最近忙于项目已经好久几天没写博客了，前2篇文章我给大家介绍了搭建基础springMvc+mybatis的maven工程，这个简单框架已经可以对付一般的小型项目。但是我们实际项目中会碰到很多复杂的场景，比如数据量很大的情况下如何保证性能。今天我就给大家介绍数据库分库分表的优化，本文介绍mybatis结合当当网的sharding-jdbc分库分表技术（原理这里不做介绍）
首先在pom文件中引入需要的依赖
<dependency>
<groupId>com.dangdang</groupId>
<artifactId>sharding-jdbc-core</artifactId>
<version>1.4.2</version>
</dependency>
<dependency>
<groupId>com.dangdang</groupId>
<artifactId>sharding-jdbc-config-spring</artifactId>
<version>1.4.0</version>
</dependency>
二、新建一个sharding-jdbc.xml文件，实现分库分表的配置
<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:context="http://www.springframework.org/schema/context"
xmlns:tx="http://www.springframework.org/schema/tx"
xmlns:rdb="http://www.dangdang.com/schema/ddframe/rdb"
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans.xsd
http://www.springframework.org/schema/tx
http://www.springframework.org/schema/tx/spring-tx.xsd
http://www.springframework.org/schema/context
http://www.springframework.org/schema/context/spring-context.xsd
http://www.dangdang.com/schema/ddframe/rdb
http://www.dangdang.com/schema/ddframe/rdb/rdb.xsd">
<rdb:strategy id="tableShardingStrategy" sharding-columns="user_id" algorithm-class="com.meiren.member.common.sharding.MemberSingleKeyTableShardingAlgorithm"/>
<rdb:data-source id="shardingDataSource">
<rdb:sharding-rule data-sources="dataSource">
<rdb:table-rules>
<rdb:table-rule logic-table="member_index" actual-tables="member_index_tbl_${[0,1,2,3,4,5,6,7,8,9]}${0..9}" table-strategy="tableShardingStrategy"/>
<rdb:table-rule logic-table="member_details" actual-tables="member_details_tbl_${[0,1,2,3,4,5,6,7,8,9]}${0..9}" table-strategy="tableShardingStrategy"/>
</rdb:table-rules>
</rdb:sharding-rule>
</rdb:data-source>
<bean id="transactionManager" class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
<property name="dataSource" ref="shardingDataSource" />
</bean>
</beans>
这里我简单介绍下一些属性的含义，
<rdb:strategy id="tableShardingStrategy" sharding-columns="user_id" algorithm-class="com.meiren.member.common.sharding.MemberSingleKeyTableShardingAlgorithm"/> 配置分表规则器 sharding-columns：分表规 则
依赖的名（根据user_id取模分表），algorithm-class:分表规则的实现类
<rdb:sharding-rule data-sources="dataSource"> 这里填写关联数据源（多个数据源用逗号隔开）,
<rdb:table-rule logic-table="member_index" actual-tables="member_index_tbl_${[0,1,2,3,4,5,6,7,8,9]}${0..9}" table-strategy="tableShardingStrategy"/> logic-table：逻辑表名（mybatis中代替的表名）actual-tables：
数据库实际的表名，这里支持inline表达式，比如：member_index_tbl_${0..2}会解析成member_index_tbl_0，member_index_tbl_1，member_index_tbl_2；member_index_tbl_${[a,b,c]}会被解析成
member_index_tbl_a,member_index_tbl_b和member_index_tbl_c,两种表达式一起使用的时候，会采取笛卡尔积的方式：member_index_tbl_${[a,b]}${0..2}解析为member_index_tbl_a0,member_index_tbl_a1 member_index_tbl_a2,member_index_tbl_b0,member_index_tbl_b1,member_index_tbl_b2;table-strategy:前面定义的分表规则器；
三、配置好改文件后，需要修改之前我们的spring-dataSource的几个地方，把sqlSessionFactory和transactionManager原来关联的dataSource统一修改为shardingDataSource（这一步作用就是把数据源全部托管给sharding去管理）
