受Raymond Chen帖子的启发,假设你有一个4x4二维数组,写一个函数使它旋转90度。Raymond链接到伪代码中的解决方案,但我想看到一些真实的东西。

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

就变成:

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

更新:Nick的答案是最直接的,但是有没有比n²更好的方法呢?如果矩阵是10000x10000呢?


当前回答

这是Java中的一个更好的版本:我已经为一个具有不同宽度和高度的矩阵制作了它

H是旋转后矩阵的高度 W是旋转后矩阵的宽度

 

public int[][] rotateMatrixRight(int[][] matrix)
{
    /* W and H are already swapped */
    int w = matrix.length;
    int h = matrix[0].length;
    int[][] ret = new int[h][w];
    for (int i = 0; i < h; ++i) {
        for (int j = 0; j < w; ++j) {
            ret[i][j] = matrix[w - j - 1][i];
        }
    }
    return ret;
}


public int[][] rotateMatrixLeft(int[][] matrix)
{
    /* W and H are already swapped */
    int w = matrix.length;
    int h = matrix[0].length;   
    int[][] ret = new int[h][w];
    for (int i = 0; i < h; ++i) {
        for (int j = 0; j < w; ++j) {
            ret[i][j] = matrix[j][h - i - 1];
        }
    }
    return ret;
}

此代码基于Nick Berardi的帖子。

其他回答

试试我图书馆的算盘——常见的:

@Test
public void test_42519() throws Exception {
    final IntMatrix matrix = IntMatrix.range(0, 16).reshape(4);

    N.println("======= original =======================");
    matrix.println();
    // print out:
    //    [0, 1, 2, 3]
    //    [4, 5, 6, 7]
    //    [8, 9, 10, 11]
    //    [12, 13, 14, 15]

    N.println("======= rotate 90 ======================");
    matrix.rotate90().println();
    // print out:
    //    [12, 8, 4, 0]
    //    [13, 9, 5, 1]
    //    [14, 10, 6, 2]
    //    [15, 11, 7, 3]

    N.println("======= rotate 180 =====================");
    matrix.rotate180().println();
    // print out:
    //    [15, 14, 13, 12]
    //    [11, 10, 9, 8]
    //    [7, 6, 5, 4]
    //    [3, 2, 1, 0]

    N.println("======= rotate 270 ======================");
    matrix.rotate270().println();
    // print out:
    //    [3, 7, 11, 15]
    //    [2, 6, 10, 14]
    //    [1, 5, 9, 13]
    //    [0, 4, 8, 12]

    N.println("======= transpose =======================");
    matrix.transpose().println();
    // print out:
    //    [0, 4, 8, 12]
    //    [1, 5, 9, 13]
    //    [2, 6, 10, 14]
    //    [3, 7, 11, 15]

    final IntMatrix bigMatrix = IntMatrix.range(0, 10000_0000).reshape(10000);

    // It take about 2 seconds to rotate 10000 X 10000 matrix.
    Profiler.run(1, 2, 3, "sequential", () -> bigMatrix.rotate90()).printResult();

    // Want faster? Go parallel. 1 second to rotate 10000 X 10000 matrix.
    final int[][] a = bigMatrix.array();
    final int[][] c = new int[a[0].length][a.length];
    final int n = a.length;
    final int threadNum = 4;

    Profiler.run(1, 2, 3, "parallel", () -> {
        IntStream.range(0, n).parallel(threadNum).forEach(i -> {
            for (int j = 0; j < n; j++) {
                c[i][j] = a[n - j - 1][i];
            }
        });
    }).printResult();
}

#转置是Ruby的Array类的标准方法,因此:

% irb
irb(main):001:0> m = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 0, 1, 2], [3, 4, 5, 6]]
=> [[1, 2, 3, 4], [5, 6, 7, 8], [9, 0, 1, 2], [3, 4, 5, 6]] 
irb(main):002:0> m.reverse.transpose
=> [[3, 9, 5, 1], [4, 0, 6, 2], [5, 1, 7, 3], [6, 2, 8, 4]]

实现是一个用c写的n^2转置函数,你可以在这里看到: http://www.ruby-doc.org/core-1.9.3/Array.html#method-i-transpose 通过选择“点击切换源”旁边的“转置”。

我记得比O(n^2)的解更好,但只适用于特殊构造的矩阵(如稀疏矩阵)

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!

#include <iostream>
#include <iomanip>

using namespace std;
const int SIZE=3;
void print(int a[][SIZE],int);
void rotate(int a[][SIZE],int);

void main()
{
    int a[SIZE][SIZE]={{11,22,33},{44,55,66},{77,88,99}};
    cout<<"the array befor rotate\n";

    print(a,SIZE);
    rotate( a,SIZE);
    cout<<"the array after rotate\n";
    print(a,SIZE);
    cout<<endl;

}

void print(int a[][SIZE],int SIZE)
{
    int i,j;
    for(i=0;i<SIZE;i++)
       for(j=0;j<SIZE;j++)
          cout<<a[i][j]<<setw(4);
}

void rotate(int a[][SIZE],int SIZE)
{
    int temp[3][3],i,j;
    for(i=0;i<SIZE;i++)
       for(j=0;j<SIZE/2.5;j++)
       {
           temp[i][j]= a[i][j];
           a[i][j]= a[j][SIZE-i-1] ;
           a[j][SIZE-i-1] =temp[i][j];

       }
}

Nick的答案也适用于NxM阵列,只需要做一点修改(与NxN相反)。

string[,] orig = new string[n, m];
string[,] rot = new string[m, n];

...

for ( int i=0; i < n; i++ )
  for ( int j=0; j < m; j++ )
    rot[j, n - i - 1] = orig[i, j];

考虑这个问题的一种方法是将轴(0,0)的中心从左上角移动到右上角。你只是简单地从一个转置到另一个。