This task is quite simple! You are given a string STR of length N where each character of string is from the set {A, B, C, D, E, F, G, H, I, J}. Now you need to delete exactly M characters from this String STR. On deleting a character from this String, STR, the neighboring characters of this just deleted character become new neighbors.

For example, Consider the String $abcd$. On deleting the second character of this String, the new resultant string is : $acd$. On deleting the first character of this resultant String, the new resultant String is : $cd$. The relative order of the rest of the non-deleted characters does not change due to any deletion. Each character of String STR can be deleted at most once.

Now, there can be many ways to delete M characters. After deleting any M characters from this String STR the rest of the non-deleted characters come together to from a String S. For example, given a String $abcd$, you need to delete any 2 characters from this String. One of the ways to do so is deleting the $1^{st}$ and the $3^{rd}$ characters. After doing so, the resultant String S is $bd$.

So, your task here is to delete any M characters from String STR, such that the resultant String S is lexicographically largest over all possible ways to delete exactly M characters.

Print this lexicographically largest String as your answer on a new line. The length of this String has to be exactly $N-M$

Input :

The first line contains two space separated positive integers N and M.
Second line contains string STR.

Output :

Output the lexicographically largest string that can be produced from STR by deleting exactly M characters from it.

Constraint:

$1 \le M \le N \le 5 \times 10^5$