前言

前面历经33篇内容的讲解，与ES的请求操作都是在Kibana平台上用Restful请求完成的，一直没发布Java或python的客户端代码，Restful才是运用、理解ES核心功能最直接的表达方式，但实际项目中肯定是以Java/python来完成ES请求的发起与数据处理的，前面理解了ES的核心功能，后面Java API的使用将会非常简单，剩余的未覆盖的功能API，自行查阅文档即可。

概要

本篇讲解Elasticsearch的客户端API开发的一些示例，以Java语言为主，介绍一些最常用，最核心的API。

代码示例

引入依赖

我们以maven项目为例，添加项目依赖

org.elasticsearch

elasticsearch

6.3.1

org.elasticsearch.client

transport

6.3.1

log4j

log4j

1.2.17

org.apache.logging.log4j

log4j-core

2.12.1

建立ES连接创建Settings对象，指定集群名称

创建TransportClient对象，手动指定IP、端口即可

Settings settings = Settings.builder().put("cluster.name", "elasticsearch").build();

TransportClient client = new PreBuiltTransportClient(settings).addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("localhost"), 9300));

如果集群的节点数比较多，为每个node分别指定IP、Port可行性不高，我们可以使用集群节点自动探查的功能，代码如下：

// 将client.transport.sniff设置为true即可打开集群节点自动探查功能

Settings settings = Settings.builder().put("client.transport.sniff", true)..put("cluster.name", "elasticsearch").build();

// 只需要指定一个node就行

TransportClient client = new PreBuiltTransportClient(settings);

transport.addTransportAddress(new TransportAddress(InetAddress.getByName("192.168.17.137"), 9300));

基本CRUD

最基本的CRUD代码，可以当作入门demo来写：

/**

* 创建员工信息(创建一个document)

* @param client

*/

private static void createEmployee(TransportClient client) throws Exception {

IndexResponse response = client.prepareIndex("company", "employee", "1")

.setSource(XContentFactory.jsonBuilder()

.startObject()

.field("name", "jack")

.field("age", 27)

.field("position", "technique")

.field("country", "china")

.field("join_date", "2017-01-01")

.field("salary", 10000)

.endObject())

.get();

System.out.println(response.getResult());

}

/**

* 获取员工信息

* @param client

* @throws Exception

*/

private static void getEmployee(TransportClient client) throws Exception {

GetResponse response = client.prepareGet("company", "employee", "1").get();

System.out.println(response.getSourceAsString());

}

/**

* 修改员工信息

* @param client

* @throws Exception

*/

private static void updateEmployee(TransportClient client) throws Exception {

UpdateResponse response = client.prepareUpdate("company", "employee", "1")

.setDoc(XContentFactory.jsonBuilder()

.startObject()

.field("position", "technique manager")

.endObject())

.get();

System.out.println(response.getResult());

}

/**

* 删除 员工信息

* @param client

* @throws Exception

*/

private static void deleteEmployee(TransportClient client) throws Exception {

DeleteResponse response = client.prepareDelete("company", "employee", "1").get();

System.out.println(response.getResult());

}

搜索

我们之前使用Restful的搜索，现在改用java实现，原有的Restful示例如下：

GET /company/employee/_search

{

"query": {

"bool": {

"must": [

{

"match": {

"position": "technique"

}

}

],

"filter": {

"range": {

"age": {

"gte": 30,

"lte": 40

}

}

}

}

},

"from": 0,

"size": 1

}

等同于这样的Java代码：

SearchResponse response = client.prepareSearch("company")

.setTypes("employee")

.setQuery(QueryBuilders.termQuery("position", "technique")) // Query

.setPostFilter(QueryBuilders.rangeQuery("age").from(30).to(40)) // Filter

.setFrom(0).setSize(60)

.get();

