ElasticSearch 6 中带有过滤器的嵌套对象聚合

Question

我有一组文档表示 ElasticSearsh 6 中的属性个单位。每个属性都有嵌套的每周费率数组：

{
   "name" : "Completely Awesome Cabin"
   "rates" : [
      {
         "start": "2018-06-09T00:00:00",
         "end": "2018-06-16T00:00:00",
         "weeklyRate": 100.0,
      },
      {
         "start": "2018-06-16T00:00:00",
         "end": "2018-06-23T00:00:00",
         "weeklyRate": 200.0,
      }
      ...
   ]    
   ...
}

我正在按包括日期在内的几个选项执行一些过滤。我需要添加聚合，以便在所有通过过滤器的单元之间提供最小和最大周率。我想它应该是某种带有过滤器的嵌套聚合。我怎样才能做到这一点？

Answer 1

这是一个完整的例子运行 NEST 6.1.0

private static void Main()
{
    var index = "default";
    var pool = new SingleNodeConnectionPool(new Uri("http://localhost:9200"));
    var connectionSettings = new ConnectionSettings(pool)
        .DefaultIndex(index);

    var client = new ElasticClient(connectionSettings);

    if (client.IndexExists(index).Exists)
        client.DeleteIndex(index);

    client.CreateIndex(index, c => c
        .Mappings(m => m
            .Map<MyDocument>(mm => mm
                .AutoMap()
                .Properties(p => p
                    .Nested<Rate>(n => n
                        .AutoMap()
                        .Name(nn => nn.Rates)
                    )
                )
            )
        )
    );

    client.Bulk(b => b
        .IndexMany(new[] {
            new MyDocument
            {
                Name = "doc 1",
                Rates = new []
                {
                    new Rate
                    {
                        Start = new DateTime(2018, 6, 9),
                        End = new DateTime(2018, 6, 16),
                        WeeklyRate = 100
                    },
                    new Rate
                    {
                        Start = new DateTime(2018, 6, 16),
                        End = new DateTime(2018, 6, 23),
                        WeeklyRate = 200
                    }
                }
            },
            new MyDocument
            {
                Name = "doc 2",
                Rates = new []
                {
                    new Rate
                    {
                        Start = new DateTime(2018, 6, 9),
                        End = new DateTime(2018, 6, 16),
                        WeeklyRate = 120
                    },
                    new Rate
                    {
                        Start = new DateTime(2018, 6, 16),
                        End = new DateTime(2018, 6, 23),
                        WeeklyRate = 250
                    }
                }
            }
        })
        .Refresh(Refresh.WaitFor)
    );

    var searchResponse = client.Search<MyDocument>(s => s
        // apply your filtering in .Query(...) e.g. applicable date range
        .Query(q => q.MatchAll())
        // don't return documents, just calculate aggregations
        .Size(0)
        .Aggregations(a => a
            .Nested("nested_start_dates", n => n
                .Path(f => f.Rates)
                .Aggregations(aa => aa
                    .DateHistogram("start_dates", dh => dh
                        .Field(f => f.Rates.First().Start)
                        .Interval(DateInterval.Day)
                        .MinimumDocumentCount(1)
                        .Aggregations(aaa => aaa
                            .Min("min_rate", m => m
                                .Field(f => f.Rates.First().WeeklyRate)
                            )
                            .Max("max_rate", m => m
                                .Field(f => f.Rates.First().WeeklyRate)
                            )
                        )
                    )
                )
            )
        )
    );

    var nested = searchResponse.Aggregations.Nested("nested_start_dates");

    var startBuckets = nested.DateHistogram("start_dates").Buckets;

    foreach(var start in startBuckets)
    {
        var min = start.Min("min_rate").Value;
        var max = start.Max("max_rate").Value;

        Console.WriteLine($"{start.KeyAsString} - min: {min}, max: {max}");
    }
}

public class MyDocument
{
    public string Name {get;set;}

    public IEnumerable<Rate> Rates {get;set;}
}

public class Rate
{
    public DateTime Start {get;set;}

    public DateTime End {get;set;}

    public double WeeklyRate {get;set;}
}

将以下内容打印到控制台

2018-06-09T00:00:00.000Z - min: 100, max: 120
2018-06-16T00:00:00.000Z - min: 200, max: 250

您可能还对其他指标聚合感兴趣，例如 Stats Agggregation

Answer 2

除了 Russ Cam 提供的非常有用和有帮助的答案之外，我还想 post 最终实现，它正是我所需要的。这是 NEST 聚合：

.Aggregations(a => a
    .Nested("budget_bound", n => n
        .Path(p => p.Rates)
        .Aggregations(aa => aa
            .Filter("by_start_date", fl => fl
                .Filter(fld => fld 
                    .DateRange(dr => dr
                        .Field(f => f.Rates.First().Start)
                        .GreaterThanOrEquals(checkIn)
                        .LessThanOrEquals(checkOut)))
                     .Aggregations(md => md
                         .Min("min_budget", m => m
                             .Field(f => f.Rates.First().WeeklyRate))
                         .Max("max_budget", m => m
                             .Field(f => f.Rates.First().WeeklyRate))
                      )
                 )
            )
       )

这里是对应的ES查询：

"aggs": {
"budget_bound": {
  "nested": {
    "path": "rates"
  },
  "aggs": {
    "by_start_date": {
      "filter": {
        "range": {
          "rates.start": {
            "gte": "2018-06-29T00:00:00+07:00", // parameter values
            "lte": "2018-07-06T00:00:00+07:00"  // parameter values
          }
        }
      },
      "aggs": {
        "min_budget": {
          "min": {
            "field": "rates.weeklyRate"
          }
        },
        "max_budget": {
          "max": {
            "field": "rates.weeklyRate"
          }
        }
      }
    }
  }
}}

对我来说，事情是弄清楚如何嵌套聚合以在获得最小和最大聚合之前添加嵌套集合的过滤。

ElasticSearch 6 中带有过滤器的嵌套对象聚合

Aggregation by nested objects with filter in ElasticSearch 6

elasticsearch

nest

elasticsearch-aggregation