Description
Announcement
- Statements
One-dimensional country has n cities, the i-th of which is located at the point xi and has population pi, and all xi, as well as all pi, are distinct. When one-dimensional country got the Internet, it was decided to place the main server in the largest city, and to connect any other city j to the city k that has bigger population than j and is the closest to it (if there are many such cities, the largest one should be chosen). City k is called a parent of city j in this case.
Unfortunately, the Ministry of Communications got stuck in determining from where and to where the Internet cables should be laid, and the population of the country is suffering. So you should solve the problem. For every city, find its parent city.
Input
The first line contains a single integer n(1 ≤ n ≤ 200000) — the number of cities.
Each of the next n lines contains two space-separated integers xi and pi(1 ≤ xi, pi ≤ 109) — the coordinate and the population of the i-th city.
Output
Output n space-separated integers. The i-th number should be the parent of the i-th city, or - 1, if the i-th city doesn't have a parent. The cities are numbered from 1 by their order in the input.
Sample Input
4
1 1000
7 10
9 1
12 100
-1 4 2 1
3
1 100
2 1
3 10
-1 1 1
3
1 10
3 100
2 1
2 -1 2
题意: 给你n个点在x轴上的位置x和权值pos
#include <iostream>
#include <cstdio>
#include <cstdlib>
#include <cmath>
#include <vector>
#include <stack>
#include <queue>
#include <cstring>
#include <string>
#include <algorithm>
using namespace std;
#define MM(a,b) memset(a,b,sizeof(a));
#define inf 0x7f7f7f7f;
#define FOR(i,n) for(int i=1;i<=n;i++)
#define CT continue;
#define PF printf
const int maxn=200000+10;
struct node{
int x,id,p,f;
}ne[maxn];
int l[maxn],r[maxn];
bool cmpx(node a,node b)
{
return a.x<b.x;
}
bool cmpid(node a,node b)
{
return a.id<b.id;
}
stack<node> stl,str;
int main()
{
int n;
while(~scanf("%d",&n))
{
FOR(i,n) scanf("%d%d",&ne[i].x,&ne[i].p),
ne[i].id=i;
while(stl.size()) stl.pop();
while(str.size()) str.pop();
sort(ne+1,ne+n+1,cmpx);
FOR(i,n){
int id=ne[i].id;
while(stl.size()&&stl.top().p<=ne[i].p) stl.pop();
l[id]=stl.empty()?-1:stl.top().id;
stl.push(ne[i]);
}
for(int i=n;i>=1;i--){
int id=ne[i].id;
while(str.size()&&str.top().p<=ne[i].p) str.pop();
r[id]=str.empty()?-1:str.top().id;
str.push(ne[i]);
}
sort(ne+1,ne+n+1,cmpid);
FOR(i,n){
if(l[i]==-1&&r[i]==-1) {PF("-1 ");CT;}
else if(l[i]==-1) {PF("%d ",r[i]);CT;}
else if(r[i]==-1) {PF("%d ",l[i]);CT;}
if(ne[r[i]].x-ne[i].x>ne[i].x-ne[l[i]].x) PF("%d ",l[i]);
else if(ne[r[i]].x-ne[i].x<ne[i].x-ne[l[i]].x) PF("%d ",r[i]);
else {
if(ne[r[i]].p>ne[l[i]].p) PF("%d ",r[i]);
else PF("%d ",l[i]);
}
}
PF("
");
}
return 0;
}
错因分析:刚开始想用lower_bound的,先按x从小到大排好序后,再利用lower_bound查找到枚举
的点的右侧,比当前枚举点人口数大的第一个点,后来写了后才想起lower_bound是基于二分查找的,
但是按x排好序后的人口数并不呈现有序性,当然不能用lower_bound;
解决:竟然忘了单调栈这么典型的算法,左右各扫一遍,记录大于当前点的最左和最右的点的id,