SpringBoot整合ES高级查询方式
目录
- 1、配置
- 2、API操作ES
- 2.1 查询索引列表
- 2.2 TermsQuery
- 2.3 WildcardQuery
- 2.4 RangeQuery
- 2.5 MatchQuery
- 2.6 MultiMatchQuery
- 2.7 ExistsQuery
- 2.8 BoolQuery
- 2.9 排序
- 2.10 结果字段过滤
- 2.11 分页
- 2.22 聚合
- springboot版本:2.0.5.RELEASE
- elasticsearch版本:7.9.1
1、配置 引入依赖:
org.elasticsearch.client elasticsearch-rest-high-level-client7.9.1 org.elasticsearch elasticsearch7.9.1
application.properties 配置文件:
elasticsearch.schema=httpelasticsearch.address=192.168.80.130:9200,192.168.80.131:9200,192.168.80.132:9200elasticsearch.connectTimeout=10000elasticsearch.socketTimeout=60000elasticsearch.connectionRequestTimeout=10000elasticsearch.maxConnectNum=200elasticsearch.maxConnectPerRoute=200# 无密码可忽略elasticsearch.userName=elasticelasticsearch.password=123456
连接配置:
import org.apache.http.HttpHost; import org.apache.http.auth.AuthScope; import org.apache.http.auth.UsernamePasswordCredentials; import org.apache.http.client.CredentialsProvider; import org.apache.http.impl.client.BasicCredentialsProvider; import org.elasticsearch.client.RestClient; import org.elasticsearch.client.RestClientBuilder; import org.elasticsearch.client.RestHighLevelClient; import org.springframework.beans.factory.annotation.Value; import org.springframework.context.annotation.Bean; import org.springframework.context.annotation.Configuration; import javax.annotation.PreDestroy; import java.io.IOException; import java.util.ArrayList; import java.util.List; @Configurationpublic class ElasticSearchConfig {/*** 协议*/@Value("${elasticsearch.schema:http}")private String schema; /*** 集群地址,如果有多个用“,”隔开*/@Value("${elasticsearch.address}")private String address; /*** 集群地址,如果有多个用“,”隔开*/@Value("${elasticsearch.userName}")private String userName; /*** 集群地址,如果有多个用“,”隔开*/@Value("${elasticsearch.password}")private String password; /*** 连接超时时间*/@Value("${elasticsearch.connectTimeout:5000}")private int connectTimeout; /*** Socket 连接超时时间*/@Value("${elasticsearch.socketTimeout:10000}")private int socketTimeout; /*** 获取连接的超时时间*/@Value("${elasticsearch.connectionRequestTimeout:5000}")private int connectionRequestTimeout; /*** 最大连接数*/@Value("${elasticsearch.maxConnectNum:100}")private int maxConnectNum; /*** 最大路由连接数*/@Value("${elasticsearch.maxConnectPerRoute:100}")private int maxConnectPerRoute; private RestHighLevelClient restHighLevelClient; @Beanpublic RestHighLevelClient restHighLevelClient() {final CredentialsProvider credentialsProvider = new BasicCredentialsProvider(); UsernamePasswordCredentials elastic = new UsernamePasswordCredentials(userName, password); credentialsProvider.setCredentials(AuthScope.ANY,elastic); // 拆分地址List hostLists = new ArrayList<>(); String[] hostList = address.split(","); for (String addr : hostList) {String host = addr.split(":")[0]; String port = addr.split(":")[1]; hostLists.add(new HttpHost(host, Integer.parseInt(port), schema)); }// 转换成 HttpHost 数组HttpHost[] httpHost = hostLists.toArray(new HttpHost[]{}); // 构建连接对象RestClientBuilder builder = RestClient.builder(httpHost); // 异步连接延时配置builder.setRequestConfigCallback(requestConfigBuilder -> {requestConfigBuilder.setConnectTimeout(connectTimeout); requestConfigBuilder.setSocketTimeout(socketTimeout); requestConfigBuilder.setConnectionRequestTimeout(connectionRequestTimeout); return requestConfigBuilder; }); // 异步连接数配置builder.setHttpClientConfigCallback(httpClientBuilder -> {httpClientBuilder.setMaxConnTotal(maxConnectNum); httpClientBuilder.setMaxConnPerRoute(maxConnectPerRoute); httpClientBuilder.setDefaultCredentialsProvider(credentialsProvider); return httpClientBuilder; }); restHighLevelClient = new RestHighLevelClient(builder); return restHighLevelClient; }@PreDestroypublic void clientClose() {try {this.restHighLevelClient.close(); } catch (IOException e) {e.printStackTrace(); }}}
2、API操作ES
2.1 查询索引列表
可以模糊匹配索引名称
@Testpublic void tset() throws IOException {GetIndexRequest getIndexRequest = new GetIndexRequest("log*"); // 获取es前缀过滤下所有索引GetIndexResponse getIndexResponse = restHighLevelClient.indices().get(getIndexRequest, RequestOptions.DEFAULT); // 将es查出的索引转换为listList elasticsearchList = new ArrayList<>(getIndexResponse.getMappings().keySet()); elasticsearchList.forEach(System.out::println); }
2.2 TermsQuery
es 的 trem query 做的是精确匹配查询,关于这里早 serviceName 字段后面加的 .keyword 说明如下:
1.es5.0 及以后的版本取消了 String 类型,将原先的 String 类型拆分为 text 和 keyword 两种类型。它们的区别在于 text 会对字段进行分词处理而 keyword 则不会。
2.当没有为索引字段预先指定 mapping 的话,es 就会使用 Dynamic Mapping ,通过推断你传入的文档中字段的值对字段进行动态映射。例如传入的文档中字段 total 的值为12,那么 total 将被映射为 long 类型;字段 addr 的值为"192.168.0.1",那么 addr 将被映射为 ip 类型。然而对于不满足 ip 和 long 格式的普通字符串来说,情况有些不同:ES 会将它们映射为 text 类型,但为了保留对这些字段做精确查询以及聚合的能力,又同时对它们做了 keyword 类型的映射,作为该字段的 fields 属性写到 _mapping 中。例如,我这里使用的字段 “serviceName”,用来存储服务名称字符串类型,会对它做如下的 Dynamic Mapping:
"serviceName" : {"type" : "text","fields" : {"keyword" : {"type" : "keyword","ignore_above" : 256}}}
在之后的查询中使用 serviceName 是将 serviceName 作为 text 类型查询,而使用 serviceName.keyword 则是将 serviceName 作为 keyword 类型查询。前者会对查询内容做分词处理之后再匹配,而后者则是直接对查询结果做精确匹配。
3.es 的 trem query 做的是精确匹配而不是分词查询,因此对 text 类型的字段做 term 查询将是查不到结果的(除非字段本身经过分词器处理后不变,未被转换或分词)。此时,必须使用 serviceName.keyword 来对 serviceName 字段以 keyword 类型进行精确匹配。
GET logdata-log-center-2021.05.06/_search{"query": {"terms": {"serviceName.keyword": ["log-center-user-portal","log-center-collect-manage"]}}}
Java API
