Virtual contest is a way to take part in past contest, as close as possible to participation on time. It is supported only ACM-ICPC mode for virtual contests.
If you've seen these problems, a virtual contest is not for you - solve these problems in the archive.
If you just want to solve some problem from a contest, a virtual contest is not for you - solve this problem in the archive.
Never use someone else's code, read the tutorials or communicate with other person during a virtual contest.

No tag edit access

C. Laboratory Work

time limit per test

1 secondmemory limit per test

256 megabytesinput

standard inputoutput

standard outputAnya and Kirill are doing a physics laboratory work. In one of the tasks they have to measure some value *n* times, and then compute the average value to lower the error.

Kirill has already made his measurements, and has got the following integer values: *x*_{1}, *x*_{2}, ..., *x*_{n}. It is important that the values are close to each other, namely, the difference between the maximum value and the minimum value is at most 2.

Anya does not want to make the measurements, however, she can't just copy the values from Kirill's work, because the error of each measurement is a random value, and this coincidence will be noted by the teacher. Anya wants to write such integer values *y*_{1}, *y*_{2}, ..., *y*_{n} in her work, that the following conditions are met:

- the average value of
*x*_{1},*x*_{2}, ...,*x*_{n}is equal to the average value of*y*_{1},*y*_{2}, ...,*y*_{n}; - all Anya's measurements are in the same bounds as all Kirill's measurements, that is, the maximum value among Anya's values is not greater than the maximum value among Kirill's values, and the minimum value among Anya's values is not less than the minimum value among Kirill's values;
- the number of equal measurements in Anya's work and Kirill's work is as small as possible among options with the previous conditions met. Formally, the teacher goes through all Anya's values one by one, if there is equal value in Kirill's work and it is not strike off yet, he strikes off this Anya's value and one of equal values in Kirill's work. The number of equal measurements is then the total number of strike off values in Anya's work.

Help Anya to write such a set of measurements that the conditions above are met.

Input

The first line contains a single integer *n* (1 ≤ *n* ≤ 100 000) — the numeber of measurements made by Kirill.

The second line contains a sequence of integers *x*_{1}, *x*_{2}, ..., *x*_{n} ( - 100 000 ≤ *x*_{i} ≤ 100 000) — the measurements made by Kirill. It is guaranteed that the difference between the maximum and minimum values among values *x*_{1}, *x*_{2}, ..., *x*_{n} does not exceed 2.

Output

In the first line print the minimum possible number of equal measurements.

In the second line print *n* integers *y*_{1}, *y*_{2}, ..., *y*_{n} — the values Anya should write. You can print the integers in arbitrary order. Keep in mind that the minimum value among Anya's values should be not less that the minimum among Kirill's values, and the maximum among Anya's values should be not greater than the maximum among Kirill's values.

If there are multiple answers, print any of them.

Examples

Input

6

-1 1 1 0 0 -1

Output

2

0 0 0 0 0 0

Input

3

100 100 101

Output

3

101 100 100

Input

7

-10 -9 -10 -8 -10 -9 -9

Output

5

-10 -10 -9 -9 -9 -9 -9

Note

In the first example Anya can write zeros as here measurements results. The average value is then equal to the average value of Kirill's values, and there are only two equal measurements.

In the second example Anya should write two values 100 and one value 101 (in any order), because it is the only possibility to make the average be the equal to the average of Kirill's values. Thus, all three measurements are equal.

In the third example the number of equal measurements is 5.

Codeforces (c) Copyright 2010-2019 Mike Mirzayanov

The only programming contests Web 2.0 platform

Server time: Mar/19/2019 09:41:42 (d3).

Desktop version, switch to mobile version.

Supported by

User lists

Name |
---|