如何在Elasticsearch索引中获取指定字段的所有唯一对象？

Hey there! Let's walk through how to properly fetch all unique values (or objects) for your tags field in Elasticsearch, building on the request you've already started with.

先说说你当前请求的小局限 #

You're already using the terms aggregation, which is the right tool for this job—but by default, Elasticsearch only returns the top 10 most frequent unique values for a terms aggregation. That means if you have more than 10 unique tags, you'll miss some results. Also, setting size: 1 only controls how many hit documents are returned, not how many aggregation buckets you get.

改进方案：获取所有唯一值（普通字段） #

If tags is a regular field (not a nested object), adjust your request to set a large enough size in the terms aggregation (bigger than the total number of unique tags you expect) and set size: 0 to skip returning hit documents (saves resources):

POST http://es5server:9200/indexname/_search
Content-Type: application/json

{
  "size": 0,
  "aggs": {
    "unique_tags": {
      "terms": {
        "field": "tags",
        "size": 10000  // 设一个比你预期的唯一标签总数更大的值
      }
    }
  }
}

如果tags是嵌套对象的情况 #

You mentioned "unique objects"—if tags is a nested object type, you'll need to use a nested aggregation first to access the nested data, then use terms on a unique identifier field inside the object (like id) to group unique objects:

POST http://es5server:9200/indexname/_search
Content-Type: application/json

{
  "size": 0,
  "aggs": {
    "nested_tags_context": {
      "nested": {
        "path": "tags"
      },
      "aggs": {
        "unique_tag_objects": {
          "terms": {
            "field": "tags.id"  // 使用标签对象中的唯一标识字段
          }
        }
      }
    }
  }
}

解读返回结果 #

你的唯一值/对象会出现在响应的aggregations部分：

• 普通字段场景：查看aggregations.unique_tags.buckets——每个bucket的key是一个唯一标签，doc_count是它的出现次数。
• 嵌套对象场景：查看aggregations.nested_tags_context.unique_tag_objects.buckets——每个key对应一个标签对象的唯一标识。

内容的提问来源于stack exchange，提问作者Marcus Haase