@Testpublic void test() throws IOException {//构建查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //termQuery只能匹配一个值,第一个入参为字段名称,第二个参数为传入的值,相当于sql中的=//searchSourceBuilder.query(QueryBuilders.termQuery("serviceName.keyword", "log-center-user-portal-web")); //termsQuery可以一次性匹配多个值,相当于sql中的insearchSourceBuilder.query(QueryBuilders.termsQuery("serviceName.keyword", "log-center-user-portal-web", "log-center-collect-manage")); //构建查询请求对象,入参为索引SearchRequest searchRequest = new SearchRequest("log-web-up-log-center-2021.10.30"); //向搜索请求对象中配置搜索源searchRequest.source(searchSourceBuilder); // 执行搜索,向ES发起http请求SearchResponse response = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT); if (RestStatus.OK.equals(response.status())) {long total = response.getHits().getTotalHits().value; //检索到符合条件的总数SearchHit[] hits = response.getHits().getHits(); //未指定size,默认查询的是10条for (SearchHit hit : hits) {String index = hit.getIndex(); //索引名称String id = hit.getId(); //文档idJSONObject jsonObject = JSON.parseObject(hit.getSourceAsString(), JSONObject.class); //文档内容System.out.println(jsonObject); }}}
2.3 WildcardQuery
es的 wildcard query 做的是模糊匹配查询,类似 sql 中的 like,而 value 值前后的 “*” 号类似与 sql 中的 ”%“ 。
GET logdata-log-center-2021.05.06/_search{"query": {"wildcard": {"serviceName.keyword": {"value": "*user-portal*"}}}}
Java API
searchSourceBuilder.query(QueryBuilders.wildcardQuery("serviceName.keyword", "*" + "user-portal" + "*"));
2.4 RangeQuery
es 的 range query 做的是范围查询,相当于 sql 中的 between … and …
GET log-web-up-log-center-2021.10.30/_search{"query": {"range": {"timestamp": {"gte": "2021-10-30 15:00:00","lte": "2021-10-30 16:00:00","format": "yyyy-MM-dd HH:mm:ss||yyyy-MM-dd HH:mm:ss.SSS"}}}}
Java API
searchSourceBuilder.query(QueryBuilders.rangeQuery("timestamp").gte("2021-10-30 15:00:00") //起始值.lte("2021-10-30 16:00:00")//结束值.format("yyyy-MM-dd HH:mm:ss||yyyy-MM-dd HH:mm:ss.SSS")); //可以指定多个格式化标准,使用||隔开
2.5 MatchQuery
es的 match query 做的是全文检索,会对关键字进行分词后匹配词条。
GET log-web-up-log-center-2021.10.30/_search{"query": {"match": {"orgName": {"query": "有限公司"}}}}
query:搜索的关键字,对于英文关键字如果有多个单词则中间要用半角逗号分隔,而对于中文关键字中间可以用逗号分隔也可以不用。
Java API
//全文检索,支持分词匹配searchSourceBuilder.query(QueryBuilders.matchQuery("orgName", "有限公司");
2.6 MultiMatchQuery
上面的 MatchQuery 有一个短板,假如用户输入了某关键字,我们在检索的时候不知道具体是哪一个字段,这时我们用什么都不合适,而 MultiMatchQuery 的出现解决了这个问题,他可以通过 fields 属性来设置多个域联合查找,具体用法如下
GET log-web-up-log-center-2021.10.30/_search{"query": {"multi_match": {"query": "user-portal","fields": ["serviceName", "systemName"]}}}
Java API
//全文检索,支持分词匹配,支持多字段检索searchSourceBuilder.query(QueryBuilders.multiMatchQuery("user-portal", "serviceName", "systemName", "description"));
2.7 ExistsQuery
es的 exists query 做的是检索某个字段存在的数据,即不为 null 的数据。其中指定的 field 可以是一个具体的字段,也可以是一个 json 结构。
GET logdata-log-center-2021.05.06/_search{"query": {"exists": {"field": "networkLogDetailInfo"}}}
Java API
//查询networkLogDetailInfo不为null的数据searchSourceBuilder.query(QueryBuilders.existsQuery("networkLogDetailInfo"));
2.8 BoolQuery
es的 bool query 做的是将多个查询组合起来去检索数据,主要的组合参数有 must、should、mustNot 等。
must
:数据必须匹配 must 所包含的查询条件,相当于 ”AND“should
:数据匹配 should 包含的一个或多个查询条件,相当于 ”OR“mustNot
:数据必须不匹配 mustNot 所包含的查询条件,相当于 ”NOT“
