Elasticsearch 聚合

聚合框架收集搜索查询选择的所有数据，并由许多构建块组成，有助于构建复杂的数据摘要。 聚合的基本结构如下所示

"aggregations" : {
   "" : {
      "" : {

      }
 
      [,"meta" : { [] } ]?
      [,"aggregations" : { []+ } ]?
   }
   [,"" : { ... } ]*
}

有不同类型的聚合，每种都有自己的目的。 本章将详细讨论它们。

Metrics 聚合函数

这些聚合有助于根据聚合文档的字段值计算矩阵，有时可以从脚本中生成一些值。

数值矩阵要么像平均聚合那样是单值的，要么像统计数据那样是多值的。

Avg 聚合

此聚合用于获取聚合文档中存在的任何数字字段的平均值。 例如，

POST /schools/_search
{
   "aggs":{
      "avg_fees":{"avg":{"field":"fees"}}
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 41,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : 1.0,
      "hits" : [
         {
            "_index" : "schools",
            "_type" : "school",
            "_id" : "5",
            "_score" : 1.0,
            "_source" : {
               "name" : "Central School",
               "description" : "CBSE Affiliation",
               "street" : "Nagan",
               "city" : "paprola",
               "state" : "HP",
               "zip" : "176115",
               "location" : [
                  31.8955385,
                  76.8380405
               ],
            "fees" : 2200,
            "tags" : [
               "Senior Secondary",
               "beautiful campus"
            ],
            "rating" : "3.3"
         }
      },
      {
         "_index" : "schools",
         "_type" : "school",
         "_id" : "4",
         "_score" : 1.0,
         "_source" : {
            "name" : "City Best School",
            "description" : "ICSE",
            "street" : "West End",
            "city" : "Meerut",
            "state" : "UP",
            "zip" : "250002",
            "location" : [
               28.9926174,
               77.692485
            ],
            "fees" : 3500,
            "tags" : [
               "fully computerized"
            ],
            "rating" : "4.5"
         }
      }
   ]
 },
   "aggregations" : {
      "avg_fees" : {
         "value" : 2850.0
      }
   }
}

Cardinality 聚合

此聚合给出特定字段的不同值的计数。

POST /schools/_search?size=0
{
   "aggs":{
      "distinct_name_count":{"cardinality":{"field":"fees"}}
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 2,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "distinct_name_count" : {
         "value" : 2
      }
   }
}

注意 - 基数的值为 2，因为费用中有两个不同的值。

扩展统计聚合

此聚合生成有关聚合文档中特定数字字段的所有统计信息。

POST /schools/_search?size=0
{
   "aggs" : {
      "fees_stats" : { "extended_stats" : { "field" : "fees" } }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 8,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "fees_stats" : {
         "count" : 2,
         "min" : 2200.0,
         "max" : 3500.0,
         "avg" : 2850.0,
         "sum" : 5700.0,
         "sum_of_squares" : 1.709E7,
         "variance" : 422500.0,
         "std_deviation" : 650.0,
         "std_deviation_bounds" : {
            "upper" : 4150.0,
            "lower" : 1550.0
         }
      }
   }
}

Max 聚合

此聚合查找聚合文档中特定数字字段的最大值。

POST /schools/_search?size=0
{
   "aggs" : {
   "max_fees" : { "max" : { "field" : "fees" } }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 16,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
  "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "max_fees" : {
         "value" : 3500.0
      }
   }
}

Min 聚合

此聚合查找聚合文档中特定数字字段的最小值。

POST /schools/_search?size=0
{
   "aggs" : {
      "min_fees" : { "min" : { "field" : "fees" } }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 2,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
  "aggregations" : {
      "min_fees" : {
         "value" : 2200.0
      }
   }
}

Sum 聚合

此聚合计算聚合文档中特定数字字段的总和。

POST /schools/_search?size=0
{
   "aggs" : {
      "total_fees" : { "sum" : { "field" : "fees" } }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 8,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "total_fees" : {
         "value" : 5700.0
      }
   }
}

还有一些其他指标聚合用于特殊情况，例如地理边界聚合和地理质心聚合，用于地理位置定位。

Stats 聚合

一种多值指标聚合，用于计算从聚合文档中提取的数值的统计信息。

POST /schools/_search?size=0
{
   "aggs" : {
      "grades_stats" : { "stats" : { "field" : "fees" } }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 2,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "grades_stats" : {
         "count" : 2,
         "min" : 2200.0,
         "max" : 3500.0,
         "avg" : 2850.0,
         "sum" : 5700.0
      }
   }
}

聚合元数据

我们可以在请求时使用元标记添加一些有关聚合的数据，并可以在响应中获取这些数据。

POST /schools/_search?size=0
{
   "aggs" : {
      "avg_fees" : { "avg" : { "field" : "fees" } ,
         "meta" :{
            "dsc" :"Lowest Fees This Year"
         }
      }
   }
}

运行上面的代码，我们得到以下结果

{
   "took" : 0,
   "timed_out" : false,
   "_shards" : {
      "total" : 1,
      "successful" : 1,
      "skipped" : 0,
      "failed" : 0
   },
   "hits" : {
      "total" : {
         "value" : 2,
         "relation" : "eq"
      },
      "max_score" : null,
      "hits" : [ ]
   },
   "aggregations" : {
      "avg_fees" : {
         "meta" : {
            "dsc" : "Lowest Fees This Year"
         },
         "value" : 2850.0
      }
   }
}