我对Java比较陌生,经常发现需要对值进行Map<Key,Value>排序。

由于这些值不是唯一的,我发现自己将keySet转换为一个数组,并使用自定义比较器通过数组排序对该数组进行排序,该比较器根据与该键关联的值进行排序。

有没有更简单的方法?


当前回答

这个问题已经有了很多答案,但没有一个能为我提供我想要的,一个返回按关联值排序的键和条目的映射实现,并在映射中修改键和值时维护这个属性。另外两个问题对此提出了具体要求。

我编写了一个通用友好的示例来解决这个用例。此实现不遵守Map接口的所有约定,例如反映原始对象中keySet()和entrySet()返回的集合中的值更改和删除。我觉得这样的解决方案太大,无法包含在堆栈溢出的答案中。如果我成功地创建了一个更完整的实现,也许我会将其发布到Github,然后在这个答案的更新版本中链接到它。

import java.util.*;

/**
 * A map where {@link #keySet()} and {@link #entrySet()} return sets ordered
 * by associated values based on the the comparator provided at construction
 * time. The order of two or more keys with identical values is not defined.
 * <p>
 * Several contracts of the Map interface are not satisfied by this minimal
 * implementation.
 */
public class ValueSortedMap<K, V> extends HashMap<K, V> {
    protected Map<V, Collection<K>> valueToKeysMap;

    // uses natural order of value object, if any
    public ValueSortedMap() {
        this((Comparator<? super V>) null);
    }

    public ValueSortedMap(Comparator<? super V> valueComparator) {
        this.valueToKeysMap = new TreeMap<V, Collection<K>>(valueComparator);
    }

    public boolean containsValue(Object o) {
        return valueToKeysMap.containsKey(o);
    }

    public V put(K k, V v) {
        V oldV = null;
        if (containsKey(k)) {
            oldV = get(k);
            valueToKeysMap.get(oldV).remove(k);
        }
        super.put(k, v);
        if (!valueToKeysMap.containsKey(v)) {
            Collection<K> keys = new ArrayList<K>();
            keys.add(k);
            valueToKeysMap.put(v, keys);
        } else {
            valueToKeysMap.get(v).add(k);
        }
        return oldV;
    }

    public void putAll(Map<? extends K, ? extends V> m) {
        for (Map.Entry<? extends K, ? extends V> e : m.entrySet())
            put(e.getKey(), e.getValue());
    }

    public V remove(Object k) {
        V oldV = null;
        if (containsKey(k)) {
            oldV = get(k);
            super.remove(k);
            valueToKeysMap.get(oldV).remove(k);
        }
        return oldV;
    }

    public void clear() {
        super.clear();
        valueToKeysMap.clear();
    }

    public Set<K> keySet() {
        LinkedHashSet<K> ret = new LinkedHashSet<K>(size());
        for (V v : valueToKeysMap.keySet()) {
            Collection<K> keys = valueToKeysMap.get(v);
            ret.addAll(keys);
        }
        return ret;
    }

    public Set<Map.Entry<K, V>> entrySet() {
        LinkedHashSet<Map.Entry<K, V>> ret = new LinkedHashSet<Map.Entry<K, V>>(size());
        for (Collection<K> keys : valueToKeysMap.values()) {
            for (final K k : keys) {
                final V v = get(k);
                ret.add(new Map.Entry<K,V>() {
                    public K getKey() {
                        return k;
                    }

                    public V getValue() {
                        return v;
                    }

                    public V setValue(V v) {
                        throw new UnsupportedOperationException();
                    }
                });
            }
        }
        return ret;
    }
}

其他回答

基于@devinmore代码,一种使用泛型并支持升序和降序排序的map排序方法。

/**
 * Sort a map by it's keys in ascending order. 
 *  
 * @return new instance of {@link LinkedHashMap} contained sorted entries of supplied map.
 * @author Maxim Veksler
 */
public static <K, V> LinkedHashMap<K, V> sortMapByKey(final Map<K, V> map) {
    return sortMapByKey(map, SortingOrder.ASCENDING);
}

