文章目录
Elastic Stack核心产品包括Elasticsearch、Kibana、Beats 和 Logstash，可用于日志、APM、安全等各种用例。
Elasticsearch
Elasticsearch是一个分布式、RESTful 风格的搜索和数据分析引擎；作为 Elastic Stack 的核心，它集中存储数据。
Node与Cluster
Elasticsearch本质上是一个分布式数据库，允许多台服务器协同工作，每台服务器可以运行多个 Elastic 实例。单个实例称为一个节点（node）；一组节点构成一个集群（cluster）。
Index
Elasticsearch会索引所有字段，经过处理后写入一个反向索引（Inverted Index）；查找数据的时候，直接查找该索引
Elasticsearch数据管理的顶层容器叫做 Index（索引）（类似数据库）。Index 的名字必须小写。
Document
Index 里面单条的记录称为 Document（文档）。Document使用 JSON 格式表示：
- _index：是索引名；
- _type：文档类型（以后会弃用）；
- _id：文档唯一标识，操作文档时就需要使用此唯一标识；
- _version：文档版本，每更新一次会加一；
{
"_index": "test-07.02",
"_type": "_doc",
"_id": "Hoaaa3oBfSQ12-KDj4pW",
"_version": 1,
"_score": 0,
"fields": {
"number": [
1175
],
"name": [
"Name 2usfD"
],
"name.keyword": [
"Name 2usfD"
],
"age": [
7
]
}
}
RestHighLevelClient
ES提供了两个JAVA REST client版本：
-
Java Low Level REST Client: 低级别的REST客户端，通过http与集群交互，用户需自己编组请求JSON串，及解析响应JSON串。兼容所有ES版本。 -
Java High Level REST Client: 高级别的REST客户端，基于低级别的REST客户端，增加了编组请求JSON串、解析响应JSON串等相关api。使用的版本需要保持和ES服务端的版本一致，否则会有版本问题。
一般推荐使用High Level版本；不同版本下接口会有些不同，本文以7.13.2为例。
依赖包
为了使用ES客户端，需要在pom.xml中增加依赖（ES的Client版本要与ES版本统一，因此要在Properties中设定所使用的版本）：
<properties>
<java.version>1.8</java.version>
<es.version>7.13.2</es.version>
</properties>
<dependencies>
<!-- https://mvnrepository.com/artifact/org.elasticsearch/elasticsearch -->
<dependency>
<groupId>org.elasticsearch</groupId>
<artifactId>elasticsearch</artifactId>
<version>7.13.2</version>
</dependency>
<!-- https://mvnrepository.com/artifact/org.elasticsearch.client/elasticsearch-rest-high-level-client -->
<dependency>
<groupId>org.elasticsearch.client</groupId>
<artifactId>elasticsearch-rest-high-level-client</artifactId>
<version>7.13.2</version>
<exclusions>
<exclusion>
<groupId>org.elasticsearch.client</groupId>
<artifactId>elasticsearch-rest-client</artifactId>
</exclusion>
</exclusions>
</dependency>
<!-- https://mvnrepository.com/artifact/org.elasticsearch.client/elasticsearch-rest-client -->
<dependency>
<groupId>org.elasticsearch.client</groupId>
<artifactId>elasticsearch-rest-client</artifactId>
<version>7.13.2</version>
</dependency>
</dependencies>
client对象
在对ES进行操作前，需要先创建与elasticsearch服务进行连接的RestHighLevelClient对象：
RestHighLevelClient getClient() {
RestHighLevelClient client = new RestHighLevelClient(
RestClient.builder(
new HttpHost("127.0.0.1", 9200, "http")
)
);
在client对象使用完后，需通过close()及时关闭
Index
Index（索引）是文档(Document)的容器，是一类文档的集合。其相关操作，主要封装client..indices()中。
创建索引
通过create来创建索引；通过返回值中的isAcknowledged判断是否成功：
public String createIndex(String index) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
CreateIndexRequest createReq = new CreateIndexRequest(index);
CreateIndexResponse resp = rhlClient.indices().create(createReq, RequestOptions.DEFAULT);
String result = resp.isAcknowledged() ? "OK" : "Fail";
System.out.println(index + " create: " + result);
} catch (Exception ex) {
System.out.println(index + " create fail: " + ex);
}
return "fail";
}
判断索引是否存在
通过exists来判断索引是否存在：
public boolean isIndexExists(String index) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
GetIndexRequest getReq = new GetIndexRequest(index);
boolean bExists = rhlClient.indices().exists(getReq, RequestOptions.DEFAULT);
System.out.println(index + " exists: " + bExists);
} catch (Exception ex) {
System.out.println(index + " check fail: " + ex);
}
return false;
}
删除索引
通过delete来删除索引；通过返回值中的isAcknowledged判断是否成功：
public String deleteIndex(String index) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
DeleteIndexRequest delReq = new DeleteIndexRequest(index);
AcknowledgedResponse resp = rhlClient.indices().delete(delReq, RequestOptions.DEFAULT);
String result = resp.isAcknowledged() ? "OK" : "Fail";
System.out.println(index + " delete: " + result);
} catch (Exception ex) {
System.out.println(index + " delete fail: " + ex);
}
return "fail";
}
获取索引
通过别名Alias来获取索引（通过*来匹配）；要匹配索引需要通过GetAliasesRequest的**indices(pattern)**来设定。返回是一个map映射：
public Set<String> listIndex(String prefix) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
GetAliasesRequest getReq = new GetAliasesRequest();
getReq.indices(prefix + "*");
