es实现多字段聚合后再分页-附JAVA工具类

雨点打透心脏的1/2处 2022-11-16 10:39 328阅读 0赞

因为产品功能需要，需要实现很多数据的统计功能。产品经理在统计某些数据时为了不失用户体验，对于数据量可能大于几十条的数据都设计为了分页查询

*好巧不巧，上面描述的这个需求的后端开发正好就落在了我的头上*

*刚开始没怎么注意，想着不就是从es提取数据嘛，看看官方API定能分分钟搞定,而且肯定有现成的聚合后再分页的api*

**产品说10天时间让开发完，我想着10天是不是瞧不起我。3天，就3天完全够了。第1天开发接口，还不能下班后在公司逗留、第2天让前端对接接口、第3天走测试。好在产品也是很给我面子，说3天就3天（此段纯属虚构，切勿模仿）**

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center]

但实际情况确实 to young to simple

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center 1]

## 常见es聚合后再分页解决方式 ##

当我翻阅了elastic search官方的文档，却没有找到关于多字段聚合后再分页的现成的搜索语句，但是看到了[terms聚合说明部分][terms]这么一段话

> If you want to retrieve all terms or all combinations of terms in a nested terms aggregation you should use the Composite aggregation which allows to paginate over all possible terms rather than setting a size greater than the cardinality of the field in the terms aggregation. The terms aggregation is meant to return the top terms and does not allow pagination.

跟着文档又看了下[composite聚合函数][composite]的使用方法,也正如terms聚合函数介绍中的那样，它确实能实现分页。但是，它的分页不能任意跳转到某个页号,仅能采用类似scroll的方式实现分页，感觉和我所期望的和数据库中实现的那种分页还是有点不一样

> Unlike the other multi-bucket aggregations, you can use the composite aggregation to paginate all buckets from a multi-level aggregation efficiently. This aggregation provides a way to stream all buckets of a specific aggregation, similar to what scroll does for documents.

但是在上面看了一个multi-bucket aggregations,随后又了解了一下[multi terms][]的介绍,看了下这个感觉确实是我想要的，可以实现多字段聚合与分页

> As most bucket aggregations the multi\_term supports sub aggregations and ordering the buckets by metrics sub-aggregation

于是乎我便打算开始用multi terms来实现其中好几个需要聚合后再分页的需求了,但实现时才发现我们的es用的不是官方的7.X版本的，而是6.X版本的，6.X版本的es压根就不支持multi terms这个聚合函数。

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center 2]

时间紧迫毕竟牛逼是不能随便吹的，后面我便只好亲自动手了，用万能的**script**搭配着[bucket sort][]来进行了实现

## 自定义多字段聚合并支持分页java工具类 ##

import org.elasticsearch.action.search.SearchRequest;
    import org.elasticsearch.action.search.SearchResponse;
    import org.elasticsearch.client.RequestOptions;
    import org.elasticsearch.client.RestHighLevelClient;
    import org.elasticsearch.index.query.BoolQueryBuilder;
    import org.elasticsearch.script.Script;
    import org.elasticsearch.script.ScriptType;
    import org.elasticsearch.search.aggregations.AggregationBuilders;
    import org.elasticsearch.search.aggregations.bucket.terms.Terms;
    import org.elasticsearch.search.aggregations.bucket.terms.TermsAggregationBuilder;
    import org.elasticsearch.search.aggregations.metrics.cardinality.Cardinality;
    import org.elasticsearch.search.aggregations.pipeline.bucketsort.BucketSortPipelineAggregationBuilder;
    import org.elasticsearch.search.builder.SearchSourceBuilder;
    
    import java.util.ArrayList;
    import java.util.Arrays;
    import java.util.HashMap;
    import java.util.List;
    
    /** * es分页查询执行器 * * @author puhaiyang * @date 2021/4/6 15:16 */
    public class EsAggPaginationQueryRunner { 
        private final RestHighLevelClient restHighLevelClient;
        /** * searchRequest对象 */
        private final SearchRequest searchRequest = new SearchRequest();
        /** * 不需要返回数据,size设为0 */
        private final SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder().size(0);
        /** * 分组统计的字段 */
        private String[] groupByFields;
        /** * 查询时的script */
        private Script queryScript;
        /** * 分隔符 */
        private final String spiltSymbol = "@#@#@";
    
        private EsAggPaginationQueryRunner(RestHighLevelClient restHighLevelClient) { 
            this.restHighLevelClient = restHighLevelClient;
        }
    
        public EsAggPaginationQueryRunner build(String esIndexName, BoolQueryBuilder boolQueryBuilder, String... groupByFields) { 
            //添加查询条件
            searchSourceBuilder.query(boolQueryBuilder);
            //指定索引
            searchRequest.indices(esIndexName);
            //指定分组的字段
            this.groupByFields = groupByFields;
            //查询script初始化
            queryScript = initScript();
            return this;
        }
    
        public static EsAggPaginationQueryRunner client(RestHighLevelClient restHighLevelClient) { 
            //建立一个新对象
            return new EsAggPaginationQueryRunner(restHighLevelClient);
        }
    
    
        public Script initScript() { 
            //定义script
            StringBuilder scriptBuilder = new StringBuilder();
            for (int i = 0; i < groupByFields.length; i++) { 
                if (i != 0) { 
                    scriptBuilder.append("+'" + spiltSymbol + "'+");
                }
                scriptBuilder.append(String.format("doc['%s'].value", groupByFields[i]));
            }
            return new Script(ScriptType.INLINE, "painless", scriptBuilder.toString(), new HashMap<>(0));
        }
    
