java - 如何对多个元素上的 Vector> 进行排序，将子项分组在一起？

Question

因此，我从 REST API 返回 JSON 数据，其中数据已按日期字段排序，然后添加到 JTable 中使用的 Vector> 中。我需要将具有相同 5 位数 key 的项目分组在一起，然后按附加标识符降序顺序 (3,2,1) 对这些项目进行排序，其中日期决定下一个 5 位数 key 是什么。

例如:

[[28696, 2, 11/15/19 17:57]]
[[28696, 1, 11/15/19 17:56]]
[[00972, 2, 11/15/19 17:55]]
[[28696, 3, 11/15/19 17:54]]
[[00972, 1, 11/15/19 17:53]]

应该是

[[28696, 3, 11/15/19 17:54]]
[[28696, 2, 11/15/19 17:57]]
[[28696, 1, 11/15/19 17:56]]
[[00972, 2, 11/15/19 17:55]]
[[00972, 1, 11/15/19 17:53]]

看到 28696 是该组中的第一个项目，也是整个数据集中最新的项目。

Answer 1

你想要的不是一个 sort 就能实现的手术。相反，您可以使用流来满足分组、组排序以及组内数据排序的要求。

首先请注意，我转换了您的 List<List<String>>进入List<Data> (根据 Max Vollmer 的建议)使其更易于使用和理解。如果您确实需要它作为 List您可以替换 Data与 List<String>和 Data.get??()与 List.get(?) :

public static void main(String[] args) throws IOException, TransformerException {
    List<Data> list = Arrays.asList( //
            new Data("28696", "2", "11/15/19 17:57"), //
            new Data("28696", "1", "11/15/19 17:56"), //
            new Data("00972", "2", "11/15/19 17:55"), //
            new Data("28696", "3", "11/15/19 17:54"), //
            new Data("00972", "1", "11/15/19 17:53"));

    Comparator<Data> sortByTimestamp = Comparator.comparing(Data::getTimestamp).reversed();
    Comparator<Data> sortByOccurrence = Comparator.comparing(Data::getOccurrence).reversed();

    // LinkedHashMap keeps the insertion order
    Collector<Data, ?, LinkedHashMap<String, List<Data>>> groupByIdKeepInsertOrder
            = Collectors.groupingBy(Data::getId, LinkedHashMap::new, Collectors.toList());

    List<Data> result = list.stream()
            .sorted(sortByTimestamp) // sort all data by timestamp, if it's already sorted by timestamp you can skip this
            .collect(groupByIdKeepInsertOrder) // group them by id keeping the timestamp order
            .values().stream() // stream the lists of data grouped together
            .peek(l -> l.sort(sortByOccurrence)) // sort each list of data by occurrence
            .flatMap(Collection::stream) // flatten the lists into a single stream
            .collect(Collectors.toList()); // collect all Data into a single list

    System.out.println(result);
    // [[28696, 3, 11/15/19 17:54],
    //  [28696, 2, 11/15/19 17:57],
    //  [28696, 1, 11/15/19 17:56],
    //  [00972, 2, 11/15/19 17:55],
    //  [00972, 1, 11/15/19 17:53]]
}

<小时/>

private static class Data {
    private final String id;
    private final String occurrence;
    private final String timestamp;

    public Data(String id, String occurrence, String timestamp) {
        this.id = id;
        this.occurrence = occurrence;
        this.timestamp = timestamp;
    }

    public String getId() { return id; }
    public String getOccurrence() { return occurrence; }
    public String getTimestamp() { return timestamp; }
    @Override public String toString() { return "[" + id + ", " + occurrence + ", " + timestamp + "]";}
}