共计 12008 个字符，预计需要花费 31 分钟才能阅读完成。
本文节选自《Spring 5 外围原理》
浏览本文之前，请先浏览以下内容：
30 个类手写 Spring 外围原理之自定义 ORM（上）（6）
30 个类手写 Spring 外围原理之自定义 ORM（下）（7）
4 动静数据源切换的底层原理
这里简略介绍一下 AbstractRoutingDataSource 的基本原理。实现数据源切换的性能就是自定义一个类扩大 AbstractRoutingDataSource 抽象类，其实相当于数据源的路由中介，能够实现在我的项目运行时依据相应 key 值切换到对应的 DataSource 上。先看看 AbstractRoutingDataSource 类的源码：
public abstract class AbstractRoutingDataSource extends AbstractDataSource implements InitializingBean {
/* 只列出局部代码 */
@Nullable
private Map<Object, Object> targetDataSources;
@Nullable
private Object defaultTargetDataSource;
private boolean lenientFallback = true;
private DataSourceLookup dataSourceLookup = new JndiDataSourceLookup();
@Nullable
private Map<Object, DataSource> resolvedDataSources;
@Nullable
private DataSource resolvedDefaultDataSource;
...
public Connection getConnection() throws SQLException {return this.determineTargetDataSource().getConnection();}
public Connection getConnection(String username, String password) throws SQLException {return this.determineTargetDataSource().getConnection(username, password);
}
...
protected DataSource determineTargetDataSource() {Assert.notNull(this.resolvedDataSources, "DataSource router not initialized");
Object lookupKey = this.determineCurrentLookupKey();
DataSource dataSource = (DataSource)this.resolvedDataSources.get(lookupKey);
if(dataSource == null && (this.lenientFallback || lookupKey == null)) {dataSource = this.resolvedDefaultDataSource;}
if(dataSource == null) {throw new IllegalStateException("Cannot determine target DataSource for lookup key [" + lookupKey + "]");
} else {return dataSource;}
}
@Nullable
protected abstract Object determineCurrentLookupKey();}
能够看出，AbstractRoutingDataSource 类继承了 AbstractDataSource 类，并实现了 InitializingBean。AbstractRoutingDataSource 类的 getConnection() 办法调用了 determineTargetDataSource() 的该办法。这里重点看 determineTargetDataSource() 办法的代码，它应用了 determineCurrentLookupKey() 办法，它是 AbstractRoutingDataSource 类的形象办法，也是实现数据源切换扩大的办法。该办法的返回值就是我的项目中所要用的 DataSource 的 key 值，失去该 key 值后就能够在 resolvedDataSource 中取出对应的 DataSource，如果找不到 key 对应的 DataSource 就应用默认的数据源。
自定义类扩大 AbstractRoutingDataSource 类时要重写 determineCurrentLookupKey() 办法来实现数据源切换。
4.1 DynamicDataSource
DynamicDataSource 类封装自定义数据源，继承原生 Spring 的 AbstractRoutingDataSource 类的数据源动静路由器。
package javax.core.common.jdbc.datasource;
import org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource;
/**
* 动静数据源
*/
public class DynamicDataSource extends AbstractRoutingDataSource {
private DynamicDataSourceEntry dataSourceEntry;
@Override
protected Object determineCurrentLookupKey() {return this.dataSourceEntry.get();
}
public void setDataSourceEntry(DynamicDataSourceEntry dataSourceEntry) {this.dataSourceEntry = dataSourceEntry;}
public DynamicDataSourceEntry getDataSourceEntry(){return this.dataSourceEntry;}
}
4.2 DynamicDataSourceEntry
DynamicDataSourceEntry 类实现对数据源的操作性能，代码如下：
package javax.core.common.jdbc.datasource;
import org.aspectj.lang.JoinPoint;
/**
* 动静切换数据源
*/
public class DynamicDataSourceEntry {
// 默认数据源
public final static String DEFAULT_SOURCE = null;
private final static ThreadLocal<String> local = new ThreadLocal<String>();
/**
* 清空数据源
*/
public void clear() {local.remove();
}
/**
* 获取以后正在应用的数据源的名字
*
* @return String
*/
public String get() {return local.get();
}
/**
* 还原指定切面的数据源
*
* @param joinPoint
*/
public void restore(JoinPoint join) {local.set(DEFAULT_SOURCE);
}
/**
* 还原以后切面的数据源
*/
public void restore() {local.set(DEFAULT_SOURCE);
}
/**
* 设置已知名字的数据源
*
* @param dataSource
*/
public void set(String source) {local.set(source);
}
/**
* 依据年份动静设置数据源
* @param year
*/
public void set(int year) {local.set("DB_" + year);
}
}
5 运行成果演示
5.1 创立 Member 实体类
创立 Member 实体类代码如下：
package com.tom.orm.demo.entity;
import lombok.Data;
import javax.persistence.Entity;
import javax.persistence.Id;
import javax.persistence.Table;
import java.io.Serializable;
@Entity
@Table(name="t_member")
@Data
public class Member implements Serializable {
@Id private Long id;
private String name;
private String addr;
private Integer age;
@Override
public String toString() {
return "Member{" +
"id=" + id +
", name='" + name + '\'' +
", addr='" + addr + '\'' +
", age=" + age +
'}';
}
}
5.2 创立 Order 实体类
创立 Order 实体类代码如下：
package com.tom.orm.demo.entity;
import lombok.Data;
import javax.persistence.Column;
import javax.persistence.Entity;
import javax.persistence.Table;
import java.io.Serializable;
@Entity
@Table(name="t_order")
@Data
public class Order implements Serializable {
private Long id;
@Column(name="mid")
private Long memberId;
private String detail;
private Long createTime;
private String createTimeFmt;
@Override
public String toString() {
return "Order{" +
"id=" + id +
", memberId=" + memberId +
", detail='" + detail + '\'' +
", createTime=" + createTime +
", createTimeFmt='" + createTimeFmt + '\'' +
'}';
}
}
5.3 创立 MemberDao
创立 MemberDao 代码如下：
package com.tom.orm.demo.dao;
import com.tom.orm.demo.entity.Member;
import com.tom.orm.framework.BaseDaoSupport;
import com.tom.orm.framework.QueryRule;
import org.springframework.stereotype.Repository;
import javax.annotation.Resource;
import javax.sql.DataSource;
import java.util.List;
@Repository
public class MemberDao extends BaseDaoSupport<Member,Long> {
@Override
protected String getPKColumn() {return "id";}
@Resource(name="dataSource")
public void setDataSource(DataSource dataSource){super.setDataSourceReadOnly(dataSource);
super.setDataSourceWrite(dataSource);
}
public List<Member> selectAll() throws Exception{QueryRule queryRule = QueryRule.getInstance();
queryRule.andLike("name","Tom%");
return super.select(queryRule);
}
}
5.4 创立 OrderDao
创立 OrderDao 代码如下：
package com.tom.orm.demo.dao;
import com.tom.orm.demo.entity.Order;
import com.tom.orm.framework.BaseDaoSupport;
import org.springframework.stereotype.Repository;
import javax.annotation.Resource;
import javax.core.common.jdbc.datasource.DynamicDataSource;
import javax.sql.DataSource;
import java.text.SimpleDateFormat;
import java.util.Date;
@Repository
public class OrderDao extends BaseDaoSupport<Order, Long> {private SimpleDateFormat yearFormat = new SimpleDateFormat("yyyy");
private SimpleDateFormat fullDataFormat = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
private DynamicDataSource dataSource;
@Override
protected String getPKColumn() {return "id";}
@Resource(name="dynamicDataSource")
public void setDataSource(DataSource dataSource) {this.dataSource = (DynamicDataSource)dataSource;
this.setDataSourceReadOnly(dataSource);
this.setDataSourceWrite(dataSource);
}
/**
* @throws Exception
*
*/
public boolean insertOne(Order order) throws Exception{
// 约定优于配置
Date date = null;
if(order.getCreateTime() == null){date = new Date();
order.setCreateTime(date.getTime());
}else {date = new Date(order.getCreateTime());
}
Integer dbRouter = Integer.valueOf(yearFormat.format(date));
System.out.println("主动调配到【DB_" + dbRouter + "】数据源");
this.dataSource.getDataSourceEntry().set(dbRouter);
order.setCreateTimeFmt(fullDataFormat.format(date));
Long orderId = super.insertAndReturnId(order);
order.setId(orderId);
return orderId > 0;
}
}
5.5 批改 db.properties 文件
批改 db.properties 文件代码如下：
#sysbase database mysql config
#mysql.jdbc.driverClassName=com.mysql.jdbc.Driver
#mysql.jdbc.url=jdbc:mysql://127.0.0.1:3306/gp-vip-spring-db-demo?characterEncoding=UTF-8&rewriteBatchedStatements=true
#mysql.jdbc.username=root
#mysql.jdbc.password=123456
db2018.mysql.jdbc.driverClassName=com.mysql.jdbc.Driver
db2018.mysql.jdbc.url=jdbc:mysql://127.0.0.1:3306/gp-vip-spring-db-2018?characterEncoding=UTF-8&rewriteBatchedStatements=true
db2018.mysql.jdbc.username=root
db2018.mysql.jdbc.password=123456
db2019.mysql.jdbc.driverClassName=com.mysql.jdbc.Driver
db2019.mysql.jdbc.url=jdbc:mysql://127.0.0.1:3306/gp-vip-spring-db-2019?characterEncoding=UTF-8&rewriteBatchedStatements=true
db2019.mysql.jdbc.username=root
db2019.mysql.jdbc.password=123456
#alibaba druid config
dbPool.initialSize=1
dbPool.minIdle=1
dbPool.maxActive=200
dbPool.maxWait=60000
dbPool.timeBetweenEvictionRunsMillis=60000
dbPool.minEvictableIdleTimeMillis=300000
dbPool.validationQuery=SELECT 'x'
dbPool.testWhileIdle=true
dbPool.testOnBorrow=false
dbPool.testOnReturn=false
dbPool.poolPreparedStatements=false
dbPool.maxPoolPreparedStatementPerConnectionSize=20
dbPool.filters=stat,log4j,wall
5.6 批改 application-db.xml 文件
批改 application-db.xml 文件代码如下：
<bean id="datasourcePool" abstract="true" class="com.alibaba.druid.pool.DruidDataSource" init-method="init" destroy-method="close">
<property name="initialSize" value="${dbPool.initialSize}" />
<property name="minIdle" value="${dbPool.minIdle}" />
<property name="maxActive" value="${dbPool.maxActive}" />
<property name="maxWait" value="${dbPool.maxWait}" />
<property name="timeBetweenEvictionRunsMillis" value="${dbPool.timeBetweenEvictionRunsMillis}" />
<property name="minEvictableIdleTimeMillis" value="${dbPool.minEvictableIdleTimeMillis}" />
<property name="validationQuery" value="${dbPool.validationQuery}" />
<property name="testWhileIdle" value="${dbPool.testWhileIdle}" />
<property name="testOnBorrow" value="${dbPool.testOnBorrow}" />
<property name="testOnReturn" value="${dbPool.testOnReturn}" />
<property name="poolPreparedStatements" value="${dbPool.poolPreparedStatements}" />
<property name="maxPoolPreparedStatementPerConnectionSize" value="${dbPool.maxPoolPreparedStatementPerConnectionSize}" />
<property name="filters" value="${dbPool.filters}" />
</bean>
<bean id="dataSource" parent="datasourcePool">
<property name="driverClassName" value="${db2019.mysql.jdbc.driverClassName}" />
<property name="url" value="${db2019.mysql.jdbc.url}" />
<property name="username" value="${db2019.mysql.jdbc.username}" />
<property name="password" value="${db2019.mysql.jdbc.password}" />
