ElasticSearch - 使用数组字段的子项聚合问题 [英] ElasticSearch - issue with sub term aggregation with array fields

问题描述

我有两个以下文件：

{  
"title":"The Avengers",
"year":2012,
"casting":[  
    {  
    "name":"Robert Downey Jr.",
    "category":"Actor",
    },
    {  
    "name":"Chris Evans",
    "category":"Actor",
    }
]
}

和：

{  
"title":"The Judge",
"year":2014,
"casting":[  
    {  
    "name":"Robert Downey Jr.",
    "category":"Producer",
    },
    {  
    "name":"Robert Duvall",
    "category":"Actor",
    }
]
}

我想执行聚合，基于两个字段：cast.name和cast.category。

I would like to perform aggregations, based on two fields : casting.name and casting.category.

我尝试使用基于cast.name字段的TermsAggregation，具有子集合，这是基于cast.category字段的另一个TermsAggregation。

I tried with a TermsAggregation based on casting.name field, with a subaggregation, which is another TermsAggregation based on the casting.category field.

问题是，对于Chris Evans条目，ElasticSearch设置所有类别的桶（Actor，Producer），而应该只设置1个桶（Actor）。

The problem is that for the "Chris Evans" entry, ElasticSearch set buckets for ALL categories (Actor, Producer) whereas it should set only 1 bucket (Actor).

似乎有一个笛卡儿所有的cast.category发生之间的产物和所有的cast.name发生。
它的行为类似于数组字段（cast），而我没有简单的字段（作为标题或年份）的问题。

It seems that there is a cartesian product between all casting.category occurences and all casting.name occurences. It behaves like this with array fields (casting), whereas I don't have the problem with simple fields (as title, or year).

我也尝试使用嵌套的聚合，但可能不正确，ElasticSearch会抛出一个错误，告诉cast.res不是嵌套字段。

I also tried to use nested aggregations, but maybe not properly, and ElasticSearch throws an error telling that casting.category is not a nested field.

这里有什么想法？ >

Any idea here?

推荐答案

弹性搜索将展开嵌套对象，因此内部您将得到：

Elasticsearch will flatten the nested objects, so internally you will get:

{  
"title":"The Judge",
"year":2014,
"casting.name": ["Robert Downey Jr.","Robert Duvall"],
"casting.category": ["Producer", "Actor"]
}

如果你想保持关系，你需要使用嵌套对象或 parent c hild关系

if you want to keep the relationship you'll need to use either nested objects or a parent child relationship

要进行嵌套映射，您需要执行以下操作：

To do a nested mapping you'd need to do something like this:

  "mappings": {
    "movies": {
      "properties": {
        "title" : { "type": "string" },
        "year" : { "type": "integer" },
        "casting": {
          "type": "nested", 
          "properties": {
            "name":    { "type": "string" },
            "category": { "type": "string" }
          }
        }
      }
    }
  }

这篇关于ElasticSearch - 使用数组字段的子项聚合问题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！