四、实现分表（分库）逻辑，我们的分表逻辑类需要实现SingleKeyTableShardingAlgorithm接口的三个方法doBetweenSharding、doEqualSharding、doInSharding
/**
* 分表逻辑
* @author zhangwentao
*
*/
public class MemberSingleKeyTableShardingAlgorithm implements SingleKeyTableShardingAlgorithm<Long> {
/**
* sql between 规则
*/
public Collection<String> doBetweenSharding(Collection<String> tableNames, ShardingValue<Long> shardingValue) {
Collection<String> result = new LinkedHashSet<String>(tableNames.size());
Range<Long> range = (Range<Long>) shardingValue.getValueRange();
for (long i = range.lowerEndpoint(); i <= range.upperEndpoint(); i++) {
Long modValue = i % 100;
String modStr = modValue < 10 ? "0" + modValue : modValue.toString();
for (String each : tableNames) {
if (each.endsWith(modStr)) {
result.add(each);
}
}
}
return result;
}
/**
* sql == 规则
*/
public String doEqualSharding(Collection<String> tableNames, ShardingValue<Long> shardingValue) {
Long modValue = shardingValue.getValue() % 100;
String modStr = modValue < 10 ? "0" + modValue : modValue.toString();
for (String each : tableNames) {
if (each.endsWith(modStr)) {
return each;
}
}
throw new IllegalArgumentException();
}
/**
* sql in 规则
*/
public Collection<String> doInSharding(Collection<String> tableNames, ShardingValue<Long> shardingValue) {
Collection<String> result = new LinkedHashSet<String>(tableNames.size());
for (long value : shardingValue.getValues()) {
Long modValue = value % 100;
String modStr = modValue < 10 ? "0" + modValue : modValue.toString();
for (String tableName : tableNames) {
if (tableName.endsWith(modStr)) {
result.add(tableName);
}
}
}
return result;
}
}
五、以上四步，我们就完成了sharding-jdbc的搭建，我们可以写一个测试demo来检查我们的成果
<select id="getDetailsById" resultType="com.meiren.member.dataobject.MemberDetailsDO"
parameterType="java.lang.Long">
select user_id userId ,qq,email from member_details where user_id =#{userId} limit 1
</select>
private static final String SERVICE_PROVIDER_XML = "/spring/member-service.xml";
private static final String BEAN_NAME = "idcacheService";
private ClassPathXmlApplicationContext context = null;
IdcacheServiceImpl bean = null;
IdcacheDao idcacheDao;
@Before
public void before() {
context= new ClassPathXmlApplicationContext(
new String[] {SERVICE_PROVIDER_XML});
idcacheDao=context.getBean("IdcacheDao", IdcacheDao.class);
}
@Test
public void getAllCreditActionTest() {
// int id = bean.insertIdcache();
Long s=100l;
MemberDetailsDO memberDetailsDO=idcacheDao.getDetailsById(s);
System.out.println("QQ---------------------"+memberDetailsDO.getQq());
}
打印sql语句，输出结果：QQ-------------------------------------100，证明成功！
注意点：这次搭建过程中，我有碰到一个小坑，就是执行的时候会报错：，官方文档是有解决方案：引入 <context:property-placeholder location="classpath:/member_service.properties" ignore-unresolvable="true" /> ，引入这行代码的时候，·必须要要把这边管理配配置文件的bean删除，换句话说，即Spring容器仅允许最多定义一个PropertyPlaceholderConfigurer(或<context:property-placeholder/>)，其余的会被Spring忽略掉（当时搞了半天啊）
小结：这次给大家分享了sharding-jdbc的配置是为了解决大数据量进行分库分表的架构，下一张，我将介绍拆分业务所需的duboo+zookeeper的配置（分布式），欢迎关注！