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?
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
在python中:
import numpy as np
a = np.array(
[
[1, 2, 3, 4],
[5, 6, 7, 8],
[9, 0, 1, 2],
[3, 4, 5, 6]
]
)
print(a)
print(b[::-1, :].T)
其他回答
这是一个空间旋转方法,由java编写,只适用于正方形。对于非正方形的2d数组,无论如何都必须创建新数组。
private void rotateInSpace(int[][] arr) {
int z = arr.length;
for (int i = 0; i < z / 2; i++) {
for (int j = 0; j < (z / 2 + z % 2); j++) {
int x = i, y = j;
int temp = arr[x][y];
for (int k = 0; k < 4; k++) {
int temptemp = arr[y][z - x - 1];
arr[y][z - x - 1] = temp;
temp = temptemp;
int tempX = y;
y = z - x - 1;
x = tempX;
}
}
}
}
通过创建新数组旋转任何大小的2d数组的代码:
private int[][] rotate(int[][] arr) {
int width = arr[0].length;
int depth = arr.length;
int[][] re = new int[width][depth];
for (int i = 0; i < depth; i++) {
for (int j = 0; j < width; j++) {
re[j][depth - i - 1] = arr[i][j];
}
}
return re;
}
正如我在上一篇文章中所说的,这里有一些c#代码,可以为任何大小的矩阵实现O(1)矩阵旋转。为了简洁性和可读性,没有错误检查或范围检查。代码:
static void Main (string [] args)
{
int [,]
// create an arbitrary matrix
m = {{0, 1}, {2, 3}, {4, 5}};
Matrix
// create wrappers for the data
m1 = new Matrix (m),
m2 = new Matrix (m),
m3 = new Matrix (m);
// rotate the matricies in various ways - all are O(1)
m1.RotateClockwise90 ();
m2.Rotate180 ();
m3.RotateAnitclockwise90 ();
// output the result of transforms
System.Diagnostics.Trace.WriteLine (m1.ToString ());
System.Diagnostics.Trace.WriteLine (m2.ToString ());
System.Diagnostics.Trace.WriteLine (m3.ToString ());
}
class Matrix
{
enum Rotation
{
None,
Clockwise90,
Clockwise180,
Clockwise270
}
public Matrix (int [,] matrix)
{
m_matrix = matrix;
m_rotation = Rotation.None;
}
// the transformation routines
public void RotateClockwise90 ()
{
m_rotation = (Rotation) (((int) m_rotation + 1) & 3);
}
public void Rotate180 ()
{
m_rotation = (Rotation) (((int) m_rotation + 2) & 3);
}
public void RotateAnitclockwise90 ()
{
m_rotation = (Rotation) (((int) m_rotation + 3) & 3);
}
// accessor property to make class look like a two dimensional array
public int this [int row, int column]
{
get
{
int
value = 0;
switch (m_rotation)
{
case Rotation.None:
value = m_matrix [row, column];
break;
case Rotation.Clockwise90:
value = m_matrix [m_matrix.GetUpperBound (0) - column, row];
break;
case Rotation.Clockwise180:
value = m_matrix [m_matrix.GetUpperBound (0) - row, m_matrix.GetUpperBound (1) - column];
break;
case Rotation.Clockwise270:
value = m_matrix [column, m_matrix.GetUpperBound (1) - row];
break;
}
return value;
}
set
{
switch (m_rotation)
{
case Rotation.None:
m_matrix [row, column] = value;
break;
case Rotation.Clockwise90:
m_matrix [m_matrix.GetUpperBound (0) - column, row] = value;
break;
case Rotation.Clockwise180:
m_matrix [m_matrix.GetUpperBound (0) - row, m_matrix.GetUpperBound (1) - column] = value;
break;
case Rotation.Clockwise270:
m_matrix [column, m_matrix.GetUpperBound (1) - row] = value;
break;
}
}
}
// creates a string with the matrix values
public override string ToString ()
{
int
num_rows = 0,
num_columns = 0;
switch (m_rotation)
{
case Rotation.None:
case Rotation.Clockwise180:
num_rows = m_matrix.GetUpperBound (0);
num_columns = m_matrix.GetUpperBound (1);
break;
case Rotation.Clockwise90:
case Rotation.Clockwise270:
num_rows = m_matrix.GetUpperBound (1);
num_columns = m_matrix.GetUpperBound (0);
break;
}
StringBuilder
output = new StringBuilder ();
output.Append ("{");
for (int row = 0 ; row <= num_rows ; ++row)
{
if (row != 0)
{
output.Append (", ");
}
output.Append ("{");
for (int column = 0 ; column <= num_columns ; ++column)
{
if (column != 0)
{
output.Append (", ");
}
output.Append (this [row, column].ToString ());
}
output.Append ("}");
}
output.Append ("}");
return output.ToString ();
}
int [,]
// the original matrix
m_matrix;
Rotation
// the current view of the matrix
m_rotation;
}
好的,我把手举起来,当旋转时,它实际上不会对原始数组做任何修改。但是,在面向对象系统中,只要对象看起来像是被旋转到类的客户端,这就无关紧要了。目前,Matrix类使用对原始数组数据的引用,因此改变m1的任何值也将改变m2和m3。对构造函数稍加更改,创建一个新数组并将值复制到该数组中,就可以将其整理出来。
这是c#的
int[,] array = new int[4,4] {
{ 1,2,3,4 },
{ 5,6,7,8 },
{ 9,0,1,2 },
{ 3,4,5,6 }
};
int[,] rotated = RotateMatrix(array, 4);
static int[,] RotateMatrix(int[,] matrix, int n) {
int[,] ret = new int[n, n];
for (int i = 0; i < n; ++i) {
for (int j = 0; j < n; ++j) {
ret[i, j] = matrix[n - j - 1, i];
}
}
return ret;
}
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();
}
我只用一个循环就能做到。时间复杂度看起来像O(K)其中K是数组中的所有元素。 下面是我用JavaScript做的:
首先,我们用一个数组来表示n^2矩阵。然后,像这样迭代它:
/**
* Rotates matrix 90 degrees clockwise
* @param arr: the source array
* @param n: the array side (array is square n^2)
*/
function rotate (arr, n) {
var rotated = [], indexes = []
for (var i = 0; i < arr.length; i++) {
if (i < n)
indexes[i] = i * n + (n - 1)
else
indexes[i] = indexes[i - n] - 1
rotated[indexes[i]] = arr[i]
}
return rotated
}
基本上,我们转换源数组下标:
[0,1,2,3,4,5,6,7,8] => [2,5,8,1,4,7,0,3 6]
然后,使用这个转换后的索引数组,我们将实际值放在最终旋转的数组中。
下面是一些测试用例:
//n=3
rotate([
1, 2, 3,
4, 5, 6,
7, 8, 9], 3))
//result:
[7, 4, 1,
8, 5, 2,
9, 6, 3]
//n=4
rotate([
1, 2, 3, 4,
5, 6, 7, 8,
9, 10, 11, 12,
13, 14, 15, 16], 4))
//result:
[13, 9, 5, 1,
14, 10, 6, 2,
15, 11, 7, 3,
16, 12, 8, 4]
//n=5
rotate([
1, 2, 3, 4, 5,
6, 7, 8, 9, 10,
11, 12, 13, 14, 15,
16, 17, 18, 19, 20,
21, 22, 23, 24, 25], 5))
//result:
[21, 16, 11, 6, 1,
22, 17, 12, 7, 2,
23, 18, 13, 8, 3,
24, 19, 14, 9, 4,
25, 20, 15, 10, 5]