如何在Java中获得两个n维数组的总和？
Posted
技术标签:
【中文标题】如何在Java中获得两个n维数组的总和？【英文标题】：How to get the sum of two n-dimensional arrays in Java? 【发布时间】：2013-11-29 23:57:43 【问题描述】：在 Java 中添加两个 n 维数组
添加向量
Java 中的两个一维数组/向量可以这样添加：
public static int[] addVectors( int[] a, int[] b )
int[] c = new int[a.length];
for ( int i = 0; i < a.length; i++ )
c[i] = a[i] + b[i];
return c;
添加矩阵
Java 中的两个二维数组/矩阵可以这样添加：
public static int[][] addMatrices( int[][] a, int[][] b )
int[][] c = new int[a.length][a[0].length];
for ( int i = 0; i < a.length; i++ )
c[i] = addVectors( a[i], b[i] );
return c;
这两个函数都要求数组大小相同以避免arrayOutOfBoundsException。
addArraysN
应该有一种方法可以使用递归添加两个未知维度的数组。
例如下面的代码使用了假设函数addArraysN( arr1, arr2 )
int[][][] a =
1, 1, 1 , 1, 1, 1 , 1, 1, 1 ,
1, 1, 1 , 1, 1, 1 , 1, 1, 1 ,
1, 1, 1 , 1, 1, 1 , 1, 1, 1
;
int[][][] b =
2, 2, 2 , 2, 2, 2 , 2, 2, 2 ,
2, 2, 2 , 2, 2, 2 , 2, 2, 2 ,
2, 2, 2 , 2, 2, 2 , 2, 2, 2
;
int[][][] sum = addArraysN( a, b );
System.out.println( java.util.Arrays.deepToString( sum ) );
应该输出
[[[3, 3, 3], [3, 3, 3], [3, 3, 3]], [[3, 3, 3], [3, 3, 3], [3, 3, 3]], [[3, 3, 3], [3, 3, 3], [3, 3, 3]]]
现在我想知道如何实现这个功能addArraysN( arr1, arr2 )。
我从以下伪代码开始：
addArraysN( arr1, arr2 )
int dimension = getDimension( arr1 );
if ( dimension == 0 ) //there are no arrays, only numbers
return arr1 + arr2;
else
//create a new arrays with the same dimension and size as arr1 / arr2
//loop through the fields with for
//call itself when adding the fields of arr1 and arr2
//return the sum
可以使用java.lang.reflect.Array 中的newInstance 方法创建新数组。
循环可以像这样实现：
for ( int i = 0; i < ((int[])arr1).length; i++ )
sum = addArraysN( ((int[])arr1)[i], ((int[])arr2)[i] );
问题
但我偶然发现了很多运行时错误和其他问题。有没有人知道如何实现这个addArrayN-method？
也应该可以使用ArrayList 或任何其他类，但我主要对如何使用数组执行此操作感兴趣...（尽管如此，如果有人知道，请发布！）
提前致谢
附录 1
我的原始代码：
import java.util.Arrays;
import java.lang.reflect.Array;
public class ArrayN
public static void main( String[] args )
//Vector
int[] vector1 = 0, 1, 2, 3, 4;
int[] vector2 = 4, 3, 2, 1, 0;
int[] vector3 = ArrayN.addVectors( vector1, vector2 );
for ( int num : vector3 )
System.out.print( num );
System.out.println();
System.out.println();
//Matrix
int[][] matrix1 = 0, 1, 2, 3, 4, 4, 3, 2, 1, 0;
int[][] matrix2 = 4, 3, 2, 1, 0, 0, 1, 2, 3, 4;
int[][] matrix3 = ArrayN.addMatrices( matrix1, matrix2 );
for ( int[] vector : matrix3 )
for ( int num : vector )
System.out.print( num );
System.out.println();
System.out.println();
//N-Array
System.out.println( Arrays.deepToString( (Object[])ArrayN.addArraysN( (Object)matrix1, (Object)matrix2, 2, 5 ) ) );
public static int[] addVectors( int[] a, int[] b )
int[] c = new int[a.length];
for ( int i = 0; i < a.length; i++ )
c[i] = a[i] + b[i];
return c;
public static int[][] addMatrices( int[][] a, int[][] b )
int[][] c = new int[a.length][a[0].length];
for ( int i = 0; i < a.length; i++ )
c[i] = ArrayN.addVectors( a[i], b[i] );
return c;
public static Object addArraysN( Object arrayN1, Object arrayN2, int dimension, int innerlength )
if ( dimension == 0 )
return (int)arrayN1 + (int)arrayN2;
else
int[] dimensions = new int[dimension];
for ( int i = 0; i < dimension; i++ )
dimensions[i] = innerlength;
