ElasticSearch calculate percentage for each bucket from total

Question

I'm using ElasticSearch v5. I'm trying to do something similar described in Elasticsearch analytics percent where I have a terms aggregation and I want to calculate a percentage which is a value from each bucket over the total of all buckets. This is my request:

{
  "query": {
    "match_all": {}
  },
  "aggs": {
    "periods": {
      "terms": { 
        "field": "periods",
        "size": 3
      },
      "aggs": {
        "balance": {
          "sum": {
            "field": "balance"
          }
        }
      }
    },
    "total_balance": {
        "sum_bucket": {
            "buckets_path": "periods>balance" 
        }
    }
  }

}

The result I get back this like this:

{
  "aggregations": {
    "periods": {
      "doc_count_error_upper_bound": 0,
      "sum_other_doc_count": 1018940846,
      "buckets": [
        {
          "key": 1177977600000,
          "doc_count": 11615418,
          "balance": {
            "value": 2492032741768.1616
       }
       },
       {
          "key": 1185926400000,
          "doc_count": 11592425,
          "balance": {
            "value": 2575365325406.6533
      }
      },
      {
          "key": 1175385600000,
          "doc_count": 11477402,
          "balance": {
            "value": 2456256695380.8306
          }
        }
      ]
    },
    "total_balance": {
      "value": 7523654762555.645
    }
  }
}

How do I calculate "balance"/"total_balance" for each item in the bucket from ElasticSearch? I tried bucket script aggregation at the bucket (periods) level, but I cannot set my buckets_path to total_balance. This post https://discuss.elastic.co/t/combining-two-aggregations-to-get-term-percentage/22201 talks about using Significant Terms Aggregation, but I need calculation of using specific fields, not doc_count. I know I can do this as a simple calculation on the client side, but I would like to do this all together in ElasticSearch if possible.

Answer 1

No, you can't do that. By the time I'm writing this post, we're in version 6.1.

According to https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-pipeline.html#buckets-path-syntax, there's only two major types of aggregations pipelines: parent and siblings.

So, in order to reference the total_balance aggregation from within the periods buckets, we should be able to reference an "uncle" aggregation from the buckets_path attribute, which is not possible.