this query.
For the second query, there is a path (0 -> 1 -> 2) of two edges with distances less than 5, thus we return true for this query.

Example 2:

Input: n = 5, edgeList = [[0,1,10],[1,2,5],[2,3,9],[3,4,13]], queries = [[0,4,14],[1,4,13]]
Output: [true,false]
Exaplanation: The above figure shows the given graph.

Constraints:

2 <= n <= 10⁵
1 <= edgeList.length, queries.length <= 10⁵
edgeList[i].length == 3
queries[j].length == 3
0 <= u_i, v_i, p_j, q_j <= n - 1
u_i != v_i
p_j != q_j
1 <= dis_i, limit_j <= 10⁹
There may be multiple edges between two nodes.

Solution: Union Find

Since queries are offline, we can reorder them to optimize time complexity. Answer queries by their limits in ascending order while union edges by weights up to the limit. In this case, we just need to go through the entire edge list at most once.

Time complexity: O(QlogQ + ElogE)
Space complexity: O(Q + E)

C++

// Author: Huahua

class Solution {

public:

vector<bool> distanceLimitedPathsExist(int n, vector<vector<int>>& E, vector<vector<int>>& Q) {

vector<int> parents(n);

iota(begin(parents), end(parents), 0);

function<int(int)> find = [&](int x) {

return parents[x] == x ? x : parents[x] = find(parents[x]);

};

const int m = Q.size();

for (int i = 0; i < m; ++i) Q[i].push_back(i);

// Sort edges by weight in ascending order.

sort(begin(E), end(E), [](const auto& a, const auto& b) { return a[2] < b[2]; });

// Sort queries by limit in ascending order

sort(begin(Q), end(Q), [](const auto& a, const auto& b) { return a[2] < b[2]; });

vector<bool> ans(m);

int i = 0;

for (const auto& q : Q) {

while (i < E.size() && E[i][2] < q[2])

parents[find(E[i++][0])] = find(E[i][1]);

ans[q[3]] = find(q[0]) == find(q[1]);

}

return ans;

}

};

花花酱 LeetCode 1632. Rank Transform of a Matrix

By zxi on October 28, 2020

Given an m x n matrix, return a new matrix answer where answer[row][col] is the rank of matrix[row][col].

The rank is an integer that represents how large an element is compared to other elements. It is calculated using the following rules:

The rank is an integer starting from 1.
If two elements p and q are in the same row or column, then:
- If p < q then rank(p) < rank(q)
- If p == q then rank(p) == rank(q)
- If p > q then rank(p) > rank(q)
The rank should be as small as possible.

It is guaranteed that answer is unique under the given rules.

Example 1:

Input: matrix = [[1,2],[3,4]]
Output: [[1,2],[2,3]]
Explanation:
The rank of matrix[0][0] is 1 because it is the smallest integer in its row and column.
The rank of matrix[0][1] is 2 because matrix[0][1] > matrix[0][0] and matrix[0][0] is rank 1.
The rank of matrix[1][0] is 2 because matrix[1][0] > matrix[0][0] and matrix[0][0] is rank 1.
The rank of matrix[1][1] is 3 because matrix[1][1] > matrix[0][1], matrix[1][1] > matrix[1][0], and both matrix[0][1] and matrix[1][0] are rank 2.

Example 2:

Input: matrix = [[7,7],[7,7]]
Output: [[1,1],[1,1]]

Example 3:

Input: matrix = [[20,-21,14],[-19,4,19],[22,-47,24],[-19,4,19]]
Output: [[4,2,3],[1,3,4],[5,1,6],[1,3,4]]

Example 4:

Input: matrix = [[7,3,6],[1,4,5],[9,8,2]]
Output: [[5,1,4],[1,2,3],[6,3,1]]

Constraints:

m == matrix.length
n == matrix[i].length
1 <= m, n <= 500
-10⁹ <= matrix[row][col] <= 10⁹

Solution: Union Find

Group cells by their values, process groups (cells that have the same value) in ascending order (smaller number has smaller rank).

For cells that are in the same row and same cols union them using union find, they should have the same rank which equals to max(max_rank_x[cols], max_rank_y[rows]) + 1.

