Question

我有一个名为“VideoData”的后端表，其中包含以下格式的数据：

    VideoID       RecordingStarted         RecordingEnded
    ==============================================================
    abc123        2013-03-01 15:30:00      2013-03-01 15:40:00        
    def123        2013-03-06 12:00:00      2013-03-06 12:40:00
    ijk123        2013-03-10 11:00:00      2013-03-10 11:05:00
    klm123        2013-03-12 10:05:00      2013-03-12 10:25:00
    And list goes on .......
    .......................
    .............................

使用实体框架我希望获得2013年3月捕获的视频总时数，以便捕获视频的总小时数为Weekwise。以下列方式举例：

    Mar 1, 2013     Mar 8, 2013      Mar 15, 2013     Mar 22, 2013
   ================================================================
   500              300              350              200

我用Google搜索了很多内容，但无法弄清楚如何正确地做到这一点。请指导。谢谢你的帮助。

Answer 1

我讨厌推翻Ilya出色的工作（对不起伙伴，至少我赞成你），但由于这是实体框架，因此可以让数据库通过SqlFunctions只用一个查询来完成工作：< / p>

context.Videos.Select(t => new
  {
      Year = SqlFunctions.DatePart("yyyy", t.RecordingStarted),
      Week = SqlFunctions.DatePart("ww", t.RecordingStarted),
      Hours = SqlFunctions.DateDiff("hh", t.RecordingStarted, t.RecordingEnded)
  })
.GroupBy(x => new { x.Year, x.Week} )
.Select (x => new { x.Key.Year, x.Key.Week, TotalHours = x.Sum(p => p.Hours)} )

输出类似于

2013  9 500
2013 10 300
...

在Sql Server中，从年份+周到日期非常困难。如果这确实是一个要求，您可以考虑建立一个参考表，其中包含年份+周数和第一天的日期。或者获取内存中的数据（.ToList()）并使用C＃转换数据。

Answer 2

好吧，这项任务似乎非常有趣。我已经使用本地数据实现了基本功能。免费评论更多功能。你有点麻烦

//local data for testing
var data = new[] {
    new {VideoId = "abc123",RecordindStarted = DateTime.Parse("2013-03-01 15:30:00"),RecordingEnded = DateTime.Parse("2013-03-01 15:40:00")},        
    new {VideoId = "def123",RecordindStarted = DateTime.Parse("2013-03-06 12:00:00"),RecordingEnded = DateTime.Parse("2013-03-06 12:40:00")},
    new {VideoId = "de1223",RecordindStarted = DateTime.Parse("2013-03-06 12:30:00"),RecordingEnded = DateTime.Parse("2013-03-06 12:50:00")},
    new {VideoId = "ijk123",RecordindStarted = DateTime.Parse("2013-03-10 11:00:00"),RecordingEnded = DateTime.Parse("2013-03-10 11:05:00")},
    new {VideoId = "klm123",RecordindStarted = DateTime.Parse("2013-03-12 10:05:00"),RecordingEnded = DateTime.Parse("2013-03-12 10:25:00")},
    new {VideoId = "klm123",RecordindStarted = DateTime.Parse("2013-03-12 10:05:00"),RecordingEnded = DateTime.Parse("2013-03-13 10:25:00")},
};


var calendar = new GregorianCalendar();

var ungroupedTotalHours = data.Select(d => GetHoursPerDays(d.RecordindStarted, d.RecordingEnded));

var groupedTotalHours = 
         ungroupedTotalHours.SelectMany(v => v)
                            .GroupBy(v=> v.Key)
                            .ToDictionary(v => v.Key, v => v.Sum(val => val.Value));

var result = 
 groupedTotalHours.GroupBy(v =>calendar.GetWeekOfYear( v.Key, CalendarWeekRule.FirstDay, DayOfWeek.Monday))
                  .ToDictionary(v => "Week "+ v.Key, row => row.Sum(val => val.Value));

Console.WriteLine ( string.Join(Environment.NewLine, result.Select(r => r.Key +" has "+r.Value+" hours")) );

核心逻辑转向此方法：

public IDictionary<DateTime, int> GetHoursPerDays(DateTime start, DateTime end)
{
    if(end.Date == start.Date)
        return new Dictionary<DateTime, int>{{start.Date, (end -  start).Minutes}};
    return Enumerable.Range(1, (int)(end -  start).TotalHours)
                    .Select(v => start.AddHours(v))
                    .GroupBy(v => v.Date)
                    .ToDictionary( v => v.Key, r => r.Count());
}

打印：

Week 9 has 10 hours
Week 10 has 65 hours
Week 11 has 44 hours

如何使用Entity Framework每周获取聚合数据

2 个答案: