手撸dynamic源码详细讲解

本文源码解析基于3.3.1版本。只截了重点代码，如果需要看完整代码，可以去github拉取。

1 自动配置的实现

一般情况下，一个starter的最好入手点就是自动配置类，在 META-INF/spring.factories文件中指定自动配置类入口

org.springframework.boot.autoconfigure.EnableAutoConfiguration=\
com.baomidou.dynamic.datasource.spring.boot.autoconfigure.DynamicDataSourceAutoConfiguration

在spring.factories配置文件中，可以看到这个项目的自动配置类，从核心配置类入手，可以说DynamicDataSourceAutoConfiguration这个是整个程序的main方法，spring启动时会去执行
下面简单的给出DynamicDataSourceAutoConfiguration这个类的核心部分：

/*** 动态数据源核心自动配置类*/
@Slf4j
@Configuration
@AllArgsConstructor
// 读取以spring.datasource.dynamic为前缀的配置
@EnableConfigurationProperties(DynamicDataSourceProperties.class)
// 需要在spring boot的DataSource bean自动配置之前注入我们的DataSource bean
@AutoConfigureBefore(DataSourceAutoConfiguration.class)
// 引入了Druid的autoConfig和各种数据源连接池的Creator
@Import(value = {DruidDynamicDataSourceConfiguration.class, DynamicDataSourceCreatorAutoConfiguration.class})
// 当含有spring.datasource.dynamic配置的时候，启用这个autoConfig
@ConditionalOnProperty(prefix = DynamicDataSourceProperties.PREFIX, name = "enabled", havingValue = "true", matchIfMissing = true)
public class DynamicDataSourceAutoConfiguration {private final DynamicDataSourceProperties properties;/*** 多数据源加载接口，默认从yml中读取多数据源配置* @return DynamicDataSourceProvider*/@Bean@ConditionalOnMissingBeanpublic DynamicDataSourceProvider dynamicDataSourceProvider() {Map<String, DataSourceProperty> datasourceMap = properties.getDatasource();return new YmlDynamicDataSourceProvider(datasourceMap);}/*** 注册自己的动态多数据源DataSource* @param dynamicDataSourceProvider 各种数据源连接池创建者* @return DataSource*/@Bean@ConditionalOnMissingBeanpublic DataSource dataSource(DynamicDataSourceProvider dynamicDataSourceProvider) {DynamicRoutingDataSource dataSource = new DynamicRoutingDataSource();dataSource.setPrimary(properties.getPrimary());dataSource.setStrict(properties.getStrict());dataSource.setStrategy(properties.getStrategy());dataSource.setProvider(dynamicDataSourceProvider);dataSource.setP6spy(properties.getP6spy());dataSource.setSeata(properties.getSeata());return dataSource;}/*** AOP切面，对DS注解过的方法进行增强，达到切换数据源的目的。* @param dsProcessor 动态参数解析数据源。如果数据源名称以#开头，就会进入这个解析器链。* @return advisor*/@Role(value = BeanDefinition.ROLE_INFRASTRUCTURE)@Bean@ConditionalOnMissingBeanpublic DynamicDataSourceAnnotationAdvisor dynamicDatasourceAnnotationAdvisor(DsProcessor dsProcessor) {// aop方法拦截器在方法调用前后做操作，设置动态参数解析器DynamicDataSourceAnnotationInterceptor interceptor = new DynamicDataSourceAnnotationInterceptor(properties.isAllowedPublicOnly(), dsProcessor);// 使用AbstractPointcutAdvisor将pointcut和advice连接构成切面DynamicDataSourceAnnotationAdvisor advisor = new DynamicDataSourceAnnotationAdvisor(interceptor);advisor.setOrder(properties.getOrder());return advisor;}/*** seata分布式事务支持**/@Role(value = BeanDefinition.ROLE_INFRASTRUCTURE)@ConditionalOnProperty(prefix = DynamicDataSourceProperties.PREFIX, name = "seata", havingValue = "false", matchIfMissing = true)@Beanpublic Advisor dynamicTransactionAdvisor() {AspectJExpressionPointcut pointcut = new AspectJExpressionPointcut();pointcut.setExpression("@annotation(com.baomidou.dynamic.datasource.annotation.DSTransactional)");return new DefaultPointcutAdvisor(pointcut, new DynamicTransactionAdvisor());}/*** 动态参数解析器链* @return DsProcessor*/@Bean@ConditionalOnMissingBeanpublic DsProcessor dsProcessor() {DsHeaderProcessor headerProcessor = new DsHeaderProcessor();DsSessionProcessor sessionProcessor = new DsSessionProcessor();DsSpelExpressionProcessor spelExpressionProcessor = new DsSpelExpressionProcessor();headerProcessor.setNextProcessor(sessionProcessor);sessionProcessor.setNextProcessor(spelExpressionProcessor);return headerProcessor;}
}

