为什么80%的码农都做不了架构师？>>> 
Elasticsearch 父子关系维护和检索案例分享，展示has_child和has_parent的基本用法。
本案例针对elasticsearch 5.x进行父子关系进行介绍，6.x版本通过join type来实现父子关系检索，参考文档后面提供的demo。
本文涉及技术点：
- 1.删除和创建公司/雇员父子关系的索引表
- 2.bulk批量导入json格式数据
- 3.ormapping方式bulk批量导入数据
- 4.采用@ESId指定文档_id
- 5.采用@ESParentId制定子文档的parent信息
- 6.基本的has_child和has_parent公司/雇员父子关系检索
1.准备工作
参考文档《高性能elasticsearch ORM开发库使用介绍》导入和配置es客户端
2.定义dsl文档
建立dsl配置文件-esmapper/indexparentchild.xml
<properties>
<!--
本案例适用于es 2.x,5.x
创建包含员工类型和公司类型的公司索引表
-->
<property name="createCompanyEmployeeIndice">
<![CDATA[{
"settings": {
"number_of_shards": 6,
"index.refresh_interval": "5s"
},
"mappings": {
"company": { ##公司
"properties": {
"name": {
"type": "text",
"fields": { ##dsl注释 定义精确查找的内部keyword字段
"keyword": {
"type": "keyword"
}
}
},
"city": {
"type": "text",
"fields": { ##dsl注释 定义精确查找的内部keyword字段
"keyword": {
"type": "keyword"
}
}
},
"country": {
"type": "text",
"fields": { ##dsl注释 定义精确查找的内部keyword字段
"keyword": {
"type": "keyword"
}
}
},
"companyId": {
"type": "keyword"
}
}
},
"employee":
{ ##雇员
"_parent": {##定义雇员和公司父子关联关系
"type": "company"
},
"_routing": {
"required": false
},
"properties": {
"name": {
"type": "text",
"fields": { ##dsl注释 定义精确查找的内部keyword字段
"keyword": {
"type": "keyword"
}
}
},
"birthday": {
"type": "date",
"format":"yyyy-MM-dd||epoch_millis"
},
"hobby": {
"type": "text",
"fields": { ##dsl注释 定义精确查找的内部keyword字段
"keyword": {
"type": "keyword"
}
}
},
"companyId": {
"type": "keyword"
},
"employId": {
"type": "keyword"
}
}
}
}
}]]>
</property>
<!--
导入公司信息：
-->
<property name="bulkImportCompanyData">
<![CDATA[
{ "index": { "_id": "london" }}
{ "name": "London Westminster", "city": "London", "country": "UK" ,"companyId":"london"}
{ "index": { "_id": "liverpool" }}
{ "name": "Liverpool Central", "city": "Liverpool", "country": "UK" ,"companyId":"liverpool"}
{ "index": { "_id": "paris" }}
{ "name": "Champs Élysées", "city": "Paris", "country": "France","companyId":"paris" }
]]>
</property>
<!--
导入雇员信息：
-->
<property name="bulkImportEmployeeData">
<![CDATA[
{ "index": { "_id": 1, "parent": "london" }}
{ "name": "Alice Smith", "birthday": "1970-10-24", "hobby": "hiking" ,"companyId":"london","employeeId":1 }
{ "index": { "_id": 2, "parent": "london" }}
{ "name": "Mark Thomas", "birthday": "1982-05-16", "hobby": "diving" ,"companyId":"london","employeeId":2}
{ "index": { "_id": 3, "parent": "liverpool" }}
{ "name": "Barry Smith", "birthday": "1979-04-01", "hobby": "hiking" ,"companyId":"liverpool","employeeId":3}
{ "index": { "_id": 4, "parent": "paris" }}
{ "name": "Adrien Grand", "birthday": "1987-05-11", "hobby": "horses" ,"companyId":"paris","employeeId":4}
{ "index": { "_id": 5, "parent": "paris" }}
{ "name": "Adrien Green", "birthday": "1977-05-12", "hobby": "dancing" ,"companyId":"paris","employeeId":5}
]]>
</property>
<!--以雇员出生日期为条件检索公司信息-->
<property name="hasChildSearchByBirthday">
<![CDATA[
{
"query": {
"has_child": {
"type": "employee",
"query": {
"range": {
"birthday": {
"gte": #[birthday]
}
}
}
}
}
}
]]>
</property>
<!--以雇员姓名为条件检索公司信息-->
<property name="hasChildSearchByName">
<![CDATA[
{
"query": {
"has_child": {
"type": "employee",
"score_mode": "max",
"query": {
"match": {
"name": #[name]
}
}
}
}
}
]]>
</property>
<!--以最小员工数为条件检索公司信息-->
<property name="hasChildSearchByMinChild">
<![CDATA[
{
"query": {
"has_child": {
"type": "employee",
"min_children": #[min_children],
"query": {
"match_all": {}
}
}
}
}
]]>
</property>
<!--根据公司所在国家检索员工信息-->
<property name="hasParentSearchByCountry">
<![CDATA[
{
"query": {
"has_parent": {
"type": "company",
"query": {
"match": {
"country": #[country]
}
}
}
}
}
]]>
</property>
</properties>3.实现has_child和has_parent检索
首先创建带公司和雇员关系的索引结构
public void createIndice(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
try {
//删除mapping
clientUtil.dropIndice("company");
} catch (ElasticSearchException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
