这个解决方案不关心正方形或矩形的尺寸，你可以旋转4x5或5x4甚至4x4，它也不关心大小。 注意，这种实现在每次调用rotate90方法时都会创建一个新数组，它根本不会改变原始数组。

public static void main(String[] args) {
    int[][] a = new int[][] { 
                    { 1, 2, 3, 4 }, 
                    { 5, 6, 7, 8 }, 
                    { 9, 0, 1, 2 }, 
                    { 3, 4, 5, 6 }, 
                    { 7, 8, 9, 0 } 
                  };
    int[][] rotate180 = rotate90(rotate90(a));
    print(rotate180);
}

static int[][] rotate90(int[][] a) {
    int[][] ret = new int[a[0].length][a.length];
    for (int i = 0; i < a.length; i++) {
        for (int j = 0; j < a[i].length; j++) {
            ret[j][a.length - i - 1] = a[i][j];
        }
    }
    return ret;
}

static void print(int[][] array) {
    for (int i = 0; i < array.length; i++) {
        System.out.print("[");
        for (int j = 0; j < array[i].length; j++) {
            System.out.print(array[i][j]);
            System.out.print(" ");
        }
        System.out.println("]");
    }
}

可以做递归相当干净，这里是我的实现在golang!

在没有额外内存的情况下递归地旋转go golang中的NXN矩阵

func rot90(a [][]int) {
    n := len(a)
    if n == 1 {
        return
    }
    for i := 0; i < n; i++ {
        a[0][i], a[n-1-i][n-1] = a[n-1-i][n-1], a[0][i]
    }
    rot90(a[1:])
}

JavaScript解决方案旋转矩阵90度的地方:

function rotateBy90(m) {
  var length = m.length;
  //for each layer of the matrix
  for (var first = 0; first < length >> 1; first++) {
    var last = length - 1 - first;
    for (var i = first; i < last; i++) {
      var top = m[first][i]; //store top
      m[first][i] = m[last - i][first]; //top = left
      m[last - i][first] = m[last][last - i]; //left = bottom
      m[last][last - i] = m[i][last]; //bottom = right
      m[i][last] = top; //right = top
    }
  }
  return m;
}

虽然旋转数据可能是必要的(也许是为了更新物理存储的表示)，但在数组访问上添加一层间接层(也许是一个接口)会变得更简单，可能更性能:

interface IReadableMatrix
{
    int GetValue(int x, int y);
}

如果你的矩阵已经实现了这个接口，那么它可以通过这样一个装饰器类来旋转:

class RotatedMatrix : IReadableMatrix
{
    private readonly IReadableMatrix _baseMatrix;

    public RotatedMatrix(IReadableMatrix baseMatrix)
    {
        _baseMatrix = baseMatrix;
    }

    int GetValue(int x, int y)
    {
        // transpose x and y dimensions
        return _baseMatrix(y, x);
    }
}

旋转+90/-90/180度，水平/垂直翻转和缩放都可以以这种方式实现。

Performance would need to be measured in your specific scenario. However the O(n^2) operation has now been replaced with an O(1) call. It's a virtual method call which is slower than direct array access, so it depends upon how frequently the rotated array is used after rotation. If it's used once, then this approach would definitely win. If it's rotated then used in a long-running system for days, then in-place rotation might perform better. It also depends whether you can accept the up-front cost.

与所有性能问题一样，测量，测量，测量!

PHP:

<?php    
$a = array(array(1,2,3,4),array(5,6,7,8),array(9,0,1,2),array(3,4,5,6));
$b = array(); //result

while(count($a)>0)
{
    $b[count($a[0])-1][] = array_shift($a[0]);
    if (count($a[0])==0)
    {
         array_shift($a);
    }
}

从PHP5.6开始，数组转位可以通过一个狡猾的array_map()调用来执行。换句话说，列被转换为行。

代码:(演示)

$array = [
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 0, 1, 2],
    [3, 4, 5, 6]
];
$transposed = array_map(null, ...$array);

美元转置:

[
    [1, 5, 9, 3],
    [2, 6, 0, 4],
    [3, 7, 1, 5],
    [4, 8, 2, 6]
]

我的旋转版本:

void rotate_matrix(int *matrix, int size)
{

int result[size*size];

    for (int i = 0; i < size; ++i)
        for (int j = 0; j < size; ++j)
            result[(size - 1 - i) + j*size] = matrix[i*size+j];

    for (int i = 0; i < size*size; ++i)
        matrix[i] = result[i];
}

在其中，我们将最后一列改为第一行，以此类推。这可能不是最理想的，但对于理解来说是清楚的。