Spring data Elasticsearch to convert Elasticsearch aggregation query into code

Question

{
  "query": {
    "bool": {
      "must": [
        {
          "term": {
            "id": {
              "value": 1111
            }
          }
        }
      ]
    }
  },
  "aggs": {
    "sequence": {
      "terms": {
        "field": "sequence.keyword"
      },
      "aggregations": {
        "amount": {
          "sum": {
            "field": "amount"
          }
        }
      }
    },
    "totalAmount": {
      "sum_bucket": {
        "buckets_path": "sequence>amount"
      }
    }
  }
}

Spring Data ElasticSearch 3.2

final SearchQuery searchQuery =
        new NativeSearchQueryBuilder()
            .withQuery(QueryHelper.boolQueryBuilder(id))
            .addAggregation(
                terms("sequence")
                    .field("sequence.keyword")
                    .subAggregations(subAggregations()))
            .addAggregation(sumBucket("sum_bucket", "sequence>amount"))
            .build();

Now, since addAggregation method of NativeSearchQueryBuilder accepts AbstractAggregationBuilder and sumBucket is of type PipelineAggregationBuilder, not able to pass this value to addAggregation(sumBucket("sum_bucket", "sequence>amount")) due to compilation error, Any idea how to use sumBucket on terms aggregation and write in java?

Answer 1

I just checked the code of Spring Data Elasticsearch and Elasticsearch itself. Using PipelineAggregationBuilder derived aggregations currently is not possible with Spring Data Elasticsearch.

I don't know, why Elasticsearch has two different class hierarchies for aggregations, but Spring Data Elasticsearch currently can only handle AggregationBuilder and not PipelineAggregationBuilder derived classes.

I created an issue in Jira for this.