Object arrayN3 = Array.newInstance( Array.class, dimensions );
for ( int i = 0; i < Array.getLength( arrayN1 ); i++ )
Array.set( arrayN3, i, ArrayN.addArraysN( Array.get( arrayN1, i ), Array.get( arrayN2, i ), dimension-1, innerlength ) );
return arrayN3;
输出：
44444
44444
44444
Exception in thread "main" java.lang.IllegalArgumentException: array element type mismatch
at java.lang.reflect.Array.set(Native Method)
at ArrayN.addArraysN(ArrayN.java:85)
at ArrayN.addArraysN(ArrayN.java:85)
at ArrayN.main(ArrayN.java:41)
附录 2
我发现了错误。它是以下行：
Object arrayN3 = Array.newInstance( Array.class, dimensions );
我不得不将Array.class 替换为int.class。更正后的行应该是：
Object arrayN3 = Array.newInstance( int.class, dimensions );
现在我意识到代码存在的另一个问题： 由于 innerlength 参数，多维数组中的每个数组都必须具有相同的大小。如果数组更短，则其他值为零：
44444
44444
44444
[[4, 4, 4, 4, 4], [4, 4, 4, 4, 4], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0]]
所以我先把matrix1 和matrix2 加长一点：
//Matrix
int[][] matrix1 = 0, 1, 2, 3, 4, 4, 3, 2, 1, 0, 0, 1, 2, 3, 4, 4, 3, 2, 1, 0, 0, 1, 2, 3, 4;
int[][] matrix2 = 4, 3, 2, 1, 0, 0, 1, 2, 3, 4, 4, 3, 2, 1, 0, 0, 1, 2, 3, 4, 4, 3, 2, 1, 0;
但这不是一个好的解决方案。
Nikoloz 写了一个方法来找出数组的维数。使用它和另一种方法arrayToString( Object ) 我现在写的最终代码是：
import java.util.Arrays;
import java.lang.reflect.Array;
import java.util.List;
import java.util.ArrayList;
public class ArrayN
public static void main( String[] args )
int[][] matrix1 = 0, 1, 2, 3, 4, 4, 3, 2, 1, 0;
int[][] matrix2 = 4, 3, 2, 1, 0, 0, 1, 2, 3, 4;
System.out.println( ArrayN.arrayToString( ArrayN.addArraysN( matrix1, matrix2 ) ) );
public static Object addArraysN( Object arrayN1, Object arrayN2 )
ArrayList<Integer> dimensions = new ArrayList<Integer>();
ArrayN.getDimensions( arrayN1, dimensions );
int[] dims = new int[dimensions.size()];
for ( int i = 0; i < dims.length; i++ )
dims[i] = dimensions.get( i );
if ( dims.length == 0 )
return (int)arrayN1 + (int)arrayN2;
else
Object arrayN3 = Array.newInstance( int.class, dims );
for ( int i = 0; i < Array.getLength( arrayN1 ); i++ )
Array.set( arrayN3, i, ArrayN.addArraysN( Array.get( arrayN1, i ), Array.get( arrayN2, i ) ) );
return arrayN3;
public static void getDimensions( Object array, List<Integer> dimensions )
if ( array != null && array.getClass().isArray() )
dimensions.add( Array.getLength( array ) );
if ( Array.getLength( array ) > 0)
ArrayN.getDimensions( Array.get( array, 0 ), dimensions );
public static String arrayToString( Object arr )
if ( arr instanceof byte[] )
return Arrays.toString( (byte[])arr );
else if ( arr instanceof short[] )
return Arrays.toString( (short[])arr );
else if ( arr instanceof int[] )
return Arrays.toString( (int[])arr );
else if ( arr instanceof long[] )
return Arrays.toString( (long[])arr );
else if ( arr instanceof float[] )
return Arrays.toString( (float[])arr );
else if ( arr instanceof double[] )
return Arrays.toString( (double[])arr );
else if ( arr instanceof char[] )
return Arrays.toString( (char[])arr );
else if ( arr instanceof boolean[] )
return Arrays.toString( (boolean[])arr );
else
return Arrays.deepToString( (Object[])arr );
另一种可能性是将维度 1 作为基本情况：
public static Object addArraysN( Object arrayN1, Object arrayN2 )
ArrayList<Integer> dimensions = new ArrayList<Integer>();
