elasticsearch查询多字段聚合

版权声明:本文为博主原创文章,遵循 CC 4.0 by-sa 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://blog.csdn.net/lln_avaj/article/details/59118980

1,需求:

在分页查询的同时,返回所查询商品的类目和品牌的聚合信息

2,mapping:
    {
        "item" : {
            "properties" : {
                "item_name" : {
                    "type" : "string",
                                        "index" : "analyzed",
                                        "analyzer" : "ik_smart"
                },
               "brand_name" : {
                    "type" : "multi_field",
                    "fields" : {
                        "brand_name" : {"type" : "string", "index" : "analyzed", "analyzer" :"ik_smart"},
                        "brand_name_agg" : {"type" : "string", "index" : "not_analyzed"}
                    }
                },
                "c_name" : {
                    "type" : "multi_field",
                    "fields" : {
                        "c_name" : {"type" : "string", "index" : "analyzed","analyzer" :"ik_smart"},
                        "c_name_agg" : {"type" : "string", "index" : "not_analyzed"}
                    }
                }
            }
        }
    }

这里需要注意的是brand_name如果设置为ik分词,聚合结果也会进行分词,导致结果不准确

可对brand_name进行multi_field设置,query时使用brand_name,agg时使用brand_name_agg

 

3,查询时需要注意如果分别统计两个不相关的字段,可设置aggregations下多个结果集,若统计有父子关系的字段,

例如某年级下某班级的学生个数,可使用subAggregation功能

search:

 {"from" : 0,
  "size" : 20,
  "query" : {
    "bool" : {
      "must" : {
        "fuzzy" : {
          "item_name" : {
            "value" : "西门子"
          }
        }
      }
    }
  },
  "explain" : true,
  "aggregations" : {
    "brand_name.brand_name_agg" : {
      "terms" : {
        "field" : "brand_name.brand_name_agg",
        "size" : 9999,
        "order" : {
          "_count" : "asc"
        }
      }
    },
    "c_name.c_name_agg" : {
      "terms" : {
        "field" : "c_name.c_name_agg",
        "size" : 9999,
        "order" : {
          "_count" : "asc"
        }
      }
    }
  }
}

 

4,相关java实现方式:

1)查询条件构建:

 

SearchRequestBuilder builder = client.prepareSearch(index).setTypes(type);
BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();
TermsBuilder termsBuilder = AggregationBuilders.terms(aggregation.getField());
termsBuilder.field(aggregation.getField());
if(SortConstant.SortASC.equals(aggregation.getOrder())){
    termsBuilder.order(Terms.Order.aggregation("_count", true));
}else{
    termsBuilder.order(Terms.Order.aggregation("_count", false));
}
termsBuilder.size(aggregation.getSize());
builder.addAggregation(termsBuilder);

2)查询结果返回:
Map<String,List<Object>> aggrMap = new HashMap<String, List<Object>>(); List<Aggregation> aggregationList = searchCondition.getAggregationList(); if(aggregationList!=null && aggregationList.size() > 0){ for(Aggregation aggregation : aggregationList){ List<Object> results = new ArrayList<Object>(); Terms result = searchResponse.getAggregations().get(aggregation.getField()); for (Terms.Bucket entry : result.getBuckets()) { results.add(entry.getKey()); } aggrMap.put(aggregation.getField(),results); } }


5,参考资料:
http://www.cnblogs.com/xing901022/p/4947436.html?utm_source=tuicool&utm_medium=referral
http://blog.csdn.net/xiaohelong2005/article/details/41513985
http://www.cnblogs.com/sha0830/p/5549331.html

 

 

 

 

 

 

展开阅读全文

没有更多推荐了,返回首页