An "$$$n-$$$Latin Square" is an $$$n\times n$$$ square, each ceil contains one integer in range $$$[1,n]$$$. All integers in a row or a column are different. Formally, an "$$$n-$$$Latin Square" can be seen as a $$$2-$$$dimensions array $$$a$$$, we have the set equation $$$\{a_{i,1},a_{i,2},...,a_{i,n}\}=\{1,2,...,n\}$$$ for $$$i=1,2,...,n$$$ and $$$\{a_{1,j},a_{2,j},...,a_{n,j}\}=\{1,2,...,n\}$$$ for $$$j=1,2,...,n$$$.
The first line of the input contains two integers $$$n, m(1\le m \lt n\le 100)$$$ separated by space, which describes the size of the square, and the number of rows Sakuyalove has written.
Next $$$m$$$ lines, the $$$i$$$-th line contains $$$n$$$ integers $$$a_{i,1}, a_{i,2},...,a_{i,n}(1\le a_{i,j}\le n)$$$, which describes the integers Sakuyalove has written on the $$$i$$$-th rows.
It is guaranteed that the rows given by Sakuyalove are correct. Formally, we have the set equation $$$\{a_{i,1},a_{i,2},...,a_{i,n}\}=\{1,2,...,n\}$$$ for $$$i=1,2,...,m$$$, and $$$a_{1,j},a_{2,j},...,a_{m, j}$$$ are different integers in range$$$[1,n]$$$ for $$$j=1,2,...,n$$$.
The output should contains $$$n$$$ lines, each line contains $$$n$$$ integers separated by space, which described the "$$$n-$$$Latin Square" you built.
You should notice that the first $$$m$$$ lines of the output should be same as the first $$$m$$$ rows given by Sakuyalove.
It can be proved that it exists a way to build the $$$n-$$$Latin Square on the basis of the first $$$m$$$ rows given by Sakuyalove.
If it has more than one solution, you can print any.
5 3 1 2 5 4 3 3 5 4 1 2 4 3 1 2 5
1 2 5 4 3 3 5 4 1 2 4 3 1 2 5 2 1 3 5 4 5 4 2 3 1
The sample input and output are like this:
