elasticsearch笔记(7) 复合查询_boolean查询 boolsting查询 filter查询

6 复合查询

6.1 bool查询

将多个查询条件以一定的逻辑组合在一起

must：表示and的意思，所有的条件都符合才能找到

must_not：把满足条件的都去掉的结果

should：表示or的意思

# 查询省份是上海或者河南
# 运营商不是联通
# smsContent中包含中国和移动
# bool查询

 1 POST /sms-logs-index/sms-logs-type/_search
 2 {
 3   "query": {
 4     "bool": {
 5       "should": [
 6         {
 7           "term": {
 8             "province": {
 9               "value": "河南"
10             }
11           }
12         },  
13         {
14           "term": {
15             "province": {
16               "value": "北京"
17              }
18             }
19          }
20       ],
21       "must_not": [
22         {
23           "term": {
24             "operatorId": {
25               "value": "2"
26             }
27           }
28         }
29       ],
30       "must": [
31         {
32           "match": {
33             "smsContent": "中国"
34           }
35         },
36         {
37           "match": {
38             "smsContent": "移动"
39           }
40         }
41       ]
42     }
43   }
44 }

java中的聚合查询

 1 @Test
 2     public void bolleanTest() throws IOException {
 3         //        1. 创建查询对象
 4         String index = "sms-logs-index";
 5         String type = "sms-logs-type";
 6         SearchRequest searchRequest = new SearchRequest(index);//指定索引
 7         searchRequest.types(type);//指定类型
 8 //    2. 封装查询条件
 9         SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
10         BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();
11         boolQueryBuilder.should(QueryBuilders.termQuery("province","河南"));
12         boolQueryBuilder.should(QueryBuilders.termQuery("province","北京"));
13         boolQueryBuilder.mustNot(QueryBuilders.termQuery("operatorId","2"));
14         boolQueryBuilder.must(QueryBuilders.matchQuery("smsContent","中国"));
15         boolQueryBuilder.must(QueryBuilders.matchQuery("smsContent","移动"));
16         searchSourceBuilder.query(boolQueryBuilder);
17         searchRequest.source(searchSourceBuilder);
18 
19 
20 
21         //        3.执行查询
22         // client执行
23         HttpHost httpHost = new HttpHost("192.168.43.30", 9200);
24         RestClientBuilder restClientBuilder = RestClient.builder(httpHost);
25         RestHighLevelClient restHighLevelClient = new RestHighLevelClient(restClientBuilder);
26         SearchResponse search = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
27 
28 //        4.获取数据(source中的数据)
29         SearchHit[] hits = search.getHits().getHits();
30         for(SearchHit searchHit : hits){
31             System.out.println(searchHit);
32         }
33     }

6.2 boolsting

分数查询，比如math查询,结果有一个匹配度分数. boolsting查询可以针对内容，让其分数大，或者小，达到排前，排后的效果

positive：指定查询条件, 只有匹配到positive的内容，才会放到结果集，也就是放查询条件的地方
negative：如果匹配到的positive和negative，就会降低文档的分数
negative_boost：对匹配到的positive和negative的内容, 指定降低分数的系数. 这个系数必须小于1.0，比如：10分这个系数为0.5就会变为5分

关于分数的计算：

关键字在文档出现的频次越高，分数越高
文档的内容越短，分数越高
搜索时候，指定的关键字会被分词，分词内容匹配分词库，匹配的个数越多，分数就越高

 1 # boosting查询
 2 POST /sms-logs-index/sms-logs-type/_search
 3 {
 4   "query": {
 5     "boosting": {
 6       "positive": { #这三个postitive. negative 和 negative_boost是在一级的.
 7         "match": {
 8           "smsContent": {
 9              "query": "电话玩家活物",
10                "operator": "or"
11           }
12          }
13       },
14       "negative": { //匹配到这个的就按negative_boost在分数中减少
15         "match": {
16           "smsContent": "订单"
17         }
18       },
19       "negative_boost": 0.1
20     }
21   }
22 }

