Trainning 1 DAY

You are given a sequence of integers of length nn and integer number kk . You should print any integer number xx in the range of [1;109][1;109] (i.e. 1≤x≤1091≤x≤109 ) such that exactly kk elements of given sequence are less than or equal to xx .

Note that the sequence can contain equal elements.

If there is no such xx , print "-1" (without quotes).

Input

The first line of the input contains integer numbers nn and kk (1≤n≤2⋅1051≤n≤2⋅105 , 0≤k≤n0≤k≤n ). The second line of the input contains nn integer numbers a1,a2,…,ana1,a2,…,an (1≤ai≤1091≤ai≤109 ) — the sequence itself.

Output

Print any integer number xx from range [1;109][1;109] such that exactly kk elements of given sequence is less or equal to xx .

If there is no such xx , print "-1" (without quotes).

Examples

Input

7 4
3 7 5 1 10 3 20

Output

Input

7 2
3 7 5 1 10 3 20

Output

-1

Note

In the first example 55 is also a valid answer because the elements with indices [1,3,4,6][1,3,4,6] is less than or equal to 55 and obviously less than or equal to 66 .

In the second example you cannot choose any number that only 22 elements of the given sequence will be less than or equal to this number because 33 elements of the given sequence will be also less than or equal to this number.

#include<stdio.h>
#include<algorithm>
#include<string.h>
#include<math.h>
#include<iostream>
using namespace std;
int main()
{long long x,n,k,a[200000];scanf("%lld%lld",&n,&k);long long i;for(i=0;i<n;i++){scanf("%lld",&a[i]);}sort(a,a+n);if(k==0)     ///之前没有考虑K为0的情况。{long long b=a[0]-1>=1?a[0]-1:-1;  ///因为X要为正整数。printf("%lld\n",b);}else{if(a[k-1]==a[k]){printf("-1\n");}else{long long t=a[k-1];printf("%lld\n",t);}}
}

Trainning 1 DAY相关推荐

[人工智能-深度学习-46]：FineTuning（微调）、Transfer Trainning（迁移学习）的理论基础与深度解析
作者主页(文火冰糖的硅基工坊):文火冰糖(王文兵)的博客_文火冰糖的硅基工坊_CSDN博客本文网址:https://blog.csdn.net/HiWangWenBing/article/detai ...
Augustus指南(Trainning部分)
目录 Augustus指南官方 Tutorial Index Input: Trainning: Prediction Output: 可能用到的其他软件 Trainning Augustus 1. ...
model trainning 模型训练
学习笔记模型的理解,构建.神经网络的训练使用封闭方程来求解,获得模型在当前训练集上最优的参数==>在训练集上使得损失函数达到最小值时模型的参数使用迭代优化的方法,梯度下降(GD),在训练集 ...
on device trainning
1. On-Device Training with TensorFlow Litehttps://tensorflow.google.cn/lite/examples/on_device_train ...
EMC trainning杂谈
EMC: 中国了解EMC从认证入手.买到欧盟要强制符合EMC标准,上升到法律标准.中国有3C 标准,但在很多行业不是强制性的 .大学很少搞EMC,以前. EMC产生的原因?产生问题,所以要解决问题,所 ...
trainning 2017-11-21
ruby trainning - decryption
2019独角兽企业重金招聘Python工程师标准>>> #author linxu #decrypt the message of cipher.txt module RubyTra ...
ssis trainning
1. 防止包打开后hang住,可以使用delay validation=false. 2.2008R2 configuration 起作用的优先级? 一是des ign time. 二是运行的时候指定 ...
android trainning(4)
一.使用剪贴板在activity间传递信息 public class MainActivity extends Activity {EditText text;@Overrideprotected v ...

Trainning 1 DAY

Trainning 1 DAY相关推荐

最新文章

热门文章