聚合查询

聚合查询稍微麻烦一些，请求的封装和响应报文的解析，都是根据实际返回的结构来做的，例如下面的查询：

需求：按照country国家来进行分组

在每个country分组内，再按照入职年限进行分组

最后计算每个分组内的平均薪资

Restful的请求如下：

GET /company/employee/_search

{

"size": 0,

"aggs": {

"group_by_country": {

"terms": {

"field": "country"

},

"aggs": {

"group_by_join_date": {

"date_histogram": {

"field": "join_date",

"interval": "year"

},

"aggs": {

"avg_salary": {

"avg": {

"field": "salary"

}

}

}

}

}

}

}

}

用Java编写的请求如下：

SearchResponse sr = node.client().prepareSearch()

.addAggregation(

AggregationBuilders.terms("by_country").field("country")

.subAggregation(AggregationBuilders.dateHistogram("by_year")

.field("dateOfBirth")

.dateHistogramInterval(DateHistogramInterval.YEAR)

.subAggregation(AggregationBuilders.avg("avg_children").field("children"))

)

)

.execute().actionGet();

对响应的处理，则需要一层一层获取数据：

Map aggrMap = searchResponse.getAggregations().asMap();

StringTerms groupByCountry = (StringTerms) aggrMap.get("group_by_country");

Iterator groupByCountryBucketIterator = groupByCountry.getBuckets().iterator();

while(groupByCountryBucketIterator.hasNext()) {

Bucket groupByCountryBucket = groupByCountryBucketIterator.next();

System.out.println(groupByCountryBucket.getKey() + "\t" + groupByCountryBucket.getDocCount());

Histogram groupByJoinDate = (Histogram) groupByCountryBucket.getAggregations().asMap().get("group_by_join_date");

Iterator groupByJoinDateBucketIterator = groupByJoinDate.getBuckets().iterator();

while(groupByJoinDateBucketIterator.hasNext()) {

org.elasticsearch.search.aggregations.bucket.histogram.Histogram.Bucket groupByJoinDateBucket = groupByJoinDateBucketIterator.next();

System.out.println(groupByJoinDateBucket.getKey() + "\t" + groupByJoinDateBucket.getDocCount());

Avg avgSalary = (Avg) groupByJoinDateBucket.getAggregations().asMap().get("avg_salary");

System.out.println(avgSalary.getValue());

}

}

client.close();

}

upsert请求

private static void upsert(TransportClient transport) {

try {

IndexRequest index = new IndexRequest("book_shop", "books", "2").source(

XContentFactory.jsonBuilder().startObject()

.field("name", "mysql从入门到删库跑路")

.field("tags", "mysql")

.field("price", 32.8)

.endObject());

UpdateRequest update = new UpdateRequest("book_shop", "books", "2")

.doc(XContentFactory.jsonBuilder()

.startObject().field("price", 31.8)

.endObject())

.upsert(index);

UpdateResponse response = transport.update(update).get();

System.out.println(response.getVersion());

} catch (IOException e) {

e.printStackTrace();

} catch (InterruptedException e) {

e.printStackTrace();

} catch (ExecutionException e) {

e.printStackTrace();

}

}

mget请求

public static void mget(TransportClient transport) {

MultiGetResponse res = transport.prepareMultiGet()

.add("book_shop", "books", "1")

.add("book_shop", "books", "2")

.get();

for (MultiGetItemResponse item : res.getResponses()) {

System.out.println(item.getResponse());

}

}

bulk请求

public static void bulk(TransportClient transport) {

try {

BulkRequestBuilder bulk = transport.prepareBulk();

bulk.add(transport.prepareIndex("book_shop", "books", "3").setSource(

XContentFactory.jsonBuilder().startObject()

.field("name", "设计模式从入门到拷贝代码")

.field("tags", "设计模式")

.field("price", 55.9)

.endObject()));

bulk.add(transport.prepareIndex("book_shop", "books", "4").setSource(

XContentFactory.jsonBuilder().startObject()

.field("name", "架构设计从入门到google搜索")

.field("tags", "架构设计")