ArrayN.getDimensions( arrayN1, dimensions );
int[] dims = new int[dimensions.size()];
for ( int i = 0; i < dims.length; i++ )
dims[i] = dimensions.get( i );
if ( dims.length == 1 )
Object arrayN3 = Array.newInstance( int.class, dims );
for ( int i = 0; i < Array.getLength( arrayN1 ); i++ )
int sum = ((int[])arrayN1)[i] + ((int[])arrayN2)[i];
Array.set( arrayN3, i, sum );
return arrayN3;
else
Object arrayN3 = Array.newInstance( int.class, dims );
for ( int i = 0; i < Array.getLength( arrayN1 ); i++ )
Array.set( arrayN3, i, (int[])ArrayN.addArraysN( Array.get( arrayN1, i ), Array.get( arrayN2, i ) ) );
return arrayN3;
相关问题
How to sum arrays in JavaIs it possible to dynamically build a multi-dimensional array in Java?Iterating over arrays by reflectionJava Reflection - Get size of array objectCreating an n-dimension Array in Java during runtimeInitialising a multidimensional array in Javafinding sum of two dimensional array javaAdding matrices JavaJava Matrices Arrays
【问题讨论】：
只是一个想法：将原始数组包装在一个接受 N 个索引值并将其转换为单个索引以从原始数组中读取的类中要容易得多。它强制执行非参差不齐，允许您通过更改尺寸来重塑形状，并允许您在不复制内容的情况下交换列的顺序。至于您在问题中提到的最后一点：不。你不会得到适用于 List 和数组的东西。它们在 java 中几乎没有关系。 我很困惑，为什么矩阵加法需要反射？您只能将 mxn 矩阵相加，知道一个矩阵的维数意味着您知道和的维数。But I stumbled over a lot of runtime errors and other problems. 究竟是哪些？
【参考方案1】：
这是完整而简单的解决方案。您可以将任何维度数组传递给copyArray 方法。
package com.azry.test;
import java.lang.reflect.Array;
import java.util.ArrayList;
import java.util.List;
public class MultiDimArray
public Object copyArray(Object srcArray1, Object srcArray2)
ArrayList<Integer> dimensions = new ArrayList<Integer>();
getDimensions(srcArray1, dimensions);
int[] dims = new int[dimensions.size()];
for (int i = 0; i < dims.length; i++)
dims[i] = dimensions.get(i);
Object dstArray = Array.newInstance(int.class, dims);
copyArray(srcArray1, srcArray2, dstArray);
return dstArray;
public void copyArray(Object srcArray1, Object srcArray2, Object dstArray)
if (srcArray1 != null && srcArray1.getClass().isArray())
if (srcArray1 instanceof int[])
int[] s1 = (int[])srcArray1;
int[] s2 = (int[])srcArray2;
int[] d = (int[])dstArray;
for (int i = 0; i < s1.length; i++)
d[i] = s1[i] + s2[i];
for (int i = 0; i < Array.getLength(srcArray1); i++)
copyArray(Array.get(srcArray1, i), Array.get(srcArray2, i), Array.get(dstArray, i));
public void getDimensions(Object array, List<Integer> dimensions)
if (array != null && array.getClass().isArray())
dimensions.add(Array.getLength(array));
if (Array.getLength(array) > 0)
getDimensions(Array.get(array, 0), dimensions);
public static void main(String[] args)
int[][][] srcArray1 = new int[2][3][4];
for (int i = 0; i < srcArray1.length; i++)
for (int j = 0; j < srcArray1[i].length; j++)
for (int k = 0; k < srcArray1[i][j].length; k++)
srcArray1[i][j][k] = 2;
int[][][] srcArray2 = new int[2][3][4];
for (int i = 0; i < srcArray2.length; i++)
for (int j = 0; j < srcArray2[i].length; j++)
for (int k = 0; k < srcArray2[i][j].length; k++)
srcArray2[i][j][k] = 3;
int[][][] dstArray = (int[][][])new MultiDimArray().copyArray(srcArray1, srcArray2);
for (int i = 0; i < dstArray.length; i++)
for (int j = 0; j < dstArray[i].length; j++)