/**
 * Sort a map by it's values in ascending order.
 *  
 * @return new instance of {@link LinkedHashMap} contained sorted entries of supplied map.
 * @author Maxim Veksler
 */
public static <K, V> LinkedHashMap<K, V> sortMapByValue(final Map<K, V> map) {
    return sortMapByValue(map, SortingOrder.ASCENDING);
}

/**
 * Sort a map by it's keys.
 *  
 * @param sortingOrder {@link SortingOrder} enum specifying requested sorting order. 
 * @return new instance of {@link LinkedHashMap} contained sorted entries of supplied map.
 * @author Maxim Veksler
 */
public static <K, V> LinkedHashMap<K, V> sortMapByKey(final Map<K, V> map, final SortingOrder sortingOrder) {
    Comparator<Map.Entry<K, V>> comparator = new Comparator<Entry<K,V>>() {
        public int compare(Entry<K, V> o1, Entry<K, V> o2) {
            return comparableCompare(o1.getKey(), o2.getKey(), sortingOrder);
        }
    };

    return sortMap(map, comparator);
}

/**
 * Sort a map by it's values.
 *  
 * @param sortingOrder {@link SortingOrder} enum specifying requested sorting order. 
 * @return new instance of {@link LinkedHashMap} contained sorted entries of supplied map.
 * @author Maxim Veksler
 */
public static <K, V> LinkedHashMap<K, V> sortMapByValue(final Map<K, V> map, final SortingOrder sortingOrder) {
    Comparator<Map.Entry<K, V>> comparator = new Comparator<Entry<K,V>>() {
        public int compare(Entry<K, V> o1, Entry<K, V> o2) {
            return comparableCompare(o1.getValue(), o2.getValue(), sortingOrder);
        }
    };

    return sortMap(map, comparator);
}

@SuppressWarnings("unchecked")
private static <T> int comparableCompare(T o1, T o2, SortingOrder sortingOrder) {
    int compare = ((Comparable<T>)o1).compareTo(o2);

    switch (sortingOrder) {
    case ASCENDING:
        return compare;
    case DESCENDING:
        return (-1) * compare;
    }

    return 0;
}

/**
 * Sort a map by supplied comparator logic.
 *  
 * @return new instance of {@link LinkedHashMap} contained sorted entries of supplied map.
 * @author Maxim Veksler
 */
public static <K, V> LinkedHashMap<K, V> sortMap(final Map<K, V> map, final Comparator<Map.Entry<K, V>> comparator) {
    // Convert the map into a list of key,value pairs.
    List<Map.Entry<K, V>> mapEntries = new LinkedList<Map.Entry<K, V>>(map.entrySet());

    // Sort the converted list according to supplied comparator.
    Collections.sort(mapEntries, comparator);

    // Build a new ordered map, containing the same entries as the old map.  
    LinkedHashMap<K, V> result = new LinkedHashMap<K, V>(map.size() + (map.size() / 20));
    for(Map.Entry<K, V> entry : mapEntries) {
        // We iterate on the mapEntries list which is sorted by the comparator putting new entries into 
        // the targeted result which is a sorted map. 
        result.put(entry.getKey(), entry.getValue());
    }

    return result;
}

/**
 * Sorting order enum, specifying request result sort behavior.
 * @author Maxim Veksler
 *
 */
public static enum SortingOrder {
    /**
     * Resulting sort will be from smaller to biggest.
     */
    ASCENDING,
    /**
     * Resulting sort will be from biggest to smallest.
     */
    DESCENDING
}

好的,这个版本使用两个新的Map对象和两次迭代,并对值进行排序。希望,虽然地图条目必须循环两次,但表现良好:

public static void main(String[] args) {
    Map<String, String> unsorted = new HashMap<String, String>();
    unsorted.put("Cde", "Cde_Value");
    unsorted.put("Abc", "Abc_Value");
    unsorted.put("Bcd", "Bcd_Value");

