将Hashmap分配给Hashmap

Question

我有一个哈希地图,我想复制其他用途.但每当我复制并重复使用它时,它也会改变原来的.这是为什么？

    do {
            Map<Integer, Map<String, Object>> map1 = originalMap; 
            //at the second iteration originalMap is the same as map1 of the last iteration, 
            //eventhough the change was nog accepted;
            //do something with map1 (change value);
            if(change is accepted) {
               originalMap = map1;
            }
        } while(iteration < 10);

提前致谢

    public static <Integer,String, Schedule>Map<Integer, Map<String, Schedule>> deepCopy(Map<Integer, Map<String, Schedule>> original) {
    Map<Integer, Map<String, Schedule>> copy = new HashMap<Integer, Map<String, Schedule>>();

    for (Map.Entry<Integer, Map<String, Schedule>> entry : original.entrySet()) {
        copy.put(entry.getKey(), deepCopy2(entry.getValue()));
    }
    return copy;
}

public static <String, Schedule>Map<String, Schedule> deepCopy2(Map<String, Schedule> original) {
    Map<String, Schedule> copy = new HashMap<String, Schedule>();
    for (Map.Entry<String, Schedule> entry : original.entrySet()) {
        copy.put(entry.getKey(), entry.getValue());
    }

    return copy;
}

Answer 1

你所做的不是创建地图的副本,而是创建地图的副本.当两个引用指向同一个对象时,对一个引用的更改将反映在另一个引用中.

解决方案1:如果这是一个从简单类型到另一个类型的Map,那么您可以这样做:

Map<SomeType, OtherType> map1 = new HashMap<SomeType, OtherType>(original); 

这称为复制构造函数.几乎所有标准的Collection和Map实现都有一个,它通常是克隆简单结构的最简单方法.这将工作,只要罚款,SomeType并OtherType是不可变的(例如,Integer和其他Number类型Boolean,String但不集合,日期,地图,阵列等)

如果没有,正如其他答复者和评论者指出的那样,您还需要复制地图值.

解决方案2:这是一个安全的快速和脏的版本:

Map<Integer, Map<String, Object>> original=new HashMap<Integer, Map<String,Object>>();
Map<Integer, Map<String, Object>> copy = 
        new HashMap<Integer, Map<String, Object>>();
for(Entry<Integer, Map<String, Object>> entry : original.entrySet()){
    copy.put(entry.getKey(), new HashMap<String, Object>(entry.getValue()));
}

但实际上,我喜欢Hunter提供深度复制方法的想法.所以这里的解决方案3:我自己的版本使用通用参数:

public static <K1, K2, V> Map<K1, Map<K2, V>> deepCopy(
    Map<K1, Map<K2, V>> original){

    Map<K1, Map<K2, V>> copy = new HashMap<K1, Map<K2, V>>();
    for(Entry<K1, Map<K2, V>> entry : original.entrySet()){
        copy.put(entry.getKey(), new HashMap<K2, V>(entry.getValue()));
    }
    return copy;
}

你可以这样称呼它:

Map<Integer, Map<String, Object>> original=new HashMap<Integer, Map<String,Object>>();
// do stuff here
Map<Integer, Map<String, Object>> copy = deepCopy(original);

更新

我已经攻击了一个为地图,集合和数组(原始和其他)执行深度克隆的类.用法:

Something clone = DeepClone.deepClone(original);

这里是:

public final class DeepClone {

    private DeepClone(){}

    public static <X> X deepClone(final X input) {
        if (input == null) {
            return input;
        } else if (input instanceof Map<?, ?>) {
            return (X) deepCloneMap((Map<?, ?>) input);
        } else if (input instanceof Collection<?>) {
            return (X) deepCloneCollection((Collection<?>) input);
        } else if (input instanceof Object[]) {
            return (X) deepCloneObjectArray((Object[]) input);
        } else if (input.getClass().isArray()) {
            return (X) clonePrimitiveArray((Object) input);
        }

        return input;
    }

    private static Object clonePrimitiveArray(final Object input) {
        final int length = Array.getLength(input);
        final Object copy = Array.newInstance(input.getClass().getComponentType(), length);
        // deep clone not necessary, primitives are immutable
        System.arraycopy(input, 0, copy, 0, length);
        return copy;
    }

    private static <E> E[] deepCloneObjectArray(final E[] input) {
        final E[] clone = (E[]) Array.newInstance(input.getClass().getComponentType(), input.length);
        for (int i = 0; i < input.length; i++) {
            clone[i] = deepClone(input[i]);
        }

        return clone;
    }

    private static <E> Collection<E> deepCloneCollection(final Collection<E> input) {
        Collection<E> clone;
        // this is of course far from comprehensive. extend this as needed
        if (input instanceof LinkedList<?>) {
            clone = new LinkedList<E>();
        } else if (input instanceof SortedSet<?>) {
            clone = new TreeSet<E>();
        } else if (input instanceof Set) {
            clone = new HashSet<E>();
        } else {
            clone = new ArrayList<E>();
        }

        for (E item : input) {
            clone.add(deepClone(item));
        }

        return clone;
    }

    private static <K, V> Map<K, V> deepCloneMap(final Map<K, V> map) {
        Map<K, V> clone;
        // this is of course far from comprehensive. extend this as needed
        if (map instanceof LinkedHashMap<?, ?>) {
            clone = new LinkedHashMap<K, V>();
        } else if (map instanceof TreeMap<?, ?>) {
            clone = new TreeMap<K, V>();
        } else {
            clone = new HashMap<K, V>();
        }

        for (Entry<K, V> entry : map.entrySet()) {
            clone.put(deepClone(entry.getKey()), deepClone(entry.getValue()));
        }

        return clone;
    }
}

Answer 2

通过做这个:

Map<Integer, Map<String, Object>> copy = originalMap;

...你没有复制地图,只创建一个引用完全相同地图的新变量,显然你使用这个变量所做的更改将反映在原始地图中 - 它们指向同一个对象记忆.最好使用接收另一个地图作为参数的构造函数复制原始地图:

Map<Integer, Map<String, Object>> copy;
copy = new HashMap<Integer, Map<String, Object>>(originalMap);

上面的代码将创建原始地图的浅表副本,这意味着:如果更改一个地图中的元素值,则更改将反映在另一个地图中,但您可以自由地添加/删除地图中的元素和其他不会受到影响.如果这还不够好,您需要在复制时对地图中的元素执行深层复制.