//创建mapping
clientUtil.createIndiceMapping("company","createCompanyEmployeeIndice");
}然后通过bulk导入测试需要的公司和雇员数据，本案例通过加载配置文件中的dsl json data导入公司和雇员数据：
public void importData(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
//导入公司数据,并且实时刷新，测试需要，实际环境不要带refresh
clientUtil.executeHttp("company/company/_bulk?refresh","bulkImportCompanyData",ClientUtil.HTTP_POST);
//导入雇员数据,并且实时刷新，测试需要，实际环境不要带refresh
clientUtil.executeHttp("company/employee/_bulk?refresh","bulkImportEmployeeData",ClientUtil.HTTP_POST);
}如果需要根据List集合批量导入测试数据，则参考以下方法：
/**
* 通过List集合导入雇员和公司数据
*/
public void importDataFromBeans() {
ClientInterface clientUtil = ElasticSearchHelper.getRestClientUtil();
//导入公司数据,并且实时刷新，测试需要，实际环境不要带refresh
List<Company> companies = buildCompanies();
clientUtil.addDocuments("company","company",companies,"refresh");
//导入雇员数据,并且实时刷新，测试需要，实际环境不要带refresh
List<Employee> employees = buildEmployees();
clientUtil.addDocuments("company","employee",employees,"refresh");
}List<Company>和List<Employee>列表分别对应需要批量导入的公司数据和雇员数据。需要特别说明的是Company和Employee这两个对象采用了注解@ESId来标注文档_id属性，采用@ESParentId属性来标注雇员和公司的关联属性：
public class Company extends ESBaseData {
private String name;
/**
* 将companyId作为索引_id的值
*/
@ESId
private String companyId;
。。。。。。public class Employee extends ESBaseData {
/**
* 通过ESId注解将employeeId指定为雇员的文档_id
*/
@ESId
private int employeeId;
/**
* 通过ESParentId注解将companyId指定为雇员的parent属性，对应Company中的文档_id的值
*/
@ESParentId
private String companyId;接下来实现has_child和has_parent检索功能
/**
* 通过雇员生日检索公司信息
*/
public void hasChildSearchByBirthday(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
Map<String,Object> params = new HashMap<String,Object>();
params.put("birthday","1980-01-01");
ESDatas<Company> escompanys = clientUtil.searchList("company/company/_search","hasChildSearchByBirthday",params,Company.class);
List<Company> companyList = escompanys.getDatas();//获取符合条件的公司
long totalSize = escompanys.getTotalSize();
}
/**
* 通过雇员姓名检索公司信息
*/
public void hasChildSearchByName(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
Map<String,Object> params = new HashMap<String,Object>();
params.put("name","Alice Smith");
ESDatas<Company> escompanys = clientUtil.searchList("company/company/_search","hasChildSearchByName",params,Company.class);
List<Company> companyList = escompanys.getDatas();//获取符合条件的公司
long totalSize = escompanys.getTotalSize();
}
/**
* 通过雇员数量检索公司信息
*/
public void hasChildSearchByMinChild(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
Map<String,Object> params = new HashMap<String,Object>();
params.put("min_children",2);
ESDatas<Company> escompanys = clientUtil.searchList("company/company/_search","hasChildSearchByMinChild",params,Company.class);
List<Company> companyList = escompanys.getDatas();//获取符合条件的公司
long totalSize = escompanys.getTotalSize();
}
/**
* 通过公司所在国家检索雇员信息
*/
public void hasParentSearchByCountry(){
ClientInterface clientUtil = ElasticSearchHelper.getConfigRestClientUtil("esmapper/indexparentchild.xml");
Map<String,Object> params = new HashMap<String,Object>();
params.put("country","UK");
ESDatas<Employee> escompanys = clientUtil.searchList("company/employee/_search","hasParentSearchByCountry",params,Employee.class);
List<Employee> companyList = escompanys.getDatas();//获取符合条件的公司
long totalSize = escompanys.getTotalSize();
}通过junit测试用例执行上述功能
@Test
public void test(){
createIndice();
importData();
hasChildSearchByBirthday();
this.hasChildSearchByName();
this.hasChildSearchByMinChild();
this.hasParentSearchByCountry();
}4.参考文档
5.x父子查询示例：
https://gitee.com/bboss/eshelloword-booter/tree/master/src/test/java/org/bboss/elasticsearchtest/parentchild
6.x父子查询示例：
https://gitee.com/bboss/eshelloword-booter/tree/master/src/test/java/org/bboss/elasticsearchtest/jointype
https://blog.csdn.net/napoay/article/details/52032931
测试用例对应的工程
https://gitee.com/bbossgroups/eshelloword-booter