一、ES基本概念
1. 节点：Elastic是一个分布式数据库，每个数据库实例是一个节点Node，一台服务器上可以有多个Node，可以多台服务器协同工作
2. 集群：一组节点Node构成一个集群Cluster
3. 文档：存储数据的基本单元称为文档Document，使用json表示
4. 索引：索引Index是一类文档的集合，ES会索引所有字段，经过处理后写入一个反向索引
5. 类型：一个索引的逻辑分区叫类型Type
6. 分词器：把text类型的数据进行分词，分词后按照分词建立索引，可以按照分词进行查询；默认分词器为ik，可以设置为中文分词器ik_max_word
二、跟普通数据库mysql在数据结构上的区别
mysql：库->表->行->字段
ES：索引->类型->文档->字段
三、数据类型
1.ES在5.*版本之后，把string字段设置为了过时字段，引入text，keyword字段
2.text会自动分词，keyword不分词，因为全部字段分词，会浪费空间
3.如果不在mapping中定义字段类型，java的String类型会自动映射成text & keyword两种
4.如果ES中已经有数据，不支持修改字段类型
a.可以删除原来索引，重新建立
b.新建一个索引，把老数据迁移到新索引上面
c.如果数据比较多或者比较重要，可以新增一个字段
三、增删改查API
1.新增或者更新，/index/type/id已存在就会更新
put /index/type/id
{
"field1":"value1",
"field2":"value2"
}
2.删除
delete /index/type/id
3.查询
term：精确匹配一个字段
match：模糊查询或者分词查询一个字段
wildcard：使用通配符进行查询
四、java使用RestHighLevelClient操作ES
1.引入jar包
<dependency>
<groupId>org.elasticsearch.client</groupId>
<artifactId>elasticsearch-rest-high-level-client</artifactId>
<version>6.7.1</version>
</dependency>
2.初始化
@Configuration public class TaskEsTopicConfig { private static final int HTTP_PORT = 9200; private static final String IP = "aaa"; private static final String SecurityUser = "aaa"; private static final String SecurityPassword = "aaa"; private static CredentialsProvider init() { CredentialsProvider credentialsProvider = new BasicCredentialsProvider(); credentialsProvider.setCredentials(AuthScope.ANY, new UsernamePasswordCredentials(SecurityUser, SecurityPassword)); return credentialsProvider; } @Bean public RestClientBuilder restClientBuilder() { //配置权限验证 RestClientBuilder restClientBuilder = RestClient.builder( new HttpHost(IP, HTTP_PORT, "http")) .setHttpClientConfigCallback(httpClientBuilder -> httpClientBuilder.setDefaultCredentialsProvider(init())); return restClientBuilder; } @Bean(name = "highLevelClient") public RestHighLevelClient highLevelClient(@Autowired RestClientBuilder restClientBuilder) { return new RestHighLevelClient(restClientBuilder); } }
3.写数据
IndexRequest request = new IndexRequest(); request.index(index).type(type).source(JSON.toJSONString(object), XContentType.JSON); bulkRequest.add(request);
// 保存数据到ES
BulkResponse bulkResponse = highLevelClient.bulk(bulkRequest, RequestOptions.DEFAULT);
4.读数据 BoolQueryBuilder的should/must，对应mysql的or/and
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder(); // 分页参数 searchSourceBuilder.from(1); searchSourceBuilder.size(10); searchSourceBuilder.sort("_id", SortOrder.DESC); BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery() // 精确匹配 .must(QueryBuilders.termQuery("field1","value1") // 关键字模糊+分词匹配
.must(QueryBuilders.multiMatchQuery("keyword","field1","field2","field3","field4","field5")); searchSourceBuilder.query(boolQueryBuilder); SearchRequest searchRequest = new SearchRequest(); searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = highLevelClient.search(searchRequest, RequestOptions.DEFAULT);
if (Objects.nonNull(searchResponse)) {
total = searchResponse.getHits().totalHits;
SearchHits searchHits = searchResponse.getHits();
for (int i = 0; i < searchHits.getHits().length; i++) {
String str = searchHits.getHits()[i].getSourceAsString();
}
}
5.删除数据
DeleteRequest deleteRequest = new DeleteRequest(); deleteRequest.index(index).type(type).id(id); try { highLevelClient.delete(deleteRequest, RequestOptions.DEFAULT); } catch (Exception ex) { log.error("删除ES中的数据异常", ex); }
参考：
https://www.cnblogs.com/javallh/p/13790486.html
https://www.cnblogs.com/dreamroute/p/8484457.html
https://www.cnblogs.com/zwt1990/p/7737747.html