这里自动配置的几个Bean都是非常重要的
先看下自动配置类上面的注解，比较重要的有如下的：

// 读取以spring.datasource.dynamic为前缀的配置
@EnableConfigurationProperties(DynamicDataSourceProperties.class)

@EnableConfigurationProperties这个注解：使使 @ConfigurationProperties 注解的类生效，主要是用来把properties或者yml配置文件转化为bean来使用，这个在实际使用中非常实用

2 配置文件注入

在跟进DynamicDataSourceProperties中：

@Slf4j
@Getter
@Setter
@ConfigurationProperties(prefix = DynamicDataSourceProperties.PREFIX)
public class DynamicDataSourceProperties {public static final String PREFIX = "spring.datasource.dynamic";public static final String HEALTH = PREFIX + ".health";/*** 必须设置默认的库,默认master*/private String primary = "master";/*** 是否启用严格模式,默认不启动. 严格模式下未匹配到数据源直接报错, 非严格模式下则使用默认数据源primary所设置的数据源*/private Boolean strict = false;/*** 是否使用p6spy输出，默认不输出*/private Boolean p6spy = false;/*** 是否使用开启seata，默认不开启*/private Boolean seata = false;/*** seata使用模式，默认AT*/private SeataMode seataMode = SeataMode.AT;/*** 是否使用 spring actuator 监控检查，默认不检查*/private boolean health = false;/*** 每一个数据源*/private Map<String, DataSourceProperty> datasource = new LinkedHashMap<>();/*** 多数据源选择算法clazz，默认负载均衡算法*/private Class<? extends DynamicDataSourceStrategy> strategy = LoadBalanceDynamicDataSourceStrategy.class;/*** aop切面顺序，默认优先级最高*/private Integer order = Ordered.HIGHEST_PRECEDENCE;/*** Druid全局参数配置*/@NestedConfigurationPropertyprivate DruidConfig druid = new DruidConfig();/*** HikariCp全局参数配置*/@NestedConfigurationPropertyprivate HikariCpConfig hikari = new HikariCpConfig();/*** 全局默认publicKey*/private String publicKey = CryptoUtils.DEFAULT_PUBLIC_KEY_STRING;/*** aop 切面是否只允许切 public 方法*/private boolean allowedPublicOnly = true;
}

可以发现，就是将spring.datasource.dynamic开头的配置文件注入并创建properties对象，需要注意的是，使用了@NestedConfigurationProperty嵌套了其他的配置类。如果不清楚嵌套的其他配置类是什么，看下DynamicDataSourceProperties这个类中的嵌套properties类。比如DruidConfig，代码较长就不全部粘贴进来了，有兴趣跟进看下。重点是它下面有个toProperties方法，为了实现yml配置中每个dataSource下面的durid可以独立配置(若不独立配置，则使用全局配置)，根据全局配置(druid数据池)和独立配置(每个数据源下单独的druid数据源配置)结合转换为Properties，然后在DruidDataSourceCreator类(下面回讲，作用是创建不同的数据源)中根据这个配置创建druid连接池

3 如何集成多种数据池并创建

集成连接池配置项是通过DynamicDataSourceProperties配置类实现的，其中除了上面提到的druid嵌套配置，还有hikari等，但是如何通过这些配置项生成真正的数据源连接池？让我们来看creator包下的类

见名知意，能看出这些就是根据DynamicDataSourceProperties中的配置，生成对应连接池，这里具体实现暂且不看，先看下不同数据源的数据池对象是怎么保存并在使用时获取的
还是最开始的自动配置类中的方法：