        /** * 进行聚合 */
        public ResultTotal getResultTotal() throws Exception { 
            //定义script
            StringBuilder scriptBuilder = new StringBuilder();
            for (int i = 0; i < groupByFields.length; i++) { 
                if (i != 0) { 
                    scriptBuilder.append(" +' '+");
                }
                scriptBuilder.append(String.format("doc['%s'].value", groupByFields[i]));
            }
    
            Script getTotalScript = new Script(ScriptType.INLINE, "painless", scriptBuilder.toString(), new HashMap<>(0));
            SearchSourceBuilder distinctByFieldAgg = searchSourceBuilder.aggregation(AggregationBuilders.cardinality("distinct_by_field").script(getTotalScript));
            //指定source
            searchRequest.source(distinctByFieldAgg);
            //执行查询
            SearchResponse searchResponse = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
            //获取出查询结果
            long groupTotal = ((Cardinality) searchResponse.getAggregations().get("distinct_by_field")).getValue();
            long totalHits = searchResponse.getHits().totalHits;
            return new ResultTotal(groupTotal, totalHits);
        }
    
        public List<KeyValueDTO<List<String>, Long>> getListByPagination(int offset, int limit) throws Exception { 
            List<KeyValueDTO<List<String>, Long>> resultList = new ArrayList<>();
    
            int size = offset + limit;
            //传入一下新的聚合函数进行聚合,使用bucket_sort实现
            TermsAggregationBuilder groupByFieldAgg = AggregationBuilders.terms("group_by_field").script(queryScript).size(size);
            //定义分页条件
            BucketSortPipelineAggregationBuilder bucketSort = new BucketSortPipelineAggregationBuilder("bucket_sort", null)
                    .from(offset)
                    .size(limit);
            //添加分页内容
            groupByFieldAgg.subAggregation(bucketSort);
            //指定agg
            searchSourceBuilder.aggregation(groupByFieldAgg);
            SearchResponse searchResponse = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
            Terms terms = searchResponse.getAggregations().get("group_by_field");
            List<? extends Terms.Bucket> buckets = terms.getBuckets();
            for (Terms.Bucket bucketItem : buckets) { 
                String[] keyArray = bucketItem.getKeyAsString().split(spiltSymbol);
    
                resultList.add(new KeyValueDTO<>(Arrays.asList(keyArray), bucketItem.getDocCount()));
            }
            return resultList;
        }
    }
    
    
    /** * KeyValueDTO * * @author puhaiyang * @date 2021/4/6 15:31 */
    public class KeyValueDTO<T, V> { 
        private T key;
        private V value;
    
        public KeyValueDTO(T key, V value) { 
            setKey(key);
            setValue(value);
        }
    
        /**getter and setter method**/
    }
    
    /** * 查询结果集-total * * @author puhaiyang * @date 2021/4/5 11:46 */
    
    public class ResultTotal { 
        /** * 记录条数,聚合后的条数 */
        private Long recordTotal;
        /** * 所有满足条件的总数 */
        private Long sumTotal;
    
        public ResultTotal(Long recordTotal, Long sumTotal) { 
            this.recordTotal = recordTotal;
            this.sumTotal = sumTotal;
        }
        
        /**getter and setter method**/
    
    }

### 使用示例 ###

//指定EsAggPaginationQueryRunner的客户端，并创建一个EsAggPaginationQueryRunner出来
            EsAggPaginationQueryRunner esAggPaginationQueryRunner = EsAggPaginationQueryRunner.client(restHighLevelClient).
                    build("esIndexName", new BoolQueryBuilder(), "product_price","product_name");
            //获取分页数据
            ResultTotal resultTotal = esAggPaginationQueryRunner.getResultTotal();
            System.out.println("总记录数:"+resultTotal.getSumTotal()+" 数据条数:"+resultTotal.getRecordTotal());
            if (resultTotal.getRecordTotal().intValue() > 0) { 
                //有数据,进行分页查询
                List<KeyValueDTO<List<String>, Long>> pageResultList = esAggPaginationQueryRunner.getListByPagination(pageOffet, pageLimit);
            }

有了支持多字段聚合后再分页的工具类后，感觉一切都那么顺畅，噼里啪啦好几个类似的需求被我一把梭掉的感觉真不错。*(感谢公司领导和产品经理将这个开发任务指派给了我)*

## 总结 ##

对于es中需要实现聚合后再分页的功能，如果是高版本的es可以考虑采用[multi terms聚合函数][multi terms]来实现

如果是稍微低一点版本的es可以使用[composite聚合函数][composite]来实现，使用滚动分页(**类scroll**)来进行分页

如果数据量不是很大，则可以考虑用传统的terms+script来实现多字段分组，并采用bucket sort实现排序

如果数据量实在太大还想实现聚合后分页查询，个人感觉得让产品改改需求了，可以来个【最后数据刷新时间】啥的，先用scroll慢慢将统计结果集放到关系型数据库中，再用sql语句直接查询关系型数据库

当然最重要的还是看具体的业务场景，在适合的情况下选择最适合自己的那个就对了

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center]: /images/20221022/4badc395cd384d8b8d77cd040d994ab9.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center 1]: /images/20221022/a8202f932ec14d938a15224b2f139735.png
[terms]: https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-terms-aggregation.html
[composite]: https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-composite-aggregation.html
[multi terms]: https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-multi-terms-aggregation.html
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3B1aGFpeWFuZw_size_16_color_FFFFFF_t_70_pic_center 2]: /images/20221022/44ee6d58893741baa9b07c523d941a03.png
[bucket sort]: https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-pipeline-bucket-sort-aggregation.html