java中用boosting查询

 1     @Test
 2     public void boostingTest() throws IOException {
 3         //        1. 创建查询对象
 4         String index = "sms-logs-index";
 5         String type = "sms-logs-type";
 6         SearchRequest searchRequest = new SearchRequest(index);//指定索引
 7         searchRequest.types(type);//指定类型
 8 //    2. 封装查询条件
 9         SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
10 
11         BoostingQueryBuilder boostingQueryBuilder = QueryBuilders.boostingQuery(
12                 QueryBuilders.matchQuery("smsContent", "电话").operator(Operator.OR),//相当于positivequery
13                 QueryBuilders.matchQuery("smsContent", "订单")
14         ).negativeBoost(0.1f);
15 
16         searchSourceBuilder.query(boostingQueryBuilder);
17         searchRequest.source(searchSourceBuilder);
18 
19 
20 
21         //        3.执行查询
22         // client执行
23         HttpHost httpHost = new HttpHost("192.168.43.30", 9200);
24         RestClientBuilder restClientBuilder = RestClient.builder(httpHost);
25         RestHighLevelClient restHighLevelClient = new RestHighLevelClient(restClientBuilder);
26         SearchResponse search = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
27 
28 //        4.获取数据(source中的数据)
29         SearchHit[] hits = search.getHits().getHits();
30         for(SearchHit searchHit : hits){
31             System.out.println(searchHit);
32         }
33     }

二, filter查询: 如果结果不需要匹配度

与query查询相比, 会计算分数, 并对结果排序

filter不计算分数, 结果也不排序

filter查询在java中应用

 1     @Test
 2     public void filterTest() throws IOException {
 3         //        1. 创建查询对象
 4         String index = "sms-logs-index";
 5         String type = "sms-logs-type";
 6         SearchRequest searchRequest = new SearchRequest(index);//指定索引
 7         searchRequest.types(type);//指定类型
 8 //    2. 封装查询条件
 9         SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
10         BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();//先创建Boolean查询对象
11         boolQueryBuilder.filter(QueryBuilders.termQuery("corpName", "网易"));
12 
13 
14         searchSourceBuilder.query(boolQueryBuilder);
15         searchRequest.source(searchSourceBuilder);
16 
17 
18 
19         //        3.执行查询
20         // client执行
21         HttpHost httpHost = new HttpHost("192.168.43.30", 9200);
22         RestClientBuilder restClientBuilder = RestClient.builder(httpHost);
23         RestHighLevelClient restHighLevelClient = new RestHighLevelClient(restClientBuilder);
24         SearchResponse search = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
25 
26 //        4.获取数据(source中的数据)
27         SearchHit[] hits = search.getHits().getHits();
28         for(SearchHit searchHit : hits){
29             System.out.println(searchHit);
30         }
31     }

三. 高亮查询: 将查询到的内容自动加到高亮标签中.

将用户输入的内容，以高亮的样式展示出来，查询的结果会附带在hits下面以单独的形式返回，不会影响查询的结果

ES提供了一个hightlight的属性，和query同级别，其属性如下：

fragment_size：指定要展示多少内容，可以看到百度的内容后面有...还有很长，默认100个
pre_tags：指定前缀标签比如：就是红色
post_tags：指定后缀标签：
fields：指定哪几个field以高亮形式返回

java中用highlight查询

 1     @Test
 2     public void highlightQuery() throws IOException {
 3         //        1. 创建查询对象
 4         String index = "sms-logs-index";
 5         String type = "sms-logs-type";
 6         SearchRequest searchRequest = new SearchRequest(index);//指定索引
 7         searchRequest.types(type);//指定类型
 8 //    2. 封装查询条件
 9         SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
10         searchSourceBuilder.query(QueryBuilders.matchQuery("smsContent", "网易"));
11         HighlightBuilder highlightBuilder = new HighlightBuilder();
12         highlightBuilder.field("smsContent") .preTags("<font color='red'>").postTags("</font>").fragmentSize(20);
13         searchSourceBuilder.highlighter(highlightBuilder);
14         searchRequest.source(searchSourceBuilder);
15 
16         //        3.执行查询
17         // client执行
18         HttpHost httpHost = new HttpHost("192.168.43.30", 9200);
19         RestClientBuilder restClientBuilder = RestClient.builder(httpHost);
20         RestHighLevelClient restHighLevelClient = new RestHighLevelClient(restClientBuilder);
21         SearchResponse search = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
22 
23 //        4.获取数据(source中的数据)
24         SearchHit[] hits = search.getHits().getHits();
25         for(SearchHit searchHit : hits){
26             System.out.println(searchHit);
27         }
28     }

posted @ 2021-02-21 12:10 当当和瓶瓶阅读(557) 评论(0) 编辑收藏举报

刷新页面返回顶部

当当和瓶瓶

elasticsearch笔记(7) 复合查询_boolean查询 boolsting查询 filter查询

6 复合查询

6.1 bool查询

6.2 boolsting

公告