Qualifying Contest

CodeForces - 659B

Very soon Berland will hold a School Team Programming Olympiad. From each of the mBerland regions a team of two people is invited to participate in the olympiad. The qualifying contest to form teams was held and it was attended by n Berland students. There were at least two schoolboys participating from each of the m regions of Berland. The result of each of the participants of the qualifying competition is an integer score from 0 to 800 inclusive.

The team of each region is formed from two such members of the qualifying competition of the region, that none of them can be replaced by a schoolboy of the same region, not included in the team and who received a greater number of points. There may be a situation where a team of some region can not be formed uniquely, that is, there is more than one school team that meets the properties described above. In this case, the region needs to undertake an additional contest. The two teams in the region are considered to be different if there is at least one schoolboy who is included in one team and is not included in the other team. It is guaranteed that for each region at least two its representatives participated in the qualifying contest.

Your task is, given the results of the qualifying competition, to identify the team from each region, or to announce that in this region its formation requires additional contests.

Input

The first line of the input contains two integers n and m (2 ≤ n ≤ 100 000, 1 ≤ m ≤ 10 000, n ≥ 2m) — the number of participants of the qualifying contest and the number of regions in Berland.

Next n lines contain the description of the participants of the qualifying contest in the following format: Surname (a string of length from 1 to 10 characters and consisting of large and small English letters), region number (integer from 1 to m) and the number of points scored by the participant (integer from 0 to 800, inclusive).

It is guaranteed that all surnames of all the participants are distinct and at least two people participated from each of the m regions. The surnames that only differ in letter cases, should be considered distinct.

Output

Print m lines. On the i-th line print the team of the i-th region — the surnames of the two team members in an arbitrary order, or a single character "?" (without the quotes) if you need to spend further qualifying contests in the region.

Examples

Input
5 2
Ivanov 1 763
Andreev 2 800
Petrov 1 595
Sidorov 1 790
Semenov 2 503
Output
Sidorov Ivanov
Andreev Semenov
Input
5 2
Ivanov 1 800
Andreev 2 763
Petrov 1 800
Sidorov 1 800
Semenov 2 503
Output
?
Andreev Semenov

Note

In the first sample region teams are uniquely determined.

In the second sample the team from region 2 is uniquely determined and the team from region 1 can have three teams: "Petrov"-"Sidorov", "Ivanov"-"Sidorov", "Ivanov" -"Petrov", so it is impossible to determine a team uniquely.

sol:STL非常好用,按照分数从大到小排序后如果第二名分数和第三名相同就无法确定了

#include <bits/stdc++.h>
using namespace std;
typedef int ll;
inline ll read()
{
ll s=;
bool f=;
char ch=' ';
while(!isdigit(ch))
{
f|=(ch=='-'); ch=getchar();
}
while(isdigit(ch))
{
s=(s<<)+(s<<)+(ch^); ch=getchar();
}
return (f)?(-s):(s);
}
#define R(x) x=read()
inline void write(ll x)
{
if(x<)
{
putchar('-'); x=-x;
}
if(x<)
{
putchar(x+''); return;
}
write(x/);
putchar((x%)+'');
return;
}
#define W(x) write(x),putchar(' ')
#define Wl(x) write(x),putchar('\n')
const int N=,B=;
int n,m;
struct Record
{
char Name[B];
int Score;
};
inline bool cmp(Record p,Record q)
{
return p.Score>q.Score;
}
vector<Record>Vec[N];
char Name[B];
int main()
{
int i;
R(n); R(m);
for(i=;i<=n;i++)
{
Record tmp;
int id;
scanf("%s",tmp.Name);
R(id);
R(tmp.Score);
Vec[id].push_back(tmp);
}
for(i=;i<=m;i++)
{
sort(Vec[i].begin(),Vec[i].end(),cmp);
if(Vec[i].size()==)
{
puts("?");
}
else if(Vec[i].size()==)
{
printf("%s",Vec[i][].Name);
putchar(' ');
printf("%s",Vec[i][].Name);
putchar('\n');
}
else
{
if(Vec[i][].Score==Vec[i][].Score) puts("?");
else
{
printf("%s",Vec[i][].Name);
putchar(' ');
printf("%s",Vec[i][].Name);
putchar('\n');
}
}
}
return ;
}
/*
input
5 2
Ivanov 1 763
Andreev 2 800
Petrov 1 595
Sidorov 1 790
Semenov 2 503
output
Sidorov Ivanov
Andreev Semenov input
5 2
Ivanov 1 800
Andreev 2 763
Petrov 1 800
Sidorov 1 800
Semenov 2 503
output
?
Andreev Semenov
*/

codeforces659B的更多相关文章

随机推荐

  1. zabbix4.0安装和配置

    以下为master端安装,开始 内核$ cat /etc/centos-release CentOS Linux release 7.4.1708 (Core) $ uname -r 3.10.0-6 ...

  2. 简单的策略模式Strategy演示

    策略模式,即规则在变化之中,结果终归为一. 公司给员工计算工资,如有加班费,差旅费,每个月的生活补帖等等其它费用需要计算.这个费的规则是不尽相同. 不管策略的规则怎样,终归需要计算出一个结果 工资: ...

  3. 学习angularjs的内置API函数

    angularjs的内置API函数有很多,如isString()判断给定的对象是否为字符串,如果是返回 true,反之返回false:isNumber()判断给定的对象是否为数字,如果是返回 true ...

  4. EZ 2018 05 20 NOIP2018 模拟赛(十五)

    这次的比赛充满着玄学的气息,玄学链接 首先讲一下为什么没有第十四场 其实今天早上9点时看到题目就叫了:原题! 没错,整套试卷都做过,我还写了题解 然后老叶就说换一套,但如果仅仅是这样就没什么 但等13 ...

  5. [Oracle]如何取Control File 的Dump

    ]如何取Control File 的Dump: SQL> alter session set events 'immediate trace name controlf level 3';SQL ...

  6. Centos 6.9下部署Oracle 11G数据库环境的操作记录

    操作系统:Centos6.9(64Bit)Oracle:11g .11.2.0.4.0版本Ip地址:172.16.220.139 废话不多说了,下面记录安装过程:1)安装桌面环境 [root@vm01 ...

  7. 分布式监控系统Zabbix-3.0.3-完整安装记录(0)

    一.Linux下开源监控系统简单介绍1)cacti:存储数据能力强,报警性能差2)nagios:报警性能差,存储数据仅有简单的一段可以判断是否在合理范围内的数据长度,储存在内存中.比如,连续采样数据存 ...

  8. Redis常用操作-------Set(集合)

    1.SADD key member [member ...] 将一个或多个 member 元素加入到集合 key 当中,已经存在于集合的 member 元素将被忽略. 假如 key 不存在,则创建一个 ...

  9. Linux内核分析作业六

    1.阅读理解task_struct数据结构 2.分析fork函数对应的内核处理过程sys_clone,理解创建一个新进程如何创建和修改task_struct数据结构: fork进程的代码 #inclu ...

  10. 我的Android之路——底部菜单栏的实现

    底部菜单栏的实现 底部菜单栏两种实现方法:ViewPager:可滑动的界面:Fragment:固定的界面. 首先,页面布局,在除去顶部toolbar之后,将主界面分为两部分,一部分为界面显示区,另一部 ...