GET logdata-log-center-2021.05.06/_search{"query": {"bool": {"must": [{"exists": {"field": "networkLogDetailInfo"}},{"range": {"timestamp": {"gte": "2021-05-05 00:00:00","lte": "2021-05-07 00:00:00","format": "yyyy-MM-dd HH:mm:ss||yyyy-MM-dd HH:mm:ss.SSS"}}}],"must_not": [{"exists": {"field": "serviceLogDetailInfo"}}]}}}
Java API
@Testpublic void test() throws IOException {//构建查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //构建bool类型查询器BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery(); //使用must连接,相当于and,构建第一个查询条件existsQuery必须包含此字段boolQueryBuilder.must(QueryBuilders.existsQuery("networkLogDetailInfo")); //使用must连接第二个条件,rangeQuery范围查找,相当于between...and...boolQueryBuilder.must(QueryBuilders.rangeQuery("timestamp").from("2021-05-05 00:00:00") //起始值.to("2021-05-07 00:00:00")//结束值.includeLower(true)//是否等于起始值.includeUpper(false)//是否等于结束值.format("yyyy-MM-dd HH:mm:ss||yyyy-MM-dd HH:mm:ss.SSS")); //格式化时间//使用mustNot连接第三个条件boolQueryBuilder.mustNot(QueryBuilders.existsQuery("serviceLogDetailInfo")); searchSourceBuilder.query(boolQueryBuilder); //构建查询请求对象,入参为索引SearchRequest searchRequest = new SearchRequest("logdata-log-center-2021.05.06"); //向搜索请求对象中配置搜索源searchRequest.source(searchSourceBuilder); // 执行搜索,向ES发起http请求SearchResponse response = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT); if (RestStatus.OK.equals(response.status())) {long total = response.getHits().getTotalHits().value; //检索到符合条件的总数SearchHit[] hits = response.getHits().getHits(); for (SearchHit hit : hits) {String index = hit.getIndex(); //索引名称String id = hit.getId(); //文档idJSONObject jsonObject = JSON.parseObject(hit.getSourceAsString(), JSONObject.class); //文档内容System.out.println(jsonObject); }}}
2.9 排序
es 使用 sort 进行排序,可以多个字段联合排序。
GET logdata-log-center-2021.05.06/_search{"query": {"bool": {"must_not": [{"exists": {"field": "serviceLogDetailInfo"}}]}},"sort": [{"serviceName.keyword": {"order": "asc"},"timestamp": {"order": "desc"}}]}
先按照第一个字段排序,第一个字段相同时按照第二个字段排序。
Java API
//升序searchSourceBuilder.sort("serviceName.keyword", SortOrder.ASC); //降序searchSourceBuilder.sort("timestamp", SortOrder.DESC);
2.10 结果字段过滤
检索数据,有时只需要其中的几个字段,es 也支持对结果集进行字段筛选过滤。字段可以使用 “*” 进行模糊匹配。
GET logdata-log-center-2021.05.06/_search{"_source": {"includes": ["messageId", "system*", "service*", "timestamp"],"excludes": []}}
Java API
//筛选字段,第一个参数为需要的字段,第二个参数为不需要的字段searchSourceBuilder.fetchSource(new String[] {"messageId", "system*", "service*", "timestamp"}, new String[] {});
2.11 分页
es 的分页方式有三种:from+ size、scroll、search_after, 默认采用的分页方式是 from+ size 的形式。
2.11.1 from+ size
GET logdata-log-center-2021.05.06/_search{"from": 0, "size": 2, "query": {"exists": {"field": "networkLogDetailInfo"}},"_source": {"includes": ["messageId", "system*", "service*", "timestamp"],"excludes": []}}
文章图片
