LINQ 基于键和以下记录对记录进行分组

Question

我想知道以下是否可以用LINQ实现。此时我按“MeetingType”分组，但结果集不是我想要的，因为我还想根据会议类型的时间顺序保留分组。

所以基本上在下面的示例中，您将看到有 4 个 PoolDiscussion 会议。在这种情况下，我不希望所有 4 条记录在第 3 次和第 4 次会议之间分组讨论 Pool Discussion 发生了另一个不同类型的会议，所以我想要三个组，其中第一个是 PoolDiscussion （有 3 条记录）其次是景观美化（1 条记录），最后是 PoolDiscussion（1 条记录）

假设我有以下数据记录 数据库记录

MeetingName	MeetingType	MeetingDateTime
With John	Pool Discussion	2020-01-02
With Emily	Pool Discussion	2020-01-02
With Ed	Pool Discussion	2020-01-03
With John	Landscaping	2020-01-03
With Ed	Pool Discussion	2020-01-04

所以在这种情况下，我要做的是在以下结构中输出 JSON（注意如何有 2 个“PoolDiscussions”分组）

{
    "Meetings": [
        {
            "PoolDiscussion": [
                {
                    "MeetingName": "With John",
                    "MeetingDate": "2020-01-02"
                },
                {
                    "MeetingName": "With Emily",
                    "MeetingDate": "2020-01-02"
                },
                {
                    "MeetingName": "With Ed",
                    "MeetingDate": "2020-01-03"
                }
            ]
        },
        {
            "Landscaping": [
                {
                    "MeetingName": "With John",
                    "MeetingDate": "2020-01-03"
                }
            ]
        },
        {
            "PoolDiscussion": [
                {
                    "MeetingName": "With Ed",
                    "MeetingDate": "2020-01-04"
                }
            ]
        }
    ]
}

Answer 1

我不认为您可以通过单个 LINQ 查询完全完成您想要的。原因是您没有唯一的分组依据。您基本上是在尝试按会议类型和日期进行分组，因此您可以接近以下内容，但是您将在 3 日的“With Ed”会议上获得额外的“PoolDiscussion”记录...

from meeting in data
group meeting by new { Type = meeting.MeetingType, Date = meeting.MeetingDateTime } into mts
select new { Key = mts.Key.Type, Value = mts.ToList() }).ToList()

这将产生以下内容：

[
        {
            "PoolDiscussion": [
                {
                    "MeetingName": "With John",
                    "MeetingDate": "2020-01-02"
                },
                {
                    "MeetingName": "With Emily",
                    "MeetingDate": "2020-01-02"
                }
             ]
        },
        {  "PoolDiscussion": [
                {
                    "MeetingName": "With Ed",
                    "MeetingDate": "2020-01-03"
                }
            ]
        },
        {
            "Landscaping": [
                {
                    "MeetingName": "With John",
                    "MeetingDate": "2020-01-03"
                }
            ]
        },
        {
            "PoolDiscussion": [
                {
                    "MeetingName": "With Ed",
                    "MeetingDate": "2020-01-04"
                }
            ]
        }
    ]

一旦你有了这个数据集，你就可以遍历它并折叠相邻的“PoolDiscussion”记录

Answer 2

这是一个简单的部分 LINQ 解决方案，但我想知道一种更高效或完全基于 LINQ 的方法：

var groupedMeetings = new List<KeyValuePair<string, List<Meeting>>>();
while (meetings.Any())
{
    var group = meetings.TakeWhile((meeting, index) =>
    {
        return index == 0 || meeting.MeetingType == meetings[index - 1].MeetingType;
    });
    
    groupedMeetings.Add(KeyValuePair.Create(group.First().MeetingType, group.ToList()));

    // Skip past the meetings we've already processed.
    meetings = meetings.SkipWhile((meeting, index) =>
    {
        return index == 0 || meeting.MeetingType == meetings[index - 1].MeetingType;
    }).ToList();
}

这是一种非 LINQ 方法进行比较：

var groupedMeetings = new List<KeyValuePair<string, List<Meeting>>>();

var i = 0;
while (i < meetings.Count)
{
    var meetingType = meetings[i].MeetingType;

    // A group will always consist of at least one meeting.
    var group = new List<Meeting>
    {
        meetings[i++]
    };

    while (i < meetings.Count - 1 && meetings[i].MeetingType == meetings[i - 1].MeetingType)
    {
        group.Add(meetings[i++]);
    }

    groupedMeetings.Add(KeyValuePair.Create(meetingType, group));
}

这两种方法都假定会议在处理之前已经正确排序，并保留该排序。

假设以下类：

public class Meeting
{
    public string MeetingName { get; set; }

    [JsonIgnore]
    public string MeetingType { get; set; }
    
    public DateTime MeetingDate { get; set; }
}

这个：

var settings = new JsonSerializerSettings
{
    Formatting = Formatting.Indented,
    DateFormatString = "yyyy-MM-dd"
};
Console.WriteLine(JsonConvert.SerializeObject(groupedMeetings, settings));

生成与您的示例等效的 JSON：

[
  {
    "Key": "Pool Discussion",
    "Value": [
      {
        "MeetingName": "With John",
        "MeetingDate": "2020-01-02"
      },
      {
        "MeetingName": "With Emily",
        "MeetingDate": "2020-01-02"
      },
      {
        "MeetingName": "With Ed",
        "MeetingDate": "2020-01-03"
      }
    ]
  },
  {
    "Key": "Landscaping",
    "Value": [
      {
        "MeetingName": "With John",
        "MeetingDate": "2020-01-03"
      }
    ]
  },
  {
    "Key": "Pool Discussion",
    "Value": [
      {
        "MeetingName": "With Ed",
        "MeetingDate": "2020-01-04"
      }
    ]
  }
]