Time complexity: O(m*n*(m+n))
Space complexity: O(m*n)

C++

// Author: Huahua

class DSU {

public:

DSU(int n): p_(n, -1) {}

int find(int x) {

return p_[x] == -1 ? x : p_[x] = find(p_[x]);

}

void merge(int x, int y) {

x = find(x), y = find(y);

if (x != y) p_[x] = y;

}

private:

vector<int> p_;

};

class Solution {

public:

vector<vector<int>> matrixRankTransform(vector<vector<int>>& matrix) {

const int m = matrix.size();

const int n = matrix[0].size();

vector<vector<int>> ans(m, vector<int>(n));

map<int, vector<pair<int, int>>> mp; // val -> {positions}

for (int y = 0; y < m; ++y)

for (int x = 0; x < n; ++x)

mp[matrix[y][x]].emplace_back(x, y);

vector<int> rx(n), ry(m);

for (const auto& [val, ps]: mp) {

DSU dsu(n + m);

vector<vector<pair<int, int>>> cc(n + m); // val -> {positions}

for (const auto& [x, y]: ps)

dsu.merge(x, y + n);

for (const auto& [x, y]: ps)

cc[dsu.find(x)].emplace_back(x, y);

for (const auto& ps: cc) {

int rank = 1;

for (const auto& [x, y]: ps)

rank = max(rank, max(rx[x], ry[y]) + 1);

for (const auto& [x, y]: ps)

rx[x] = ry[y] = ans[y][x] = rank;

}

return ans;

}

};

花花酱 LeetCode 1627. Graph Connectivity With Threshold

By zxi on October 18, 2020

We have n cities labeled from 1 to n. Two different cities with labels x and y are directly connected by a bidirectional road if and only if x and y share a common divisor strictly greater than some threshold. More formally, cities with labels x and y have a road between them if there exists an integer z such that all of the following are true:

x % z == 0,
y % z == 0, and
z > threshold.

Given the two integers, n and threshold, and an array of queries, you must determine for each queries[i] = [a_i, b_i] if cities a_i and b_i are connected (i.e. there is some path between them).

Return an array answer, where answer.length == queries.length and answer[i] is true if for the i^th query, there is a path between a_i and b_i, or answer[i] is false if there is no path.

Example 1:

Input: n = 6, threshold = 2, queries = [[1,4],[2,5],[3,6]]
Output: [false,false,true]
Explanation: The divisors for each number:
1:   1
2:   1, 2
3:   1, 3
4:   1, 2, 4
5:   1, 5
6:   1, 2, 3, 6
Using the underlined divisors above the threshold, only cities 3 and 6 share a common divisor, so they are the
only ones directly connected. The result of each query:
[1,4]   1 is not connected to 4
[2,5]   2 is not connected to 5
[3,6]   3 is connected to 6 through path 3--6

Example 2:

Input: n = 6, threshold = 0, queries = [[4,5],[3,4],[3,2],[2,6],[1,3]]
Output: [true,true,true,true,true]
Explanation: The divisors for each number are the same as the previous example. However, since the threshold is 0,
all divisors can be used. Since all numbers share 1 as a divisor, all cities are connected.

Example 3:

Input: n = 5, threshold = 1, queries = [[4,5],[4,5],[3,2],[2,3],[3,4]]
Output: [false,false,false,false,false]
Explanation: Only cities 2 and 4 share a common divisor 2 which is strictly greater than the threshold 1, so they are the only ones directly connected.
Please notice that there can be multiple queries for the same pair of nodes [x, y], and that the query [x, y] is equivalent to the query [y, x].

Constraints:

2 <= n <= 10⁴
0 <= threshold <= n
1 <= queries.length <= 10⁵
queries[i].length == 2
1 <= a_i, b_i <= cities
a_i != b_i

Solution: Union Find

For x, merge 2x, 3x, 4x, ..,
If a number is already “merged”, skip it.

Time complexity: O(nlogn? + queries)?
Space complexity: O(n)

C++

// Author: Huahua

class Solution {

public:

vector<bool> areConnected(int n, int threshold, vector<vector<int>>& queries) {

if (threshold == 0) return vector<bool>(queries.size(), true);

vector<int> ds(n + 1);

iota(begin(ds), end(ds), 0);

function<int(int)> find = [&](int x) {

return ds[x] == x ? x : ds[x] = find(ds[x]);

};

for (int x = threshold + 1; x <= n; ++x)

if (ds[x] == x)

for (int y = 2 * x; y <= n; y += x)

ds[max(find(x), find(y))] = min(find(x), find(y));

vector<bool> ans;

for (const vector<int>& q : queries)

ans.push_back(find(q[0]) == find(q[1]));

return ans;

}

};

