elasticsearch - ElasticSearch-使用路径标记器按父分组

Question

我有以下文件:

PUT /my_index/topic/1
{
"path" : "fruits"
}

PUT /my_index/topic/2
{
"path" : "fruits/apple"
}

PUT /my_index/topic/3
{
"path" : "fruits/pear"
}

PUT /my_index/topic/4
{
"path" : "vegetables"
}

PUT /my_index/topic/5
{
"path" : "vegetables/carrot"
}

PUT /my_index/topic/6
{
"path" : "vegetables/broccoli"
}

我正在努力弄清楚如何汇总这些文档，以便获得以下输出:

水果 {
苹果，
pear
}

蔬菜{
胡萝卜，
西兰花
}

Answer 1

我发现做到这一点的一种方法是使用 path_hierarchy tokenizer和 token_count 字段。首先，我们像这样创建my_index:

curl -XPUT localhost:9200/my_index -d '{
  "settings": {
    "analysis": {
      "analyzer": {
        "path-analyzer": {
          "type": "custom",
          "tokenizer": "path-tokenizer"
        }
      },
      "tokenizer": {
        "path-tokenizer": {
          "type": "path_hierarchy"
        }
      }
    }
  },
  "mappings": {
    "topic": {
      "properties": {
        "path": {
          "type": "string",
          "index_analyzer": "path-analyzer",
          "fields": {
            "tokens": {
              "type": "token_count",
              "store": "yes",
              "analyzer": "standard"
            }
          }
        }
      }
    }
  }
}'

然后，我们使用与您的问题相同的PUT查询为您的文档建立索引。

最后，可以输出如下期望的搜索查询:

curl -XPOST localhost:9200/my_index/topic/_search?pretty -d '{
  "size": 0,
  "aggs": {
    "first_level": {
      "terms": {
        "field": "path",
        "exclude": ".*/.*"
      },
      "aggs": {
        "second_level": {
          "filter": {
            "term": {
              "path.tokens": 2
            }
          },
          "aggs": {
            "type": {
              "terms": {
                "field": "path",
                "include": ".*/.*"
              }
            }
          }
        }
      }
    }
  }
}'

并输出:

{
  ...
  "aggregations" : {
    "first_level" : {
      "doc_count_error_upper_bound" : 0,
      "sum_other_doc_count" : 0,
      "buckets" : [ {
        "key" : "fruits",
        "doc_count" : 3,
        "second_level" : {
          "doc_count" : 2,
          "type" : {
            "doc_count_error_upper_bound" : 0,
            "sum_other_doc_count" : 0,
            "buckets" : [ {
              "key" : "fruits/apple",
              "doc_count" : 1
            }, {
              "key" : "fruits/pear",
              "doc_count" : 1
            } ]
          }
        }
      }, {
        "key" : "vegetables",
        "doc_count" : 3,
        "second_level" : {
          "doc_count" : 2,
          "type" : {
            "doc_count_error_upper_bound" : 0,
            "sum_other_doc_count" : 0,
            "buckets" : [ {
              "key" : "vegetables/broccoli",
              "doc_count" : 1
            }, {
              "key" : "vegetables/carrot",
              "doc_count" : 1
            } ]
          }
        }
      } ]
    }
  }
}