ElasticSearch入门之彼行我释(四)

时间:2021-04-10 19:59:08
ElasticSearch入门之彼行我释(四)

散仙在上篇文章中,介绍了关于ElasticSearch基本的增删改查的基本粒子,本篇呢,我们来学下稍微高级一点的知识: 


(1)如何在ElasticSearch中批量提交索引 ? 
(2)如何使用高级查询(包括,检索,排序,过滤,分页) ? 
(3)如何组合多个查询 ? 
(4)如何使用翻页深度查询 ? 
(5)如何使用基本的聚合查询 ? 



(一)首先,我们思考下,为什么要使用批量添加,这个毫无疑问,因为效率问题,举个在生活中的例子,假如我们有50个人,要去美国旅游,不使用批处理的方式是,给每一个人派一架飞机送到美国,那么这就需要50次飞机的来回往来,假如使用了批处理,现在的情况就是一个飞机坐50个人,只需一次即可把所有人都送到美国,效率可想而知,生活也有很多实际的例子,大家可以自己想想。 

在原生的lucene中,以及solr中,这个批处理方式,实质是控制commit的时机,比如多少个提交一次,或者超过ranbuffersize的大小后自动提交,es封装了lucene的api提供bulk的方式来批量添加,原理也是,聚集一定的数量doc,然后发送一次添加请求。 


(二)只要我们使用了全文检索,我们的业务就会有各种各样的api操作,包括,任意维度的字段查询,过滤掉某些无效的信息,然后根据某个字段排序,再取topN的结果集返回,使用数据库的小伙伴们,相信大家都不陌生,在es中,这些操作都是支持的,而且还非常高效,它能满足我们大部分的需求 


(三)在es中,我们可以查询多个index,以及多个type,这一点是非常灵活地,我们,我们可以一次组装两个毫无关系的查询,发送到es服务端进行检索,然后获取结果。 


(四)es中,通过了scorll的方式,支持深度分页查询,在数据库里,我们使用的是一个cursor游标来记录读取的偏移量,同样的在es中也支持,这样的查询方式,它通过一个scrollid记录了上一次查询的状态,能轻而易举的实现深度翻页,本质上是对了Lucene的SearchAfter的封装。 

(五)es中,也提供了对聚合函数的支持,比如一些max,min,avg,count,sum等支持,除此之外还支持group,facet等操作,这些功能,在电商中应用非常广泛,基于lucene的solr和es都有很好的支持。 

下面截图看下散仙的测试数据值: 