@Bean@ConditionalOnMissingBeanpublic DataSource dataSource(DynamicDataSourceProvider dynamicDataSourceProvider) {DynamicRoutingDataSource dataSource = new DynamicRoutingDataSource();dataSource.setPrimary(properties.getPrimary());dataSource.setStrict(properties.getStrict());dataSource.setStrategy(properties.getStrategy());dataSource.setProvider(dynamicDataSourceProvider);dataSource.setP6spy(properties.getP6spy());dataSource.setSeata(properties.getSeata());return dataSource;}

这里创建了个DynamicRoutingDataSource，该类实现了InitializingBean接口，在bean初始化时做一些操作。

@Slf4j
public class DynamicRoutingDataSource extends AbstractRoutingDataSource implements InitializingBean, DisposableBean {private static final String UNDERLINE = "_";/*** 所有数据库*/private final Map<String, DataSource> dataSourceMap = new ConcurrentHashMap<>();/*** 分组数据库*/private final Map<String, GroupDataSource> groupDataSources = new ConcurrentHashMap<>();@Setterprivate DynamicDataSourceProvider provider;@Setterprivate Class<? extends DynamicDataSourceStrategy> strategy = LoadBalanceDynamicDataSourceStrategy.class;@Setterprivate String primary = "master";@Setterprivate Boolean strict = false;@Setterprivate Boolean p6spy = false;@Setterprivate Boolean seata = false;@Overridepublic DataSource determineDataSource() {return getDataSource(DynamicDataSourceContextHolder.peek());}private DataSource determinePrimaryDataSource() {log.debug("dynamic-datasource switch to the primary datasource");return groupDataSources.containsKey(primary) ? groupDataSources.get(primary).determineDataSource() : dataSourceMap.get(primary);}/*** 获取当前所有的数据源** @return 当前所有数据源*/public Map<String, DataSource> getCurrentDataSources() {return dataSourceMap;}/*** 获取的当前所有的分组数据源** @return 当前所有的分组数据源*/public Map<String, GroupDataSource> getCurrentGroupDataSources() {return groupDataSources;}/*** 获取数据源** @param ds 数据源名称* @return 数据源*/public DataSource getDataSource(String ds) {if (StringUtils.isEmpty(ds)) {return determinePrimaryDataSource();} else if (!groupDataSources.isEmpty() && groupDataSources.containsKey(ds)) {log.debug("dynamic-datasource switch to the datasource named [{}]", ds);return groupDataSources.get(ds).determineDataSource();} else if (dataSourceMap.containsKey(ds)) {log.debug("dynamic-datasource switch to the datasource named [{}]", ds);return dataSourceMap.get(ds);}if (strict) {throw new RuntimeException("dynamic-datasource could not find a datasource named" + ds);}return determinePrimaryDataSource();}/*** 添加数据源** @param ds         数据源名称* @param dataSource 数据源*/public synchronized void addDataSource(String ds, DataSource dataSource) {DataSource oldDataSource = dataSourceMap.put(ds, dataSource);// 新数据源添加到分组this.addGroupDataSource(ds, dataSource);// 关闭老的数据源if (oldDataSource != null) {try {closeDataSource(oldDataSource);} catch (Exception e) {log.error("dynamic-datasource - remove the database named [{}]  failed", ds, e);}}log.info("dynamic-datasource - load a datasource named [{}] success", ds);}/*** 新数据源添加到分组** @param ds         新数据源的名字* @param dataSource 新数据源*/private void addGroupDataSource(String ds, DataSource dataSource) {if (ds.contains(UNDERLINE)) {String group = ds.split(UNDERLINE)[0];GroupDataSource groupDataSource = groupDataSources.get(group);if (groupDataSource == null) {try {groupDataSource = new GroupDataSource(group, strategy.getDeclaredConstructor().newInstance());groupDataSources.put(group, groupDataSource);} catch (Exception e) {throw new RuntimeException("dynamic-datasource - add the datasource named " + ds + " error", e);}}groupDataSource.addDatasource(ds, dataSource);}}/*** 删除数据源** @param ds 