for (int k = 0; k < dstArray[i][j].length; k++)
System.out.println("[" + i + "," + j + "," + k + "] = " + dstArray[i][j][k]);
【讨论】：
对数字求和后copyArray不应该返回吗？看起来您正在继续递归并检查每个 Integer 是否都是数组。此外，非常冗长：尝试使其采用另一种类型的数字。而且您还假设所有数组的长度都相同。【参考方案2】：问题是你不能像现在这样处理数组：你在addVectorsN(...) 中收到Object vectorN1，实际上是(Object)matrice1，所以int[][]。但是，您像 (int)vectorN1 一样访问它，这是错误的，因为它实际上是一个数组，而不是一个 int。因此，您应该像这样访问vectorN1：int i = ...; vectorN1[i];，然后我会将vectorN1 保留为int[][]，而不是将其转换为Object。我也会对 vectorN2 做同样的事情。
编辑：
我会改变
addVectorsN( (Object)matrice1, (Object)matrice2, 2, 5 ) )是
addVectorsN( matrice1, matrice2, 2, 5 ) )
你正在像(int)vectorN1一样访问它
public static Object addVectorsN( Object vectorN1, Object vectorN2,
int dimension, int innerlength )
if ( dimension == 0 )
return (int)vectorN1 + (int)vectorN2; (...)
我想改成：
public static Object addVectorsN(int[][] vectorN1, int[][] vectorN2,
int dimension, int innerlength )
EDIT2：
也许你需要这样的东西：
if (vectorN1.length == 1 && vectorN2.length == 1)
if (vectorN1[0].length == 1 && vectorN2[0].length == 1)
return vectorN1[0][0] + vectorN2[0][0];
【讨论】：
这是错误的。如果 vectorN1 和 vectorN2 维度为 0 (vectorN1.length == 0 && vectorN2.length == 0) 那么你不能访问它们，它们是空的。
我编辑了我的代码 (EDIT2) 以反映当 vectorN1 和 vectorN2 都在他们【参考方案3】：
这实际上会满足您的要求：
public static Object sumArbitraryDimensions(Object arrayOne, Object arrayTwo)
if (!arrayOne.getClass().equals(arrayTwo.getClass()) ||
!arrayOne.getClass().isArray())
throw new IllegalArgumentException("not an array");
int len = Array.getLength(arrayOne);
if (len != Array.getLength(arrayTwo))
throw new IllegalArgumentException("incomparable length");
Object sum;
if (arrayOne instanceof Object[])
sum = new Object[len];
for (int i = 0; i < len; i++)
((Object[])sum)[i] = sumArbitraryDimensions(
Array.get(arrayOne, i), Array.get(arrayTwo, i));
else if (arrayOne instanceof byte[])
sum = new byte[len];
for (int i = 0; i < len; i++)
((byte[])sum)[i] = (byte)(((byte[])arrayOne)[i] + ((byte[])arrayTwo)[i]);
else if (arrayOne instanceof short[])
sum = new short[len];
for (int i = 0; i < len; i++)
((short[])sum)[i] = (short)(((short[])arrayOne)[i] + ((short[])arrayTwo)[i]);
else if (arrayOne instanceof int[])
sum = new int[len];
for (int i = 0; i < len; i++)
((int[])sum)[i] = ((int[])arrayOne)[i] + ((int[])arrayTwo)[i];
else if (arrayOne instanceof long[])
sum = new long[len];
for (int i = 0; i < len; i++)
((long[])sum)[i] = ((long[])arrayOne)[i] + ((long[])arrayTwo)[i];
else if (arrayOne instanceof float[])
sum = new float[len];
for (int i = 0; i < len; i++)
((float[])sum)[i] = ((float[])arrayOne)[i] + ((float[])arrayTwo)[i];
else if (arrayOne instanceof double[])
sum = new double[len];
for (int i = 0; i < len; i++)
((double[])sum)[i] = ((double[])arrayOne)[i] + ((double[])arrayTwo)[i];
else
throw new IllegalArgumentException("cannot sum, non-numerical");
return sum;
或者稍微不那么冗长但不那么简洁：
public static Object sumArbitraryDimensions(Object arrayOne, Object arrayTwo)
// exceptions unchecked
int len = Array.getLength(arrayOne);
if (arrayOne instanceof Object[])
Object[] sum = new Object[len];
for (int i = 0; i < len; i++)
((Object[])sum)[i] = sumArbitraryDimensions(
Array.get(arrayOne, i), Array.get(arrayTwo, i));
return sum;
else