    Comparator<String> comparer = new Comparator<String>() {
        @Override
        public int compare(String o1, String o2) {
            return o1.compareTo(o2);
        }};

    System.out.println(sortByValue(unsorted, comparer));

}

public static <K, V> Map<K,V> sortByValue(Map<K, V> in, Comparator<? super V> compare) {
    Map<V, K> swapped = new TreeMap<V, K>(compare);
    for(Entry<K,V> entry: in.entrySet()) {
        if (entry.getValue() != null) {
            swapped.put(entry.getValue(), entry.getKey());
        }
    }
    LinkedHashMap<K, V> result = new LinkedHashMap<K, V>();
    for(Entry<V,K> entry: swapped.entrySet()) {
        if (entry.getValue() != null) {
            result.put(entry.getValue(), entry.getKey());
        }
    }
    return result;
}

该解决方案使用带有比较器的TreeMap,并对所有空键和值进行排序。首先,使用TreeMap中的排序功能对值进行排序,然后使用排序后的Map创建一个结果,因为LinkedHashMap保留了相同的值顺序。

格里兹,GHad

因为地图是无序的要对其进行排序,我们可以执行以下操作

Map<String, String> map= new TreeMap<String, String>(unsortMap);

您应该注意,与哈希映射不同,树映射保证其元素将按升序键排序。

对键进行排序需要Comparator为每个比较查找每个值。一个更具可扩展性的解决方案将直接使用entrySet,因为这样每次比较都会立即获得该值(尽管我没有用数字来支持)。

这是这样一件事的通用版本:

public static <K, V extends Comparable<? super V>> List<K> getKeysSortedByValue(Map<K, V> map) {
    final int size = map.size();
    final List<Map.Entry<K, V>> list = new ArrayList<Map.Entry<K, V>>(size);
    list.addAll(map.entrySet());
    final ValueComparator<V> cmp = new ValueComparator<V>();
    Collections.sort(list, cmp);
    final List<K> keys = new ArrayList<K>(size);
    for (int i = 0; i < size; i++) {
        keys.set(i, list.get(i).getKey());
    }
    return keys;
}

private static final class ValueComparator<V extends Comparable<? super V>>
                                     implements Comparator<Map.Entry<?, V>> {
    public int compare(Map.Entry<?, V> o1, Map.Entry<?, V> o2) {
        return o1.getValue().compareTo(o2.getValue());
    }
}

有一些方法可以减少上述解决方案的内存旋转。例如,创建的第一个ArrayList可以重新用作返回值;这将需要抑制一些泛型警告,但对于可重用的库代码来说,这可能是值得的。此外,Comparator不必在每次调用时重新分配。

这里有一个更有效但不太吸引人的版本:

public static <K, V extends Comparable<? super V>> List<K> getKeysSortedByValue2(Map<K, V> map) {
    final int size = map.size();
    final List reusedList = new ArrayList(size);
    final List<Map.Entry<K, V>> meView = reusedList;
    meView.addAll(map.entrySet());
    Collections.sort(meView, SINGLE);
    final List<K> keyView = reusedList;
    for (int i = 0; i < size; i++) {
        keyView.set(i, meView.get(i).getKey());
    }
    return keyView;
}

private static final Comparator SINGLE = new ValueComparator();

最后,如果您需要连续访问已排序的信息(而不是偶尔排序一次),可以使用额外的多重映射。如果你需要更多细节,请告诉我。。。

    Map<String, Integer> map = new HashMap<>();
    map.put("b", 2);
    map.put("a", 1);
    map.put("d", 4);
    map.put("c", 3);
    
    // ----- Using Java 7 -------------------
    List<Map.Entry<String, Integer>> entries = new ArrayList<>(map.entrySet());
    Collections.sort(entries, (o1, o2) -> o1.getValue().compareTo(o2.getValue()));
    System.out.println(entries); // [a=1, b=2, c=3, d=4]


    // ----- Using Java 8 Stream API --------
   map.entrySet().stream().sorted(Map.Entry.comparingByValue()).forEach(System.out::println); // {a=1, b=2, c=3, d=4}