通过查询结果可以发现,我们设置了分页参数之后, hits.total 返回的是数据总数7149,而按照分页规则,我们设置的size=2,因此 hits.hits 里面只有两条数据。
Java API
@Testpublic void test() throws IOException {//构建查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //查询条件searchSourceBuilder.query(QueryBuilders.existsQuery("networkLogDetailInfo")); int page = 1; // 页码int size = 2; // 每页显示的条数int index = (page - 1) * size; searchSourceBuilder.from(index); //设置查询起始位置searchSourceBuilder.size(size); //结果集返回的数据条数//筛选字段,第一个参数为需要的字段,第二个参数为不需要的字段searchSourceBuilder.fetchSource(new String[] {"messageId", "system*", "service*", "timestamp"}, new String[] {}); //构建查询请求对象,入参为索引SearchRequest searchRequest = new SearchRequest("logdata-log-center-2021.05.06"); //向搜索请求对象中配置搜索源searchRequest.source(searchSourceBuilder); // 执行搜索,向ES发起http请求SearchResponse response = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT); if (RestStatus.OK.equals(response.status())) {long total = response.getHits().getTotalHits().value; //检索到符合条件的总数SearchHit[] hits = response.getHits().getHits(); //未指定size,默认查询的是10条for (SearchHit hit : hits) {String index = hit.getIndex(); //索引名称String id = hit.getId(); //文档idJSONObject jsonObject = JSON.parseObject(hit.getSourceAsString(), JSONObject.class); //文档内容System.out.println(jsonObject); }}}
2.11.2 scroll
一种可满足深度分页的方式,es 提供了 scroll 的方式进行分页读取。原理上是对某次查询生成一个游标 scroll_id , 后续的查询只需要根据这个游标去取数据,每次只能拿到下一页的数据,直到结果集中返回的 hits 字段为空,就表示遍历结束。这里scroll=1m是scroll_id的有效期,表示1分钟,过期后会被es自动清理,每次查询会更新此值。
GET logdata-log-center-2021.05.06/_search?scroll=1m{"size": 2, "query": {"exists": {"field": "networkLogDetailInfo"}},"_source": {"includes": ["messageId", "system*", "service*", "timestamp"],"excludes": []}}
文章图片
后续的查询中查询条件不需要指定,只需要携带 scroll_id 即可它会按照首次查询条件进行分页展示,下一次查询(两种方式):
POST /_search/scroll{"scroll": "1m","scroll_id": "FGluY2x1ZGVfY29udGV4dF91dWlkDXF1ZXJ5QW5kRmV0Y2gBFFp0bGhXbjBCQU55Q3EtSDcxaWF4AAAAAACF-OYWV0liWUNLUHVTN09DS1ZtUl9SSHhVdw=="}
GET /_search/scroll?scroll=1m&scroll_id=FGluY2x1ZGVfY29udGV4dF91dWlkDXF1ZXJ5QW5kRmV0Y2gBFFp0bGhXbjBCQU55Q3EtSDcxaWF4AAAAAACF-OYWV0liWUNLUHVTN09DS1ZtUl9SSHhVdw==
Java API
public void testScroll(String scrollId) throws IOException {//查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //每页显示2条searchSourceBuilder.size(2); //查询条件searchSourceBuilder.query(QueryBuilders.existsQuery("networkLogDetailInfo")); //筛选字段,第一个参数为需要的字段,第二个参数为不需要的字段searchSourceBuilder.fetchSource(new String[] {"messageId", "system*", "service*", "timestamp"}, new String[] {}); SearchRequest request = new SearchRequest("logdata-log-center-2021.05.06"); request.source(searchSourceBuilder); Scroll scroll = new Scroll(TimeValue.timeValueMinutes(1L)); request.scroll(scroll); //滚动翻页SearchResponse response; if (!StringUtils.isBlank(scrollId)) {//Scroll查询SearchScrollRequest scrollRequest = new SearchScrollRequest(scrollId); scrollRequest.scroll(scroll); response = restHighLevelClient.scroll(scrollRequest, RequestOptions.DEFAULT); } else {//首次查询使用普通查询response = restHighLevelClient.search(request, RequestOptions.DEFAULT); }//更新scrollIdscrollId = response.getScrollId(); System.out.println(scrollId); if (RestStatus.OK.equals(response.status())) {//设置查询总量SearchHit[] hits = response.getHits().getHits(); for (SearchHit hit : hits) {String index = hit.getIndex(); String id = hit.getId(); JSONObject jsonObject = JSON.parseObject(hit.getSourceAsString(), JSONObject.class); System.out.println(jsonObject); }}}
2.11.3 search_after
search_after 是 ES5.0 及之后版本提供的新特性,search_after查询时需要指定sort排序字段,可以指定多个排序字段,后续查询有点类似 scroll ,但是和 scroll 又不一样,它提供一个活动的游标,通过上一次查询的最后一条数据的来进行下一次查询。 这里需要说明一下,使用search_after查询需要将from设置为0或-1,当然你也可以不写
第一次查询:
POST logdata-log-center-2021.05.06/_search{"size": 2, "query": {"exists": {"field": "networkLogDetailInfo"}},"_source": {"includes": ["messageId", "system*", "service*", "timestamp"],"excludes": []},"sort": [{"timestamp": {"order": "desc"}}]}
查询结果:可以看到每一条数据都有一个sort部分,而下一页的查询需要本次查询结果最后一条的sort值作为游标,实现分页查询
文章图片
第二次查询:
POST logdata-log-center-2021.05.06/_search{"search_after": [1620374316433],"size": 2, "query": {"exists": {"field": "networkLogDetailInfo"}},"_source": {"includes": ["messageId", "system*", "service*", "timestamp"],"excludes": []},"sort": [{"timestamp": {"order": "desc"}}]}
Java API
public void testSearchAfter(Object[] values) throws IOException {//查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); searchSourceBuilder.size(2); searchSourceBuilder.from(0); //searchAfter需要将from设置为0或-1,当然也可以不写//查询条件searchSourceBuilder.query(QueryBuilders.existsQuery("networkLogDetailInfo")); //筛选字段,第一个参数为需要的字段,第二个参数为不需要的字段searchSourceBuilder.fetchSource(new String[] {"messageId", "system*", "service*", "timestamp"}, new String[] {}); //以时间戳排序searchSourceBuilder.sort("timestamp", SortOrder.DESC); if (values != null)searchSourceBuilder.searchAfter(values); SearchRequest request = new SearchRequest("logdata-log-center-2021.05.06"); request.source(searchSourceBuilder); SearchResponse response = restHighLevelClient.search(request, RequestOptions.DEFAULT); if (RestStatus.OK.equals(response.status())) {//设置查询总量SearchHit[] hits = response.getHits().getHits(); for(int i = 0; i < hits.length; i++) {String index = hits[i].getIndex(); String id = hits[i].getId(); JSONObject jsonObject = JSON.parseObject(hits[i].getSourceAsString(), JSONObject.class); System.out.println(jsonObject); if (i == hits.length-1) {//最后一条数据的sortValue作为下一次查询的游标值values = hits[i].getSortValues(); System.out.println(Arrays.toString(values)); }}}}
2.11.4 三种分页方式特点
from+size
比较适合浅分页模式,在深度分页的情况下,这种使用方式效率是非常低的,随着分页页码的不断增大,查询的效率会直线下降。比如from = 5000, size=20, es 需要在各个分片上匹配排序并得到5000*20 条有效数据,然后在结果集中取最后20条。除了效率上的问题,还有一个无法解决的问题是,es 目前支持最大的 skip 值是 max_result_window ,默认为 10000 。也就是当 from + size > max_result_window 时,es 将返回错误。scroll