Double[] sum = new Double[len];
// works for any numerical primitive type because
// getDouble will perform a widening conversion
for (int i = 0; i < len; i++)
sum[i] = Array.getDouble(arrayOne, i) + Array.getDouble(arrayTwo, i);
return sum;
或者，如果您真的想降低到可能的元素级别：
public static Object sumArbitraryDimensions(Object objectOne, Object objectTwo)
if (!arrayOne.getClass().equals(arrayTwo.getClass()))
throw new IllegalArgumentException("incomparable types");
// reflection cannot retrieve primitive types
// here the returned arrays will be boxed
if (objectOne.getClass().isArray())
int len = Array.getLength(objectOne);
if (len != Array.getLength(objectTwo))
throw new IllegalArgumentException("incomparable lengths");
Object[] sum = new Object[len];
for (int i = 0; i < len; i++)
sum[i] = sumArbitraryDimensions(
Array.get(arrayOne, i), Array.get(arrayTwo, i));
return sum;
else if (objectOne instanceof Byte)
return (byte)((Byte)objectOne + (Byte)objectTwo));
else if (objectOne instanceof Short)
return (short)((Short)objectOne + (Short)objectTwo));
else if (objectOne instanceof Integer)
return (Integer)objectOne + (Integer)objectTwo);
else if (objectOne instanceof Long)
return (Long)objectOne + (Long)objectTwo);
else if (objectOne instanceof Float)
return (Float)objectOne + (Float)objectTwo);
else if (objectOne instanceof Double)
return (Double)objectOne + (Double)objectTwo);
else
throw new IllegalArgumentException("cannot sum, non-numerical");
与常规处理相比，所有这些都将非常缓慢，因为需要进行所有检查。我个人建议不要以任何形式或形式这样做。我看不出它的用途。只需编写最多 N 维的重载方法：
public static int[] sumArrays(int[] arrayOne, int[] arrayTwo)
assert arrayOne != null && arrayTwo != null && arrayOne.length == arrayTwo.length;
int[] sum = new int[Math.min(arrayOne.length, arrayTwo.length)];
for (int i = 0; i < sum.length; i++)
sum[i] = arrayOne[i] + arrayTwo[i];
return sum;
public static int[][] sumArrays(int[][] arrayOne, int[][] arrayTwo)
assert arrayOne != null && arrayTwo != null && arrayOne.length == arrayTwo.length;
int[][] sum = new int[Math.min(arrayOne.length, arrayTwo.length)][];
for (int i = 0, k; i < sum.length; i++)
assert arrayOne[i] != null && arrayTwo[i] != null && arrayOne[i].length == arrayTwo[i].length;
sum[i] = new int[Math.min(arrayOne[i].length, arrayTwo[i].length)];
for (k = 0; k < sum[i].length; k++)
sum[i][k] = arrayOne[i][k] + arrayTwo[i][k];
return sum;
public static int[][][] sumArrays(int[][][] arrayOne, int[][][] arrayTwo)
assert arrayOne != null && arrayTwo != null && arrayOne.length == arrayTwo.length;
int[][][] sum = new int[Math.min(arrayOne.length, arrayTwo.length)][][];
for (int i = 0, k, h; i < sum.length; i++)
assert arrayOne[i] != null && arrayTwo[i] != null && arrayOne[i].length == arrayTwo[i].length;
sum[i] = new int[Math.min(arrayOne[i].length, arrayTwo[i].length)][];
for (k = 0; k < sum[i].length; k++)
assert arrayOne[i][k] != null && arrayTwo[i][k] != null && arrayOne[i][k].length == arrayTwo[i][k].length;
sum[i][k] = new int[Math.min(arrayOne[i][k].length, arrayTwo[i][k].length)];
for (h = 0; h < sum[i][k].length; h++)
sum[i][k][h] = arrayOne[i][k][h] + arrayTwo[i][k][h];
return sum;
【讨论】：
以上是关于如何在Java中获得两个n维数组的总和？的主要内容，如果未能解决你的问题，请参考以下文章