Sorting 按与文档相同的属性对Elasticsearch Bucket进行排序

有一种方法可以按多个属性对文档进行排序。 我想按相同的属性以相同的方式对聚合的bucket进行排序。这可能吗

以下是我的示例文档：

    curl -XPOST 'localhost:9200/_bulk?pretty' -H 'Content-Type: 
application/json' -d'
    { "index" : { "_index" : "cars", "_type" : "cars", "_id" : "1" } }
    { "make":"BMW", "series":"3er", "doorCount": 4}
    { "index" : { "_index" : "cars", "_type" : "cars", "_id" : "2" } }
    { "make":"BMW", "series":"3er", "doorCount": 5}
    { "index" : { "_index" : "cars", "_type" : "cars", "_id" : "3" } }
    { "make":"Opel", "series":"Astra", "doorCount": 2}
    { "index" : { "_index" : "cars", "_type" : "cars", "_id" : "4" } }
    { "make":"Opel", "series":"Omega", "doorCount": 2}
    '

在这里，我将fielddata设置为true，以便能够聚合：

curl -XPUT 'localhost:9200/cars/_mapping/cars?pretty' -H 'Content-Type: application/json' -d'
{
  "properties": {
    "make": { 
      "type":     "text",
      "fielddata": true
    },
    "series": { 
      "type":     "text",
      "fielddata": true
    }
  }
}
'

以下是查询聚合存储桶：

curl -XGET 'localhost:9200/cars/cars/_search?pretty' -d'
{
  "size": 0,
  "aggregations" : {
    "CARS" : {
      "terms" : {
        "script" : {
          "inline" : "doc[\"make\"].value + \"_\" + doc[\"series\"].value",
          "lang" : "painless"
        }      
      },
      "aggregations" : {
        "make" : {
          "terms" : { "field" : "make"}
        },
        "series" : {
          "terms" : { "field" : "series" }
        }
      }
    }
  }
}
'

---

你知道如何按make和series对bucket进行排序吗？

首先，你可能想指定

make

和

series

为类型，而不是

text

，因为它不需要分析（你需要这些字段上的精确匹配）

您的输出是什么样子的？当我尝试它时，它似乎被正确地分类了

  [
    {
      "key": "bmw_3er",
      "doc_count": 2,
      "series": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "3er",
            "doc_count": 2
          }
        ]
      },
      "make": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "bmw",
            "doc_count": 2
          }
        ]
      }
    },
    {
      "key": "opel_astra",
      "doc_count": 1,
      "series": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "astra",
            "doc_count": 1
          }
        ]
      },
      "make": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "opel",
            "doc_count": 1
          }
        ]
      }
    },
    {
      "key": "opel_omega",
      "doc_count": 1,
      "series": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "omega",
            "doc_count": 1
          }
        ]
      },
      "make": {
        "doc_count_error_upper_bound": 0,
        "sum_other_doc_count": 0,
        "buckets": [
          {
            "key": "opel",
            "doc_count": 1
          }
        ]
      }
    }
  ]

---

我想在我的查询中添加一种

sort

参数，比如在对普通文档进行排序时：

“sort”：[{“series”：“desc”}，{“make”：“asc”}]

，这样订单桶将首先按序列排序，然后按make排序