是一种滚屏形式的分页检索,满足深度分页的场景。查询的时候生成一个游标 scroll_id,有效期内每次返回的值是一样的,后续的查询只需要根据这个游标去取数据即可。scroll查询是很耗性能的方式,scroll_id 的生成可以理解为建立了一个临时的历史快照, 系统会耗费大量的资源来保存一份当前查询结果集映像,并且会占用文件描述符,在此之后的增删改查等操作不会影响到这个快照的结果,因此不建议在实时查询中运用。这种方式往往用于非实时处理大量数据的情况,比如要进行数据迁移或者索引变更之类的。search_after
适用于深度分页+ 排序,分页是根据上一页最后一条数据来定位下一页的位置,所以无法跳页请求,同时在分页请求的过程中,如果有索引数据的增删改,这些变更也会实时的反映到游标上。在选择search_after的排序字段时尽量使用比如文档的id或者时间戳等具有唯一性的字段。search_after 相比 from+size 的浅分页以及 scroll 滚屏查询会有很大的性能提升。
2.22 聚合
es 的 aggs 对数据进行聚合查询统计,查询方式如下:
## 统计各系统一个月的日志采集数量POST log*/_search{"size": 0,"query": {"range": {"timestamp": {"gte": "2021-10-24 00:00:00","lte": "2021-11-24 00:00:00","format": "yyyy-MM-dd HH:mm:ss"}} }, "aggs": {"allLog": {"terms": {"field": "systemName.keyword","size": 10}} }}
文章图片
Java API
@Testpublic void test() throws IOException {//按照systemName字段聚合统计各个系统的日志数量TermsAggregationBuilder bySystemName = AggregationBuilders.terms("allLog").field("systemName.keyword"); RangeQueryBuilder timestamp = QueryBuilders.rangeQuery("timestamp").gte("2021-10-24 00:00:00").lte("2021-11-24 00:00:00").format("yyyy-MM-dd HH:mm:ss"); //查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //配置聚合条件searchSourceBuilder.aggregation(bySystemName); //配置查询条件searchSourceBuilder.query(timestamp); //设置查询结果不返回,只返回聚合结果searchSourceBuilder.size(0); //创建查询请求对象,将查询条件配置到其中SearchRequest request = new SearchRequest("log*"); request.source(searchSourceBuilder); // 执行搜索,向ES发起http请求SearchResponse response = restHighLevelClient.search(request, RequestOptions.DEFAULT); Aggregations aggregations = response.getAggregations(); if (aggregations != null) {Terms terms = aggregations.get("allLog"); //解析桶for (Terms.Bucket bucket : terms.getBuckets()) {System.out.print("系统名称:" + bucket.getKeyAsString()); System.out.println("\t总日志数量:" + bucket.getDocCount()); }}}
多层嵌套聚合
## 统计各个系统的总日志数量,按系统统计各种类型日志数量POST log*/_search{ "size": 0, "query": {"range": {"timestamp": {"gte": "2021-10-24 00:00:00","lte": "2021-11-24 00:00:00","format": "yyyy-MM-dd HH:mm:ss"}} }, "aggs": {"allLog": {"terms": {"field": "systemName.keyword","size": 10},"aggs": {"errorLogNum": {"filter": {"terms": {"level.keyword": ["ERROR","FATAL"]}}},"dbLogNum": {"filter": {"exists": {"field": "dataLogDetailInfo"}}},"interfaceLogNum": {"filter": {"exists": {"field": "networkLogDetailInfo"}}},"serviceLogNum": {"filter": {"exists": {"field": "serviceLogDetailInfo"}}},"webLogNum": {"filter": {"exists": {"field": "browserModel"}}}}} }}
Java API