数据源名称*/public synchronized void removeDataSource(String ds) {if (!StringUtils.hasText(ds)) {throw new RuntimeException("remove parameter could not be empty");}if (primary.equals(ds)) {throw new RuntimeException("could not remove primary datasource");}if (dataSourceMap.containsKey(ds)) {DataSource dataSource = dataSourceMap.remove(ds);try {closeDataSource(dataSource);} catch (Exception e) {log.error("dynamic-datasource - remove the database named [{}]  failed", ds, e);}if (ds.contains(UNDERLINE)) {String group = ds.split(UNDERLINE)[0];if (groupDataSources.containsKey(group)) {DataSource oldDataSource = groupDataSources.get(group).removeDatasource(ds);if (oldDataSource == null) {if (log.isWarnEnabled()) {log.warn("fail for remove datasource from group. dataSource: {} ,group: {}", ds, group);}}}}log.info("dynamic-datasource - remove the database named [{}] success", ds);} else {log.warn("dynamic-datasource - could not find a database named [{}]", ds);}}/*** 关闭数据源。* <pre>*    从3.2.0开启，如果是原生或使用 DataSourceCreator 创建的数据源会包装成ItemDataSource。*    ItemDataSource保留了最原始的数据源，其可直接关闭。*    如果不是DataSourceCreator创建的数据源则只有尝试解包装再关闭。* </pre>*/private void closeDataSource(DataSource dataSource) throws Exception {if (dataSource instanceof ItemDataSource) {((ItemDataSource) dataSource).close();} else {if (seata && dataSource instanceof DataSourceProxy) {DataSourceProxy dataSourceProxy = (DataSourceProxy) dataSource;dataSource = dataSourceProxy.getTargetDataSource();}if (p6spy && dataSource instanceof P6DataSource) {Field realDataSourceField = P6DataSource.class.getDeclaredField("realDataSource");realDataSourceField.setAccessible(true);dataSource = (DataSource) realDataSourceField.get(dataSource);}Class<? extends DataSource> clazz = dataSource.getClass();Method closeMethod = clazz.getDeclaredMethod("close");closeMethod.invoke(dataSource);}}@Overridepublic void destroy() throws Exception {log.info("dynamic-datasource start closing ....");for (Map.Entry<String, DataSource> item : dataSourceMap.entrySet()) {closeDataSource(item.getValue());}log.info("dynamic-datasource all closed success,bye");}@Overridepublic void afterPropertiesSet() throws Exception {// 检查开启了配置但没有相关依赖checkEnv();// 添加并分组数据源Map<String, DataSource> dataSources = provider.loadDataSources();for (Map.Entry<String, DataSource> dsItem : dataSources.entrySet()) {addDataSource(dsItem.getKey(), dsItem.getValue());}// 检测默认数据源是否设置if (groupDataSources.containsKey(primary)) {log.info("dynamic-datasource initial loaded [{}] datasource,primary group datasource named [{}]", dataSources.size(), primary);} else if (dataSourceMap.containsKey(primary)) {log.info("dynamic-datasource initial loaded [{}] datasource,primary datasource named [{}]", dataSources.size(), primary);} else {throw new RuntimeException("dynamic-datasource Please check the setting of primary");}}private void checkEnv() {if (p6spy) {try {Class.forName("com.p6spy.engine.spy.P6DataSource");log.info("dynamic-datasource detect P6SPY plugin and enabled it");} catch (Exception e) {throw new RuntimeException("dynamic-datasource enabled P6SPY ,however without p6spy dependency", e);}}if (seata) {try {Class.forName("io.seata.rm.datasource.DataSourceProxy");log.info("dynamic-datasource detect ALIBABA SEATA and enabled it");} catch (Exception e) {throw new RuntimeException("dynamic-datasource enabled ALIBABA SEATA,however without seata dependency", e);}}}
}

这个类就是核心动态数据源组件。它将DataSource维护在map里，这里重点看如何创建数据源连接池。它所做的操作就是：afterPropertiesSet这个方法在当前Bean对象所有属性设置之后执行的，从provider获取创建好的数据源map，然后解析这个map对其分组。下面来看看这个provider里面是如何创建这个数据源map的
还是返回最开始的自动配置类，看下provider对象创建:

