elasticsearch中的桶脚本java api

Question

我做了 elasticsearch 查询，得到每个 oid 名称，它的计数除以 1000。（数字（1000）可以更改）

    "aggregations" : {
        "agg_oid" : {
          "terms" : {
            "field" : "oid",
            "order" : {
              "_count" : "desc"
            }
          },
          "aggregations" : {
              "agg_values" : {
                  "bucket_script": {
                        "buckets_path": {
                          "count": "_count"
                        },
                        "script": "count / 1000"
                    }
              }
          }
        }
      }

效果很好，并尝试用java api（目前使用jest）。

AggregationBuilders
    .terms( "agg_oid")
    .field( "oid")
    .subAggregation(
        AggregationBuilders
             // bucket script and path for count
    );

如何在java中实现'bucket_script'？如果没有，有没有办法通过java api中的计算来查询聚合的计数？

Answer 1

应该这样做：

Map<String, String> bucketsPathsMap = new HashMap<>();
bucketsPathsMap.put("count", "_count");
Script script = new Script("count / 1000");

BucketScriptPipelineAggregationBuilder bs = PipelineAggregatorBuilders
    .bucketScript("agg_values", bucketsPathsMap, script);

TermsAggregationBuilder oid = AggregationBuilders.terms("agg_oid")
    .field("oid")
    .order(InternalOrder.COUNT_DESC).
    subAggregation(bs);