Python3

# Author: Huahua

class Solution:

def areConnected(self, n: int, threshold: int,

queries: List[List[int]]) -> List[bool]:

if threshold == 0: return [True] * len(queries)

ds = list(range(n + 1))

def find(x: int) -> int:

if x != ds[x]: ds[x] = find(ds[x])

return ds[x]

for x in range(threshold + 1, n + 1):

if ds[x] == x:

for y in range(2 * x, n + 1, x):

ds[max(find(x), find(y))] = min(find(x), find(y))

return [find(x) == find(y) for x, y in queries]

花花酱 LeetCode 1579. Remove Max Number of Edges to Keep Graph Fully Traversable

By zxi on September 6, 2020

Alice and Bob have an undirected graph of n nodes and 3 types of edges:

Type 1: Can be traversed by Alice only.
Type 2: Can be traversed by Bob only.
Type 3: Can by traversed by both Alice and Bob.

Given an array edges where edges[i] = [type_i, u_i, v_i] represents a bidirectional edge of type type_i between nodes u_i and v_i, find the maximum number of edges you can remove so that after removing the edges, the graph can still be fully traversed by both Alice and Bob. The graph is fully traversed by Alice and Bob if starting from any node, they can reach all other nodes.

Return the maximum number of edges you can remove, or return -1 if it’s impossible for the graph to be fully traversed by Alice and Bob.

Example 1:

Input: n = 4, edges = [[3,1,2],[3,2,3],[1,1,3],[1,2,4],[1,1,2],[2,3,4]]
Output: 2
Explanation: If we remove the 2 edges [1,1,2] and [1,1,3]. The graph will still be fully traversable by Alice and Bob. Removing any additional edge will not make it so. So the maximum number of edges we can remove is 2.

Example 2:

Input: n = 4, edges = [[3,1,2],[3,2,3],[1,1,4],[2,1,4]]
Output: 0
Explanation: Notice that removing any edge will not make the graph fully traversable by Alice and Bob.

Example 3:

Input: n = 4, edges = [[3,2,3],[1,1,2],[2,3,4]]
Output: -1
Explanation: In the current graph, Alice cannot reach node 4 from the other nodes. Likewise, Bob cannot reach 1. Therefore it's impossible to make the graph fully traversable.

Constraints:

1 <= n <= 10^5
1 <= edges.length <= min(10^5, 3 * n * (n-1) / 2)
edges[i].length == 3
1 <= edges[i][0] <= 3
1 <= edges[i][1] < edges[i][2] <= n
All tuples (type_i, u_i, v_i) are distinct.

Solution: Greedy + Spanning Tree / Union Find

Use type 3 (both) edges first.

Time complexity: O(E)
Space complexity: O(n)

C++

class DSU {

public:

DSU(int n): p_(n + 1), e_(0) {

iota(begin(p_), end(p_), 0);

}

int find(int x) {

if (p_[x] == x) return x;

return p_[x] = find(p_[x]);

}

int merge(int x, int y) {

int rx = find(x);

int ry = find(y);

if (rx == ry) return 1;

p_[rx] = ry;

++e_;

return 0;

}

int edges() const { return e_; }

private:

vector<int> p_;

int e_;

};

class Solution {

public:

int maxNumEdgesToRemove(int n, vector<vector<int>>& edges) {

int ans = 0;

DSU A(n), B(n);

for (const auto& e: edges) {

if (e[0] != 3) continue;

ans += A.merge(e[1], e[2]);

B.merge(e[1], e[2]);

}

for (const auto& e: edges) {

if (e[0] == 3) continue;

DSU& d = e[0] == 1 ? A : B;

ans += d.merge(e[1], e[2]);

}

return (A.edges() == n - 1 && B.edges() == n - 1) ? ans : -1;

}

};

python3

class DSU:

def __init__(self, n: int):

self.p = list(range(n))

self.e = 0

def find(self, x: int) -> int:

if x != self.p[x]: self.p[x] = self.find(self.p[x])

return self.p[x]

def merge(self, x: int, y: int) -> int:

rx, ry = self.find(x), self.find(y)

if rx == ry: return 1

self.p[rx] = ry

self.e += 1

return 0

class Solution:

def maxNumEdgesToRemove(self, n: int, edges: List[List[int]]) -> int:

A, B = DSU(n + 1), DSU(n + 1)

ans = 0

for t, x, y in edges:

if t != 3: continue

ans += A.merge(x, y)

B.merge(x, y)

for t, x, y in edges:

if t == 3: continue

d = A if t == 1 else B

ans += d.merge(x, y)

return ans if A.e == B.e == n - 1 else -1