ElasticSearch入门之彼行我释(四)
源码demo如下:  
Java代码  ElasticSearch入门之彼行我释(四)
  1. package com.dongliang.es;  
  2.   
  3. import java.util.Date;  
  4. import java.util.Map;  
  5. import java.util.Map.Entry;  
  6.   
  7. import org.apache.lucene.index.Terms;  
  8. import org.elasticsearch.action.bulk.BulkRequestBuilder;  
  9. import org.elasticsearch.action.bulk.BulkResponse;  
  10. import org.elasticsearch.action.search.MultiSearchResponse;  
  11. import org.elasticsearch.action.search.SearchRequestBuilder;  
  12. import org.elasticsearch.action.search.SearchResponse;  
  13. import org.elasticsearch.action.search.SearchType;  
  14. import org.elasticsearch.client.Client;  
  15. import org.elasticsearch.client.transport.TransportClient;  
  16. import org.elasticsearch.common.transport.InetSocketTransportAddress;  
  17. import org.elasticsearch.common.unit.TimeValue;  
  18. import org.elasticsearch.common.xcontent.XContentBuilder;  
  19. import org.elasticsearch.common.xcontent.XContentFactory;  
  20. import org.elasticsearch.index.query.FilterBuilders;  
  21. import org.elasticsearch.index.query.QueryBuilders;  
  22. import org.elasticsearch.index.query.QueryStringQueryBuilder;  
  23. import org.elasticsearch.search.SearchHit;  
  24. import org.elasticsearch.search.aggregations.AggregationBuilders;  
  25. import org.elasticsearch.search.aggregations.bucket.filters.InternalFilters.Bucket;  
  26. import org.elasticsearch.search.sort.SortOrder;  
  27.   
  28. /** 
  29.  * @author 三劫散仙 
  30.  * 搜索技术交流群:324714439  
  31.  * 一个关于elasticsearch批量提交 
  32.  * 和search query的的例子 
  33.  * **/  
  34. public class ElasticSearchDao {  
  35.       
  36.       
  37.     //es的客户端实例  
  38.     Client client=null;  
  39.     {  
  40.         //连接单台机器,注意ip和端口号,不能写错  
  41.         client=new TransportClient().  
  42.                 addTransportAddress(new InetSocketTransportAddress("192.168.46.16"9300));  
  43.           
  44.     }  
  45.       
  46.       
  47.     public static void main(String[] args)throws Exception {  
  48.         ElasticSearchDao es=new ElasticSearchDao();  
  49.         //es.indexdata();//索引数据  
  50.         //es.queryComplex();  
  51.         es.querySimple();  
  52.         //es.scorllQuery();  
  53.         //es.mutilCombineQuery();  
  54.         //es.aggregationQuery();  
  55.     }  
  56.       
  57.       
  58.     /**组合分组查询*/  
  59.     public void aggregationQuery()throws Exception{  
  60.         SearchResponse sr = client.prepareSearch()  
  61.                 .setQuery(QueryBuilders.matchAllQuery())  
  62.                 .addAggregation(  
  63.                         AggregationBuilders.terms("1").field("type")  
  64.                 )  
  65. //              .addAggregation(  
  66. //                      AggregationBuilders.dateHistogram("agg2")  
  67. //                              .field("birth")  
  68. //                              .interval(DateHistogram.Interval.YEAR)  
  69. //              )  
  70.                 .execute().actionGet();  
  71.   
  72.             // Get your facet results  
  73.             org.elasticsearch.search.aggregations.bucket.terms.Terms a = sr.getAggregations().get("1");  
  74.               
  75.             for(org.elasticsearch.search.aggregations.bucket.terms.Terms.Bucket bk:a.getBuckets()){  
  76.                 System.out.println("类型: "+bk.getKey()+"  分组统计数量 "+bk.getDocCount()+"  ");  
  77.             }  
  78.               
  79.             System.out.println("聚合数量:"+a.getBuckets().size());  
  80.             //DateHistogram agg2 = sr.getAggregations().get("agg2");  
  81.             //结果:  
  82. //          类型: 1  分组数量 2    
  83. //          类型: 2  分组数量 1    
  84. //          类型: 3  分组数量 1    
  85. //          聚合数量:3  
  86.     }  
  87.       
  88.       
  89.       
  90.       
  91.     /**多个不一样的请求组装*/  
  92.     public void mutilCombineQuery(){  
  93.           
  94.         //查询请求1  
  95.         SearchRequestBuilder srb1 =client.prepareSearch().setQuery(QueryBuilders.queryString("eng").field("address")).setSize(1);  
  96.         //查询请求2//matchQuery  
  97.         SearchRequestBuilder srb2 = client.prepareSearch().setQuery(QueryBuilders.matchQuery("title""标题")).setSize(1);  
  98.         //组装查询  
  99.         MultiSearchResponse sr = client.prepareMultiSearch().add(srb1).add(srb2).execute().actionGet();  
  100.   
  101.             // You will get all individual responses from MultiSearchResponse#getResponses()  
