Elasticsearch-如何过滤嵌套的聚合桶？ [英] Elasticsearch - How to filter nested aggregation bucket?

问题描述

我正在尝试过滤存储桶中的嵌套聚合。

I'm trying to filter buckets for nested aggregations.

映射：

{
  "dev.directory.3" : {
    "mappings" : {
      "profile" : {
        "properties" : {
          "events" : {
            "type" : "nested",
            "properties" : {
              "id" : {
                "type" : "integer"
              },
              "name" : {
                "type" : "string",
                "index" : "not_analyzed"
              },
            }
          },
          "title" : {
            "type" : "string"
          }
        }
      }
    }
  }
}

索引数据：

"hits" : {
    "total" : 1,
    "max_score" : 1.0,
    "hits" : [ {
      "_index" : "dev.directory.3",
      "_type" : "profile",
      "_id" : "1",
      "_score" : 1.0,
      "_source" : {
        "title" : "Project manager",
        "events" : [ 
          {
            "id" : 1,
            "name" : "Event A",
          }, 
          {
            "id" : 2,
            "name" : "Event B",
          },
          {
            "id" : 3,
            "name" : "Event C",
          },
          {
            "id" : 4,
            "name" : "Event D",
          } 
        ],
      }
    }
  ]
}

我正在使用此查询和聚合定义

I'm using this query and aggregation definition

{
    "query": {
        "nested": {
            "path": "events",
            "query": {
                "bool": {
                    "filter": [{
                        "terms": {
                            "events.id": [1, 2]
                        }
                    }]
                }
            },
            "inner_hits": {}
        }
    },
    "aggs": {
        "events.name12": {
            "filter": {},
            "aggs": {
                "inner": {
                    "nested": {
                        "path": "events"
                    },
                    "aggs": {
                        "events.name": {
                            "terms": {
                                "field": "events.name"
                            }
                        },
                        "events.name_count": {
                            "cardinality": {
                                "field": "events.name"
                            }
                        }
                    }
                }
            }
        }
    },
    "size": 20,
    "_source": ["email", "company_name", "events"]
}

我从聚合结果中得到的是：

What I am getting is from aggregation result is:

"aggregations": {
    "events.name12": {
      "doc_count": 2,
      "filtered": {
        "doc_count": 4,
        "events.name": {
          "buckets": [
            {
              "key": "Event A",
              "doc_count": 1
            },
            {
              "key": "Event B",
              "doc_count": 1
            },
            {
              "key": "Event C",
              "doc_count": 1
            },
            {
              "key": "Event D",
              "doc_count": 1
            }
          ]
        },
        "events.name_count": {
          "value": 4
        }
      }
    }

我在过滤存储桶列表时遇到的困难仅在于提供的事件ID，因此结果应该像：

I struggling with filtering buckets list only to provided event ids, so the result should be like:

"aggregations": {
    "events.name12": {
      "doc_count": 2,
      "filtered": {
        "doc_count": 2,
        "events.name": {
          "buckets": [
            {
              "key": "Event A",
              "doc_count": 1
            },
            {
              "key": "Event B",
              "doc_count": 1
            }
          ]
        },
        "events.name_count": {
          "value": 2
        }
      }
    }

推荐答案

您已经快到了，您只需要还要在聚合中的 events.id 上添加过滤器，如下所示：

You're almost there, you simply need to add the filter on events.id in your aggregation as well, like this:

{
  "query": {
    "nested": {
      "path": "events",
      "query": {
        "bool": {
          "filter": [
            {
              "terms": {
                "events.id": [
                  1,
                  2
                ]
              }
            }
          ]
        }
      },
      "inner_hits": {}
    }
  },
  "aggs": {
    "events.name12": {
      "nested": {
        "path": "events"
      },
      "aggs": {
        "inner": {
          "filter": {
            "terms": {
              "events.id": [
                1,
                2
              ]
            }
          },
          "aggs": {
            "events.name": {
              "terms": {
                "field": "events.name"
              }
            },
            "events.name_count": {
              "cardinality": {
                "field": "events.name"
              }
            }
          }
        }
      }
    }
  },
  "size": 20,
  "_source": [
    "email",
    "company_name",
    "events"
  ]
}

原因是您的查询将正确选择所有包含具有指定事件ID的嵌套事件的文档，但是，您的聚合将对所有选定的所有嵌套事件起作用文件。因此，您需要从聚合中也没有正确ID的那些文档中过滤掉所有嵌套事件。

The reason is that your query will correctly select all documents for which there are nested events with the specified event IDs, however, your aggregation will then work on all nested events from all selected documents. So you need to filter out all nested events from those documents that don't have the right IDs in the aggregation as well.

这篇关于Elasticsearch-如何过滤嵌套的聚合桶？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！