.field("price", 68.9)

.endObject()));

bulk.add(transport.prepareUpdate("book_shop", "books", "1").setDoc((XContentFactory.jsonBuilder()

.startObject().field("price", 32.8)

.endObject())));

BulkResponse bulkRes = bulk.get();

if (bulkRes.hasFailures()) {

System.out.println("Error...");

}

} catch (IOException e) {

e.printStackTrace();

}

}

scorll请求

public static void scorll(TransportClient client) {

SearchResponse bookShop = client.prepareSearch("book_shop").setScroll(new TimeValue(60000)).setSize(1).get();

int batchCnt = 0;

do {

// 循环读取scrollid信息，直到结果为空

for(SearchHit hit: bookShop.getHits().getHits()) {

System.out.println("batchCnt:" + ++batchCnt);

System.out.println(hit.getSourceAsString());

}

bookShop = client.prepareSearchScroll(bookShop.getScrollId()).setScroll(new TimeValue(60000)).execute().actionGet();

} while (bookShop.getHits().getHits().length != 0);

}

搜索模板

public static void searchTemplates(TransportClient client) {

Map params = new HashMap<>(10);

params.put("from",0);

params.put("size",10);

params.put("tags","java");

SearchTemplateResponse str = new SearchTemplateRequestBuilder(client)

.setScript("page_query_by_tags")

.setScriptType(ScriptType.STORED)

.setScriptParams(params)

.setRequest(new SearchRequest())

.get();

for(SearchHit hit:str.getResponse().getHits().getHits()) {

System.out.println(hit.getSourceAsString());

}

}

多条件组合查询

public static void otherSearch(TransportClient client) {

SearchResponse response1 = client.prepareSearch("book_shop").setQuery(QueryBuilders.termQuery("tags", "java")).get();

SearchResponse response2 = client.prepareSearch("book_shop").setQuery(QueryBuilders.multiMatchQuery("32.8", "price","tags")).get();

SearchResponse response3 = client.prepareSearch("book_shop").setQuery(QueryBuilders.commonTermsQuery("name", "入门")).get();

SearchResponse response4 = client.prepareSearch("book_shop").setQuery(QueryBuilders.prefixQuery("name", "java")).get();

System.out.println(response1.getHits().getHits()[0].getSourceAsString());

System.out.println(response2.getHits().getHits()[0].getSourceAsString());

System.out.println(response3.getHits().getHits()[0].getSourceAsString());

System.out.println(response4.getHits().getHits()[0].getSourceAsString());

// 多个条件组合

SearchResponse response5 = client.prepareSearch("book_shop").setQuery(QueryBuilders.boolQuery()

.must(QueryBuilders.termQuery("tags", "java"))

.mustNot(QueryBuilders.matchQuery("name", "跑路"))

.should(QueryBuilders.matchQuery("name", "入门"))

.filter(QueryBuilders.rangeQuery("price").gte(23).lte(55))).get();

System.out.println(response5.getHits().getHits()[0].getSourceAsString());

}

地理位置查询

public static void geo(TransportClient client) {

GeoBoundingBoxQueryBuilder query1 = QueryBuilders.geoBoundingBoxQuery("location").setCorners(23, 112, 21, 114);

List points = new ArrayList<>();

points.add(new GeoPoint(23,115));

points.add(new GeoPoint(25,113));

points.add(new GeoPoint(21,112));

GeoPolygonQueryBuilder query2 = QueryBuilders.geoPolygonQuery("location",points);

GeoDistanceQueryBuilder query3 = QueryBuilders.geoDistanceQuery("location").point(22.523375, 113.911231).distance(500, DistanceUnit.METERS);

SearchResponse response = client.prepareSearch("location").setQuery(query3).get();

for(SearchHit hit:response.getHits().getHits()) {

System.out.println(hit.getSourceAsString());

}

}

小结

专注Java高并发、分布式架构，更多技术干货分享与心得，请关注公众号：Java架构社区