标签:val 无法 转换 统计 聚合函数 timeout bag 个人 imp
public class ElasticSearchBuilderQuery {
/**
- 构建基本查询 - 搜索关键字 (数据分词。按逗号)
- @param searchMap
-
@return
*/
public static BoolQueryBuilder buildBasicQuery(PageData searchMap) {
// 构建布尔查询
BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();
// 时间区间查询
if(searchMap.get("startTime")!= null && searchMap.get("endTime") !=null){
boolQueryBuilder.filter(QueryBuilders.rangeQuery("date").gte(searchMap.get("startTime")).lte(searchMap.get("endTime")).timeZone("Asia/Shanghai")); //timeZone这个参数很有必要,由于ES不管设置什么,都会转换为UTC格式,很搞
}// 项目过滤 if (StringUtils.isNotBlank(searchMap.get("project")+"") && searchMap.containsKey("project")){ BoolQueryBuilder boolQueryBuilderProject = QueryBuilders.boolQuery(); String[] projects = (String[]) searchMap.get("project"); for(String project : projects){ QueryBuilder queryBuilder = QueryBuilders.matchPhraseQuery("project_ids",project); boolQueryBuilderProject.should().add(queryBuilder); } boolQueryBuilder.filter(boolQueryBuilderProject); } // 供应商过滤 if(StringUtils.isNotBlank(searchMap.get("company")+"") && searchMap.containsKey("company")){ BoolQueryBuilder boolQueryBuilderCompany = QueryBuilders.boolQuery(); String[] companys = (String[]) searchMap.get("company"); for(String company : companys){ QueryBuilder queryBuilder = QueryBuilders.matchPhraseQuery("company_ids",company); boolQueryBuilderCompany.should().add(queryBuilder); } boolQueryBuilder.filter(boolQueryBuilderCompany); } // 部门过滤 if(StringUtils.isNotBlank(searchMap.get("dept")+"") && searchMap.containsKey("dept")){ BoolQueryBuilder boolQueryBuilderDept = QueryBuilders.boolQuery(); String[] depts = (String[]) searchMap.get("dept"); for(String dept : depts){ QueryBuilder queryBuilder = QueryBuilders.matchPhraseQuery("dept_ids",dept); boolQueryBuilderDept.should().add(queryBuilder); } boolQueryBuilder.filter(boolQueryBuilderDept); } // 角色过滤 if(StringUtils.isNotBlank(searchMap.get("role")+"") && searchMap.containsKey("role")){ BoolQueryBuilder boolQueryBuilderRole = QueryBuilders.boolQuery(); String[] roles = (String[]) searchMap.get("role"); for(String role : roles){ QueryBuilder queryBuilder = QueryBuilders.matchPhraseQuery("role_ids",role); boolQueryBuilderRole.should().add(queryBuilder); } boolQueryBuilder.filter(boolQueryBuilderRole); } // 用户名过滤 if(StringUtils.isNotBlank(searchMap.get("user_name")+"") && searchMap.containsKey("user_name")){ QueryBuilder queryBuilder = QueryBuilders.termQuery("user_name.keyword",searchMap.get("user_name")); boolQueryBuilder.filter(queryBuilder); } return boolQueryBuilder;
}
}
// 连接的index
SearchRequest searchRequest = new SearchRequest();
searchRequest.indices(searchIndex);
searchRequest.types("analysed");
SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();
// 自己组装查询
BoolQueryBuilder queryBuilder = ElasticSearchBuilderQuery.buildBasicQuery(pageData);
SimpleDateFormat df = new SimpleDateFormat("yyyy-MM-dd");//格式化为年月
// 基础查询
QueryBuilder query = QueryBuilders.boolQuery().must(queryBuilder);
//script = "doc[‘date‘] +‘####‘+doc[‘user_name.keyword‘]";
// 根据日期分组
TermsAggregationBuilder aggByDate = AggregationBuilders.terms("by_date")
.field("date").size(365);
// 根据用户名分组
TermsAggregationBuilder aggByUser = AggregationBuilders.terms("by_userName")
.field("user_name.keyword").size(10000);
// 根据有效时间统计
SumAggregationBuilder sum = AggregationBuilders.sum("sum").field("valid_time");
// 根据持续时间统计
SumAggregationBuilder duringSum = AggregationBuilders.sum("during_sum").field("during_time_sec");
// 定义需要的返回值字段
String[] returnField = new String[]{"user_name","date","sum","nickname"};
TopHitsAggregationBuilder source = AggregationBuilders.topHits("source").fetchSource(returnField, new String[]{}).size(1);
aggByDate.subAggregation(aggByUser.subAggregation(sum).subAggregation(duringSum).subAggregation(source));
sourceBuilder.query(query).size(0);
sourceBuilder.aggregation(aggByDate);
sourceBuilder.timeout(new TimeValue(60, TimeUnit.SECONDS));
searchRequest.source(sourceBuilder);
try {
SearchResponse searchResponse = highLevelClient.search(searchRequest, RequestOptions.DEFAULT);
//
ParsedLongTerms byDate =(ParsedLongTerms) searchResponse.getAggregations().get("by_date");
// 获取桶
List<ParsedLongTerms.ParsedBucket> buckets = (List<ParsedLongTerms.ParsedBucket>) byDate.getBuckets();
buckets.forEach(bucket -> {
// 桶中数据通过自己的业务进行获取 后组装返回
// 通过的数据可以通过调试,获取sourceBuilder的值,通过Kb去查询查看,更好理解
});
} catch (IOException e) {
e.printStackTrace();
}
- 总结
在初次使用es时,有很多概念不懂,在查询时,将查询的数据全部返回了,然后自己通过代码给做统计,这样的查询出来的所花的时间很多。通过使用聚合函数,查询出自己想要的数据,速度很快。
注意点: (1)如果某个字段要进行分组时,那需要在建立mapping时,设置type为keyword。在没有设置keyword时,类型默认text,分组这个字段将没有结果。
(2)通过日期区间查询时,需要设置timezone这个参数,又可能是我的配置不对,这个看个人,如果是date的日期相差一天的话,建立设置时区。
(3)使用SumAggregationBuilder聚合后,该参数不可设置sub
(4) 聚合后的数据无法分页,但是有size参数,官网是说不可以。
elasticsearch 的查询,聚合,统计总结
标签:val 无法 转换 统计 聚合函数 timeout bag 个人 imp
原文地址:https://blog.51cto.com/12947828/2563499