  102.             long nbHits = 0;  
  103.             for (MultiSearchResponse.Item item : sr.getResponses()) {  
  104.                 SearchResponse response = item.getResponse();  
  105.                 for(SearchHit hits:response.getHits().getHits()){  
  106.                     String sourceAsString = hits.sourceAsString();//以字符串方式打印  
  107.                     System.out.println(sourceAsString);  
  108.                 }  
  109.                 nbHits += response.getHits().getTotalHits();  
  110.             }  
  111.         System.out.println("命中数据量:"+nbHits);  
  112.         //输出:  
  113. //      {"title":"我是标题","price":25.65,"type":1,"status":true,"address":"血落星域风阳星","createDate":"2015-03-16T09:56:20.440Z"}  
  114. //      命中数据量:2  
  115.   
  116.         client.close();  
  117.     }  
  118.       
  119.       
  120.     /** 
  121.      * 翻页查询 
  122.      * */  
  123.     public void scorllQuery()throws Exception{  
  124.         QueryStringQueryBuilder queryString = QueryBuilders.queryString("标题").field("title");  
  125.         //TermQueryBuilder qb=QueryBuilders.termQuery("title", "我是标题");  
  126.         SearchResponse scrollResp = client.prepareSearch("collection1")  
  127.                  .setSearchType(SearchType.SCAN)  
  128.                  .setScroll(new TimeValue(60000))  
  129.                  .setQuery(queryString)  
  130.                  .setSize(100).execute().actionGet(); //100 hits per shard will be returned for each scroll  
  131.            
  132.           
  133.         while (true) {  
  134.             for (SearchHit hit : scrollResp.getHits().getHits()) {  
  135.                 //Handle the hit...  
  136.                 String sourceAsString = hit.sourceAsString();//以字符串方式打印  
  137.                 System.out.println(sourceAsString);  
  138.             }  
  139.             //通过scrollid来实现深度翻页  
  140.             scrollResp = client.prepareSearchScroll(scrollResp.getScrollId()).setScroll(new TimeValue(600000)).execute().actionGet();  
  141.             //Break condition: No hits are returned  
  142.             if (scrollResp.getHits().getHits().length == 0) {  
  143.                 break;  
  144.             }  
  145.         }  
  146.         //输出  
  147. //      {"title":"我是标题","price":25.65,"type":1,"status":true,"address":"血落星域风阳星","createDate":"2015-03-16T09:56:20.440Z"}  
  148. //      {"title":"标题","price":251.65,"type":1,"status":true,"address":"美国东部","createDate":"2015-03-16T10:33:58.743Z"}  
  149.         client.close();  
  150.           
  151.     }  
  152.       
  153.     /**简单查询*/  
  154.     public void querySimple()throws Exception{  
  155.           
  156.         SearchResponse sp = client.prepareSearch("collection1").execute().actionGet();  
  157.         for(SearchHit hits:sp.getHits().getHits()){  
  158.             String sourceAsString = hits.sourceAsString();//以字符串方式打印  
  159.             System.out.println(sourceAsString);  
  160.         }  
  161.           
  162.           
  163.     //结果  
  164. //              {"title":"我是标题","price":25.65,"type":1,"status":true,"address":"血落星域风阳星","createDate":"2015-03-16T09:56:20.440Z"}  
  165. //              {"title":"中国","price":205.65,"type":2,"status":true,"address":"河南洛阳","createDate":"2015-03-16T10:33:58.740Z"}  
  166. //              {"title":"标题","price":251.65,"type":1,"status":true,"address":"美国东部","createDate":"2015-03-16T10:33:58.743Z"}  
  167. //              {"title":"elasticsearch是一个搜索引擎","price":25.65,"type":3,"status":true,"address":"china","createDate":"2015-03-16T10:33:58.743Z"}  
  168.   
  169.           
  170.     }  
  171.     /**组合查询**/  
  172.     public void queryComplex()throws Exception{  
  173.         SearchResponse sp=client.prepareSearch("collection1")//检索的目录  
  174.                 .setTypes("core1")//检索的索引  
  175.                 .setSearchType(SearchType.DFS_QUERY_THEN_FETCH)//Query type  
  176.                 .setQuery(QueryBuilders.termQuery("type""1"))//查询--Query   
  177.                 .setPostFilter(FilterBuilders.rangeFilter("price").from(10).to(550.23))//过滤 --Filter  
  178.                 .addSort("price",SortOrder.DESC) //排序 -- sort  
  179.                 .setFrom(0).setSize(20).setExplain(true)//topN方式  
  180.                 .execute().actionGet();//执行  
  181.                 System.out.println("本次查询命中条数: "+sp.getHits().getTotalHits());  
  182.                 for(SearchHit hits:sp.getHits().getHits()){  
  183.                     //String sourceAsString = hits.sourceAsString();//以字符串方式打印  
