java8按自定义值自定义映射值分组

时间:2018-05-17 18:44:54

标签: java java-8 java-stream

java8

public class CustomDate{    
    LocalDateTime datetime;   
}

List<CustomDate>  list = // 2 dates with multiple times

我想按天分组日期和以下时间的值将CustomDate作为值?

Map<LocalDate, List<CustomDate>> map = 
         list.stream().groupingBy(d-> d.datetime.toLocalDate())

感谢任何帮助。

3 个答案:

答案 0 :(得分:4)

您似乎希望按日期分组,并将地图值设为LocalTime而不是CustomDate

Map<LocalDate, List<LocalTime>> result = list.stream()
                .collect(Collectors.groupingBy(e -> e.getDatetime().toLocalDate(),
                        Collectors.mapping(e -> e.getDatetime().toLocalTime(),
                                Collectors.toList())));

答案 1 :(得分:0)

正如Aominè的回答一样,您应该使用Collectors.groupingBy()。但我建议你做一些额外的步骤。

1 使用一些常规代码teamplates创建GroupUtils类:

public final class GroupUtils {

    public static <K, V> Map<K, List<V>> groupMultipleBy(Collection<V> data, Function<V, K> classifier) {
        return groupMultipleBy(data, classifier, Function.identity());
    }

    public static <K, V, S> Map<K, List<S>> groupMultipleBy(Collection<V> data, Function<V, K> classifier, Function<V, S> mapper) {
        return Optional.ofNullable(data).orElse(Collections.emptyList()).stream()
                       .collect(Collectors.groupingBy(classifier, Collectors.mapping(mapper, Collectors.toList())));
    }
}

这是我的一个项目的例子。在这里,我有更多类似的方法。

2 在本地文件中,创建具体方法,使客户端代码更具可读性。

class Client {

    public static void main(String... args) {
        List<CustomDate> list = Collections.emptyList();
        Map<LocalDateTime, List<CustomDate>> map = groupByDatetime(list);
    }

    private static Map<LocalDateTime, List<CustomDate>> groupByDatetime(List<CustomDate> list) {
        return GroupUtils.groupMultipleBy(list, CustomDate::getDatetime);        
    }
}

我相信这是一些代码,但更具可读性:

Map<LocalDate, List<LocalTime>> result = list.stream()
                .collect(Collectors.groupingBy(e -> e.getDatetime().toLocalDate(),
                        Collectors.mapping(e -> e.getDatetime().toLocalTime(),
                                Collectors.toList())));

答案 2 :(得分:0)

public class Dates {

    private static List< CustomDate > list = new ArrayList<>(  );

    public static void main( String[] args ) {

        Map< LocalDate, List< LocalTime > > result2 = list.stream()
            .map( CustomDate::getDatetime )
            .collect(
                HashMap::new,
                Dates::accept,
                Dates::merge
            );

    }

    private static void accept( HashMap< LocalDate, List< LocalTime > > map, LocalDateTime date ) {

        map.putIfAbsent( date.toLocalDate( ), new ArrayList<>())
           .add( date.toLocalTime( ) );
    }

    private static void merge( Map< LocalDate, List< LocalTime > > map1, Map< LocalDate, List< LocalTime > > map2 ) {

        map2.forEach( (date,timeList) -> {
            map1.merge( date, timeList, ( l1, l2 ) -> {
                l1.addAll( l2 );
                return l1;
            } );
        });

    }
}

或者简单地说:

Map< LocalDate, List< LocalTime > > result2 = list.stream()
    .map( CustomDate::getDatetime )
    .collect(
        HashMap::new,
        (map,date) -> {
             map.putIfAbsent( date.toLocalDate( ), new ArrayList<>(   )).add( date.toLocalTime( ) );
        },
        (map1, map2) -> {
            map2.forEach( (date,timeList) -> {
                map1.merge( date, timeList, ( l1, l2 ) -> {
                    l1.addAll( l2 );
                    return l1;
                } );
            }); 
        }
    );

这样做不像声明那样调用收集器静态方法的3倍,给你更多的控制来指示行为,比如使用数组的初始大小,使用特定的列表等,而不会丢失来自声明的糖风格。

Map< LocalDate, List< LocalTime > > result2 = list.stream()
    .map( CustomDate::getDatetime )
    .collect(
        TreeMap::new,
        (map,date) -> {
             map.putIfAbsent( date.toLocalDate( ), new ArrayList<>( 800  )).add( date.toLocalTime( ) );
        },
        (map1, map2) -> {
            map2.forEach( (date,timeList) -> {
                map1.merge( date, timeList, ( l1, l2 ) -> {
                    l1.addAll( l2 );
                    return l1;
                } );
            }); 
        }
    );

当然,如果默认行为可以接受,那就更简单了:

    Map<LocalDate, List<LocalTime>> result = list.stream()
            .map( CustomDate::getDatetime )
            .collect(groupingBy(LocalDateTime::toLocalDate,
                     mapping(LocalDateTime::toLocalTime,toList())));

另一种方法是定义自己的收藏家,这是我通常喜欢的。通过这样做,您可以获得控制权和以声明方式编写逻辑的能力,并在需要时重用。

这会将您的代码缩减为:

Map< LocalDate, List< LocalTime > > result2 = list.stream()
            .collect(
                    customDateCollector()
            );

如果你有类似的话:

class CustomDateCollector implements Collector<CustomDate, Map< LocalDate, List< LocalTime > >, Map< LocalDate, List< LocalTime > >> {

    @Override
    public Supplier< Map< LocalDate, List< LocalTime > > > supplier( ) {

        return HashMap::new;
    }

    @Override
    public BiConsumer< Map< LocalDate, List< LocalTime > >, CustomDate > accumulator( ) {
        int goodStartUsingProblemKnowledge = 30;
        return (map, customDate) -> map.putIfAbsent( customDate.getDatetime().toLocalDate(), new ArrayList<>( goodStartUsingProblemKnowledge) )
                                       .add(  customDate.getDatetime().toLocalTime() );
    }

    @Override
    public BinaryOperator< Map< LocalDate, List< LocalTime > > > combiner( ) {

        return (map1, map2) -> {
                map2.forEach( (date,timeList) -> {
                    map1.merge( date, timeList, ( list1, list2 ) -> {
                        list1.addAll( list2 );
                        return list1;
                    } );
                });
                return map1;
        };
    }

    @Override
    public Function< Map< LocalDate, List< LocalTime > >, Map< LocalDate, List< LocalTime > > > finisher( ) {

        return Function.identity();
    }

    @Override
    public Set< Characteristics > characteristics( ) {

        return Collections.unmodifiableSet( EnumSet.of( IDENTITY_FINISH, UNORDERED,CONCURRENT) );
    }

    public static CustomDateCollector customDateCollector(){
        return new CustomDateCollector();
    }
}