[CodeForces - 197C] C - Lexicographically Maximum Subsequence

C - Lexicographically Maximum Subsequence

You've got string s, consisting of only lowercase English letters. Find its lexicographically maximum subsequence.

We'll call a non-empty string s[p₁p₂... p_k] = s_p1s_p2... s_pk(1 ≤ p₁ < p₂ < ... < p_k ≤ |s|) a subsequence of string s = s₁s₂... s_|s|.

String x = x₁x₂... x_|x| is lexicographically larger than string y = y₁y₂... y_|y|, if either |x| > |y| and x₁ = y₁, x₂ = y₂, ... , x_|y| = y_|y|, or exists such number r (r < |x|, r < |y|), that x₁ = y₁, x₂ = y₂, ... , x_r = y_r and x_r + 1 > y_r + 1. Characters in lines are compared like their ASCII codes.

Input

The single line contains a non-empty string s, consisting only of lowercase English letters. The string's length doesn't exceed 10⁵.

Output

Print the lexicographically maximum subsequence of string s.

Example

Input

ababba

Output

bbba

Input

abbcbccacbbcbaaba

Output

cccccbba

Note

Let's look at samples and see what the sought subsequences look like (they are marked with uppercase bold letters).

The first sample: aBaBBA

The second sample: abbCbCCaCbbCBaaBA

这个题目就是要求出子串中字典序最大的.

由于本题有许多方法,那我就讲一讲我自己的好了.

由于我们要尽量挑字典序最大的字符作为开头(首关键字并不是长度!),所以,我们要记录当前字符串的最大字符,然后,如果剩下的字符里面,已经不存在这个最大的字符,那么这个字符就下调,知道存在.

这样一个纯模拟的过程就好了,不过不知道为什么我在比赛中时间用的稍微有点多......

 #include<cstdio>
 #include<cstring>
 #include<algorithm>
 using namespace std;
 ],nxt[],R[];
 ],c;
 int read(){
     ,f=; char ch=getchar();
     '){if (ch=='-') f=-f; ch=getchar();}
     +ch-',ch=getchar();
     return x*f;
 }
 int main(){
     scanf("%s",a),len=strlen(a),c='a';
     memset(las,,sizeof las);
     memset(nxt,,sizeof nxt);
     ; i>=; i--) ) las[a[i]-'a']=i,nxt[i]=i,R[a[i]-'a']=i;
     else nxt[i]=las[a[i]-'a'],las[a[i]-'a']=i;
     ; i<len; i++) if (c<a[i]) c=a[i];
     ; i<len; i++) if (a[i]==c){
         putchar(c); if (nxt[i]!=i) continue;
         );
     }
     ;
 }