  184.                     //System.out.println(sourceAsString);  
  185.                     Map<String, Object> sourceAsMap = hits.sourceAsMap();  
  186.                     for(Entry<String, Object> k:sourceAsMap.entrySet()){  
  187.                         System.out.println("name: "+k.getKey()+"     value: "+k.getValue());  
  188.                     }  
  189.                       
  190.                     System.out.println("=============================================");  
  191.                       
  192.                 }  
  193.           
  194.                 //结果  
  195. //              本次查询命中条数: 2  
  196. //              name: title     value: 标题  
  197. //              name: price     value: 251.65  
  198. //              name: address     value: 美国东部  
  199. //              name: status     value: true  
  200. //              name: createDate     value: 2015-03-16T10:33:58.743Z  
  201. //              name: type     value: 1  
  202. //              =============================================  
  203. //              name: title     value: 我是标题  
  204. //              name: price     value: 25.65  
  205. //              name: address     value: 血落星域风阳星  
  206. //              name: status     value: true  
  207. //              name: createDate     value: 2015-03-16T09:56:20.440Z  
  208. //              name: type     value: 1  
  209. //              =============================================  
  210.           
  211.         client.close();  
  212.     }  
  213.       
  214.       
  215.       
  216.     /**索引数据*/  
  217.     public void indexdata()throws Exception{  
  218.           
  219.         BulkRequestBuilder bulk=client.prepareBulk();  
  220.           
  221.         XContentBuilder doc=XContentFactory.jsonBuilder()  
  222.                 .startObject()  
  223.                 .field("title","中国")  
  224.                 .field("price",205.65)  
  225.                 .field("type",2)  
  226.                 .field("status",true)  
  227.                 .field("address""河南洛阳")  
  228.                 .field("createDate"new Date()).endObject();  
  229.         //collection为索引库名,类似一个数据库,索引名为core,类似一个表  
  230. //       client.prepareIndex("collection1", "core1").setSource(doc).execute().actionGet();  
  231.           
  232.         //批处理添加  
  233.         bulk.add(client.prepareIndex("collection1""core1").setSource(doc));  
  234.           
  235.         doc=XContentFactory.jsonBuilder()  
  236.                 .startObject()  
  237.                 .field("title","标题")  
  238.                 .field("price",251.65)  
  239.                 .field("type",1)  
  240.                 .field("status",true)  
  241.                 .field("address""美国东部")  
  242.                 .field("createDate"new Date()).endObject();  
  243.         //collection为索引库名,类似一个数据库,索引名为core,类似一个表  
  244. //      client.prepareIndex("collection1", "core1").setSource(doc).execute().actionGet();  
  245.         //批处理添加  
  246.         bulk.add(client.prepareIndex("collection1""core1").setSource(doc));  
  247.           
  248.         doc=XContentFactory.jsonBuilder()  
  249.                 .startObject()  
  250.                 .field("title","elasticsearch是一个搜索引擎")  
  251.                 .field("price",25.65)  
  252.                 .field("type",3)  
  253.                 .field("status",true)  
  254.                 .field("address""china")  
  255.                 .field("createDate"new Date()).endObject();  
  256.         //collection为索引库名,类似一个数据库,索引名为core,类似一个表  
  257.         //client.prepareIndex("collection1", "core1").setSource(doc).execute().actionGet();  
  258.         //批处理添加  
  259.         bulk.add(client.prepareIndex("collection1""core1").setSource(doc));         
  260.           
  261.           
  262.         //发一次请求,提交所有数据  
  263.           BulkResponse bulkResponse = bulk.execute().actionGet();  
  264.           if (!bulkResponse.hasFailures()) {  
  265.               System.out.println("创建索引success!");  
  266.           } else {  
  267.               System.out.println("创建索引异常:"+bulkResponse.buildFailureMessage());  
  268.           }  
  269.           
  270.           
  271.           
  272.         client.close();//释放资源  
  273. //      System.out.println("索引成功!");  
  274.                   
  275.           
  276.           
  277.           
  278.     }  
  279.       
  280.       
  281.       
  282.       
  283.   
  284. }