@Testpublic void test() throws IOException {//错误日志聚合条件FilterAggregationBuilder errorLogNum = AggregationBuilders.filter("errorLogNum", QueryBuilders.termsQuery("level.keyword", "ERROR", "FATAL")); //数据库日志聚合条件FilterAggregationBuilder dataLogNum = AggregationBuilders.filter("dbLogNum", QueryBuilders.existsQuery("dataLogDetailInfo")); //接口日志聚合条件FilterAggregationBuilder networkLogNum = AggregationBuilders.filter("interfaceLogNum", QueryBuilders.existsQuery("networkLogDetailInfo")); //应用日志聚合条件FilterAggregationBuilder serviceLogNum = AggregationBuilders.filter("serviceLogNum", QueryBuilders.existsQuery("serviceLogDetailInfo")); //前端日志聚合条件FilterAggregationBuilder webUpLogNum = AggregationBuilders.filter("webLogNum", QueryBuilders.existsQuery("browserModel")); //最外层聚合条件,第一次聚合的条件TermsAggregationBuilder bySystemName = AggregationBuilders.terms("allLog").field("systemName.keyword").size(10); //内部多个条件的子聚合,在系统聚合后的结果上二次聚合bySystemName.subAggregation(errorLogNum).subAggregation(dataLogNum).subAggregation(networkLogNum).subAggregation(serviceLogNum).subAggregation(webUpLogNum); RangeQueryBuilder timestamp = QueryBuilders.rangeQuery("timestamp").gte("2021-10-24 00:00:00").lte("2021-11-24 00:00:00").format("yyyy-MM-dd HH:mm:ss"); //查询源构建器SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); //配置聚合条件searchSourceBuilder.aggregation(bySystemName); //配置查询条件searchSourceBuilder.query(timestamp); //设置查询结果不返回,只返回聚合结果searchSourceBuilder.size(0); //创建查询请求对象,将查询条件配置到其中SearchRequest request = new SearchRequest("log*"); request.source(searchSourceBuilder); // 执行搜索,向ES发起http请求SearchResponse response = restHighLevelClient.search(request, RequestOptions.DEFAULT); Aggregations aggregations = response.getAggregations(); if (aggregations != null) {Terms terms = aggregations.get("allLog"); for (Terms.Bucket bucket : terms.getBuckets()) {ParsedFilter dbFilter = bucket.getAggregations().get("dbLogNum"); ParsedFilter serviceFilter = bucket.getAggregations().get("serviceLogNum"); ParsedFilter webFilter = bucket.getAggregations().get("webLogNum"); ParsedFilter interfaceFilter = bucket.getAggregations().get("interfaceLogNum"); ParsedFilter errorFilter = bucket.getAggregations().get("errorLogNum"); System.out.print("系统名称:" + bucket.getKeyAsString()); System.out.print("\t总日志:" + bucket.getDocCount()); System.out.print("\t数据库日志:" + dbFilter.getDocCount()); System.out.print("\t服务执行日志:" + serviceFilter.getDocCount()); System.out.print("\t前端操作日志:" + webFilter.getDocCount()); System.out.print("\t接口日志:" + interfaceFilter.getDocCount()); System.out.println("\t错误日志:" + errorFilter.getDocCount()); }}}
聚合查询还提供了许多查询规则,按时间date聚合、count聚合、avg聚合、sum聚合、min聚合、max聚合等等,这里就不一一列举了。
【SpringBoot整合ES高级查询方式】以上为个人经验,希望能给大家一个参考,也希望大家多多支持脚本之家。
推荐阅读
- SpringBoot详解如何实现读写分离
- Java开发之ssm三大框架整合
- @SpringBootApplication注解的理解
- SpringBootSpringApplicationRunListener 是干啥的
- SpringBoot整合Mybatis进行单元测试mapper和xml路径不一致出现的诡异问题
- SSM整合-Spring里面applicationContext.xml文件能不能改成其他文件名()
- 10个用于流量大网站的高级WordPress托管
- win|win server 部署前后端分离项目(springboot + vue + android)
- SpringBoot org.apache.ibatis.binding.BindingException: Invalid bound statement (not found): mapper.x
- 改进SEO的10个高级网站抓取工具