There are n people in a social group labeled from 0 to n - 1. You are given an array logs where logs[i] = [timestampi, xi, yi] indicates that xi and yi will be friends at the time timestampi.
Friendship is symmetric. That means if a is friends with b, then b is friends with a. Also, person a is acquainted with a person b if a is friends with b, or a is a friend of someone acquainted with b.
Return the earliest time for which every person became acquainted with every other person. If there is no such earliest time, return -1.
Example 1:
Input: logs = [[20190101,0,1],[20190104,3,4],[20190107,2,3],[20190211,1,5],[20190224,2,4],[20190301,0,3],[20190312,1,2],[20190322,4,5]], n = 6
Output: 20190301
Explanation:
The first event occurs at timestamp = 20190101, and after 0 and 1 become friends, we have the following friendship groups [0,1], [2], [3], [4], [5].
The second event occurs at timestamp = 20190104, and after 3 and 4 become friends, we have the following friendship groups [0,1], [2], [3,4], [5].
The third event occurs at timestamp = 20190107, and after 2 and 3 become friends, we have the following friendship groups [0,1], [2,3,4], [5].
The fourth event occurs at timestamp = 20190211, and after 1 and 5 become friends, we have the following friendship groups [0,1,5], [2,3,4].
The fifth event occurs at timestamp = 20190224, and as 2 and 4 are already friends, nothing happens.
The sixth event occurs at timestamp = 20190301, and after 0 and 3 become friends, we all become friends.
Example 2:
Input: logs = [[0,2,0],[1,0,1],[3,0,3],[4,1,2],[7,3,1]], n = 4
Output: 3
Explanation: At timestamp = 3, all the persons (i.e., 0, 1, 2, and 3) become friends.
Constraints:
2 <= n <= 100
1 <= logs.length <= 104
logs[i].length == 3
0 <= timestampi <= 109
0 <= xi, yi <= n - 1
xi != yi
All the values timestampi are unique.
All the pairs (xi, yi) occur at most one time in the input.
Solutions
Solution 1: Sorting + Union-Find
Thinking
We need the earliest time when all \(n\) people lie in one connected component. Logs arrive unordered, so sort them by timestamp and process in order.
If the two people are still in different sets, union them and decrement the component count. Union-find makes the find/merge step nearly constant. When the count reaches \(1\), everyone is connected and that timestamp is the answer; if logs run out first, return \(-1\).
We sort all the logs in ascending order by timestamp, then traverse the sorted logs. Using a union-find set, we check whether the two people in the current log are already friends. If they are not friends, we merge them into one friend circle, until everyone is in one friend circle, then return the timestamp of the current log.
If we have traversed all the logs and not everyone is in one friend circle, then return \(-1\).
The time complexity is \(O(n \times \log n)\), and the space complexity is \(O(n)\). Here, \(n\) is the number of logs.
Method 1 inlines path compression and does not union by size. Method 2 wraps a UnionFind with path compression and union-by-size; union reports whether a merge happened, and the main loop only tracks the component count. The moment everyone becomes friends is unchanged, with shallower trees.