</bean>
<bean id="dataSource2018" parent="datasourcePool">
<property name="driverClassName" value="${db2018.mysql.jdbc.driverClassName}" />
<property name="url" value="${db2018.mysql.jdbc.url}" />
<property name="username" value="${db2018.mysql.jdbc.username}" />
<property name="password" value="${db2018.mysql.jdbc.password}" />
</bean>
<bean id="dynamicDataSourceEntry" class="javax.core.common.jdbc.datasource.DynamicDataSourceEntry" />
<bean id="dynamicDataSource" class="javax.core.common.jdbc.datasource.DynamicDataSource" >
<property name="dataSourceEntry" ref="dynamicDataSourceEntry"></property>
<property name="targetDataSources">
<map>
<entry key="DB_2019" value-ref="dataSource"></entry>
<entry key="DB_2018" value-ref="dataSource2018"></entry>
</map>
</property>
<property name="defaultTargetDataSource" ref="dataSource" />
</bean>
5.7 编写测试用例
编写测试用例代码如下：
package com.tom.orm.test;
import com.tom.orm.demo.dao.MemberDao;
import com.tom.orm.demo.dao.OrderDao;
import com.tom.orm.demo.entity.Member;
import com.tom.orm.demo.entity.Order;
import org.junit.Ignore;
import org.junit.Test;
import org.junit.runner.RunWith;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.test.context.ContextConfiguration;
import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;
import java.text.SimpleDateFormat;
import java.util.Arrays;
import java.util.Date;
import java.util.List;
@ContextConfiguration(locations = {"classpath:application-context.xml"})
@RunWith(SpringJUnit4ClassRunner.class)
public class OrmTest {private SimpleDateFormat sdf = new SimpleDateFormat("yyyyMMddHHmmdd");
@Autowired private MemberDao memberDao;
@Autowired private OrderDao orderDao;
@Test
public void testSelectAllForMember(){
try {List<Member> result = memberDao.selectAll();
System.out.println(Arrays.toString(result.toArray()));
} catch (Exception e) {e.printStackTrace();
}
}
@Test
@Ignore
public void testInsertMember(){
try {for (int age = 25; age < 35; age++) {Member member = new Member();
member.setAge(age);
member.setName("Tom");
member.setAddr("Hunan Changsha");
memberDao.insert(member);
}
}catch (Exception e){e.printStackTrace();
}
}
@Test
// @Ignore
public void testInsertOrder(){
try {Order order = new Order();
order.setMemberId(1L);
order.setDetail("历史订单");
Date date = sdf.parse("20180201123456");
order.setCreateTime(date.getTime());
orderDao.insertOne(order);
}catch (Exception e){e.printStackTrace();
}
}
}
所谓 ORM 就是，对象关系映射，Object Relation Mapping，市面上 ORM 框架也十分多，比方 Hibernate、Spring JDBC、MyBatis、JPA，它们都有对象关系治理的机制比方一对多、多对多、一对一关系。以上思路仅供参考，有趣味的小伙伴能够参考本文提供的思维，约定优于配置，先制订顶层接口，参数返回值全副对立，比方：
//List<?> Page<?> select(QueryRule queryRule)
//Int delete(T entity) entity 中的 ID 不能为空，如果 ID 为空，其余条件不能为空，都为空不予执行
//ReturnId insert(T entity) 只有 entity 不等于 null
//Int update(T entity) entity 中的 ID 不能为空，如果 ID 为空，其余条件不能为空，都为空不予执行 而后在此基础上进行扩大，基于 Spring JDBC 封装一套，基于 Redis 封装一套，基于 MongoDB 封装一套，基于 ElasticSearch 封装一套，基于 Hive 封装一套，基于 HBase 封装一套。本文残缺地演示了自研 ORM 框架的原理，以及数据源动静切换的基本原理，并且理解了 Spring JdbcTemplate 的 API 利用。心愿通过本章的学习，“小伙伴们”在日常工作中可能有更好的解决问题的思路，进步工作效率。
关注微信公众号『Tom 弹架构』回复“Spring”可获取残缺源码。
本文为“Tom 弹架构”原创，转载请注明出处。技术在于分享，我分享我高兴！如果您有任何倡议也可留言评论或私信，您的反对是我保持创作的能源。关注微信公众号『Tom 弹架构』可获取更多技术干货！
原创不易，保持很酷，都看到这里了，小伙伴记得点赞、珍藏、在看，一键三连加关注！如果你感觉内容太干，能够分享转发给敌人滋润滋润！
正文完