GetAliasesResponse resp = rhlClient.indices().getAlias(getReq, RequestOptions.DEFAULT);
Map<String, Set<AliasMetadata>> mapAlias = resp.getAliases();
Set<String> result = mapAlias.keySet();
System.out.println("list index of " + prefix + ": " + result);
} catch (Exception ex) {
System.out.println(prefix + " list fail: " + ex);
}
return new HashSet<>();
}
Document
Index 里面单条的记录称为Document（文档）。
添加记录
通过index来添加索引；内容可通过Map<String, Object>设定，返回值为添加记录的基本信息（id等）：
public String addDocument(String index, Map<String, Object> data) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
IndexRequest request = new IndexRequest(index);
request.source(data);
IndexResponse resp = rhlClient.index(request, RequestOptions.DEFAULT);
String result = resp.toString();
System.out.println(result);
return result;
} catch (Exception ex) {
System.out.println(index + " add doc fail: " + ex);
return ex.getMessage();
}
}
判断记录是否存在
通过exists判断记录是否存在：
public boolean isDocExists(String index, String id) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
GetRequest request = new GetRequest(index);
request.id(id);
boolean bExists = rhlClient.exists(request, RequestOptions.DEFAULT);
System.out.println(index + " exists: " + bExists);
return bExists;
} catch (Exception ex) {
System.out.println(index + " add doc fail: " + ex);
}
return false;
}
删除记录
通过delete来删除记录（根据记录id）：
public String delDocument(String index, String id) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
DeleteRequest request = new DeleteRequest(index);
request.id(id);
DeleteResponse resp = rhlClient.delete(request, RequestOptions.DEFAULT);
System.out.println(index + " exists: " + resp.toString());
return resp.toString();
} catch (Exception ex) {
System.out.println(index + " add doc fail: " + ex);
return ex.toString();
}
}
更新记录
通过update来更新记录；记录通过id标识，新的内容通过Map<String, Object>设定：
public String updateDocument(String index, String id, Map<String, Object> data) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
UpdateRequest request = new UpdateRequest(index, id);
request.doc(data);
UpdateResponse resp = rhlClient.update(request, RequestOptions.DEFAULT);
String result = resp.toString();
System.out.println(result);
return result;
} catch (Exception ex) {
System.out.println(ex);
return ex.getMessage();
}
}
获取记录
通过get可获取记录的详细信息：
public String getDocument(String index, String id){
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
GetRequest request = new GetRequest(index);
request.id(id);
GetResponse resp = rhlClient.get(request, RequestOptions.DEFAULT);
System.out.println(resp.getIndex() + " # " + resp.getVersion() + " # " + resp.getId());
System.out.println(resp.getSourceAsString());
return resp.getSourceAsString();
} catch (Exception ex) {
System.out.println(index + " add doc fail: " + ex);
}
return "Fail";
}
获取记录条数
通过count可获取记录的条数：
public long getDocCount(String index) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
CountRequest request = new CountRequest(index);
CountResponse resp = rhlClient.count(request, RequestOptions.DEFAULT);
System.out.println(index + " document count: " + resp.getCount());
return resp.getCount();
} catch (Exception ex) {
System.out.println(index + " document count fail: " + ex);
return 0L;
}
}
获取记录Id
通过search来获取所需记录（通过fetchSource(new String[]{""}, null)来设定只获取信息，不获取内容）：
public List<String> listDocId(String index, int count) {
try (RestHighLevelClient rhlClient = ESClient.getClient()) {
SearchRequest request = new SearchRequest(index);
SearchSourceBuilder sourceBuilder = SearchSourceBuilder.searchSource();
sourceBuilder.from(0);
sourceBuilder.size(count);
sourceBuilder.fetchSource(new String[]{""}, null);
request.source(sourceBuilder);
SearchResponse response = rhlClient.search(request, RequestOptions.DEFAULT);
SearchHits gotHits = response.getHits();
List<String> lstId = new ArrayList<>(gotHits.getHits().length);
for (SearchHit doc : gotHits) {
lstId.add(doc.getId());
}
System.out.println(index + " documents: " + lstId);
return lstId;
} catch (Exception ex) {
System.out.println(index + " update doc fail: " + ex);
return new ArrayList<>(0);
}
}
1万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
