Codefroces #279 div2 B.Queue



B. Queue

time limit per test

2 seconds

memory limit per test

256 megabytes

input

standard input

output

standard output

During the lunch break all n Berland State University students lined up in the food court. However, it turned out that the food court, too, has a lunch break and it temporarily stopped working.

Standing in a queue that isn't being served is so boring! So, each of the students wrote down the number of the student ID of the student that stands in line directly in front of him, and the student that stands in line directly behind him. If no one stands before or after a student (that is, he is the first one or the last one), then he writes down number 0 instead (in Berland State University student IDs are numerated from 1).

After that, all the students went about their business. When they returned, they found out that restoring the queue is not such an easy task.

Help the students to restore the state of the queue by the numbers of the student ID's of their neighbors in the queue.

Input

The first line contains integer n (2 ≤ n ≤ 2·10⁵) — the number of students in the queue.

Then n lines follow, i-th line contains the pair of integers a_i, b_i (0 ≤ a_i, b_i ≤ 10⁶), where a_i is the ID number of a person in front of a student and b_i is the ID number of a person behind a student. The lines are given in the arbitrary order. Value 0 is given instead of a neighbor's ID number if the neighbor doesn't exist.

The ID numbers of all students are distinct. It is guaranteed that the records correspond too the queue where all the students stand in some order.

Output

Print a sequence of n integers x₁, x₂, ..., x_n — the sequence of ID numbers of all the students in the order they go in the queue from the first student to the last one.

Sample test(s)

Input

4
92 31
0 7
31 0
7 141

Output

92 7 31 141 

Note

The picture illustrates the queue for the first sample.

 解题思路：这道题其实就是模拟，但是也有一定的技巧，那就是匹配了，首先，我们把队伍分成奇数号和偶数号，然后很容易发现，在匹配偶数位置的队员的时候，往往是从 0开始匹配的，而匹配奇数位置的队员的时候是从1开始匹配的，这样每次的跳表，就能匹配结束。


代码：

#include<cstdio>
#include<iostream>

using namespace std;

#define N 1000010

int a[N], b[N], ans[N], vis[N], tmp[N];
int n;

int main() {
  scanf("%d", &n);
  for (int i = 1; i <= n; i++) {
    int fr, ba;
    scanf("%d%d", &fr, &ba);
    a[fr] = ba;
    b[ba] = fr;
    vis[fr] = vis[ba] = 1;
  }
  
  //对于偶数的查找
  int k = 2;
  for (int i = a[0]; i; i = a[i], k += 2) {
    ans[k] = i;
    vis[i] = 0;
  }
  for (k = 1;; k++) {
    if (vis[k])
      break;
  }
  while (b[k])
  {
    k = b[k];
  }
  
  
  //对于奇数的查找
  int fzc = k;
  k = 1;
  for (int i = fzc; i; i = a[i], k += 2) {
    ans[k] = i;
  }
  for (int i = 1; i <= n; i++)
    printf("%d%s", ans[i], (i == n) ? "\n" : " ");
  return 0;
}