Elasticsearch 对嵌套字段的复合聚合查询

Question

I've a question on an aggregation on nested objects. Document is like:

{
    "features": [{
            "key": "key1",
            "values": ["A", "B"]
        },
        {
            "key": "key2",
            "values": ["C", "D"]
        },
        {
            "key": "key2",
            "values": ["E"]
        }
    ]
}

where 'features' is a nested object. I can aggregate and get distinct values from key and values, but I need to get a combined bucket aggregation, where I need:

key1 -> A,B key2 -> C,D,E

Is composite aggregation that has to be used? Or which is the proper aggregation to use? Java samples are also welcome!

Thanks!!!

Answer 1

You don't really need composite for this. The following should be fine:

{
  "size": 0,
  "aggs": {
    "nested_aggs": {
      "nested": {
        "path": "features"
      },
      "aggs": {
        "by_key": {
          "terms": {
            "field": "features.key.keyword"
          },
          "aggs": {
            "by_values": {
              "terms": {
                "field": "features.values.keyword"
              }
            }
          }
        }
      }
    }
  }
}

assuming your mapping looks like this

{
  "mappings":{
    "properties":{
      "features":{
        "type":"nested",
        "properties":{
          "key":{
            "type":"text",
            "fields":{
              "keyword":{
                "type":"keyword",
                "ignore_above":256
              }
            }
          },
          "values":{
            "type":"text",
            "fields":{
              "keyword":{
                "type":"keyword",
                "ignore_above":256
              }
            }
          }
        }
      }
    }
  }
}