    @Bean@ConditionalOnMissingBeanpublic DynamicDataSourceProvider dynamicDataSourceProvider() {Map<String, DataSourceProperty> datasourceMap = properties.getDatasource();return new YmlDynamicDataSourceProvider(datasourceMap);}

在自动装配中注入的这个YmlDynamicDataSourceProvider，是通过yml读取配置文件生成的，调用构造方法时传递进去DynamicDataSourceProperties中的数据源名称和数据源下面配置(url,username,password等等)的map映射集合
继续跟进，看下YmlDynamicDataSourceProvider中的内容，除了传递进来的**Map<String, DataSourceProperty>**外，有且只有一个方法：

@Override
public Map<String, DataSource> loadDataSources() {return createDataSourceMap(dataSourcePropertiesMap);
}

我们继续跟进，一直到DefaultDataSourceCreator这个类的createDataSource方法为止，上问提到过creator就是用来根据properties配置创建数据池对象的，下面看下这个类的代码，重点是createDataSource方法：

@Slf4j
@Setter
public class DefaultDataSourceCreator implements DataSourceCreator {private DynamicDataSourceProperties properties;private List<DataSourceCreator> creators;@Overridepublic DataSource createDataSource(DataSourceProperty dataSourceProperty) {return createDataSource(dataSourceProperty, properties.getPublicKey());}@Overridepublic DataSource createDataSource(DataSourceProperty dataSourceProperty, String publicKey) {DataSourceCreator dataSourceCreator = null;for (DataSourceCreator creator : this.creators) {if (creator.support(dataSourceProperty)) {dataSourceCreator = creator;break;}}if (dataSourceCreator == null) {throw new IllegalStateException("creator must not be null,please check the DataSourceCreator");}DataSource dataSource = dataSourceCreator.createDataSource(dataSourceProperty, publicKey);this.runScrip(dataSource, dataSourceProperty);return wrapDataSource(dataSource, dataSourceProperty);}private void runScrip(DataSource dataSource, DataSourceProperty dataSourceProperty) {String schema = dataSourceProperty.getSchema();String data = dataSourceProperty.getData();if (StringUtils.hasText(schema) || StringUtils.hasText(data)) {ScriptRunner scriptRunner = new ScriptRunner(dataSourceProperty.isContinueOnError(), dataSourceProperty.getSeparator());if (StringUtils.hasText(schema)) {scriptRunner.runScript(dataSource, schema);}if (StringUtils.hasText(data)) {scriptRunner.runScript(dataSource, data);}}}private DataSource wrapDataSource(DataSource dataSource, DataSourceProperty dataSourceProperty) {String name = dataSourceProperty.getPoolName();DataSource targetDataSource = dataSource;Boolean enabledP6spy = properties.getP6spy() && dataSourceProperty.getP6spy();if (enabledP6spy) {targetDataSource = new P6DataSource(dataSource);log.debug("dynamic-datasource [{}] wrap p6spy plugin", name);}Boolean enabledSeata = properties.getSeata() && dataSourceProperty.getSeata();SeataMode seataMode = properties.getSeataMode();if (enabledSeata) {if (SeataMode.XA == seataMode) {targetDataSource = new DataSourceProxyXA(dataSource);} else {targetDataSource = new DataSourceProxy(dataSource);}log.debug("dynamic-datasource [{}] wrap seata plugin transaction mode [{}]", name, seataMode);}return new ItemDataSource(name, dataSource, targetDataSource, enabledP6spy, enabledSeata, seataMode);}public void setDataSourceCreators(List<DataSourceCreator> dataSourceCreator) {this.creators = dataSourceCreator;}@Overridepublic boolean support(DataSourceProperty dataSourceProperty) {return true;}
}

上文提到过creator见名知意，这里的createDataSource方法，就是根据对应的properties对象中的属性来判断使用哪个creator去创建datasource对象，拿到这里所有的过程都明朗了，这也就是dynamic集成多种数据池创建数据源的关键

4 @DS注解是如何被拦截增强的

先总结下，上面聊了，DynamicRoutingDataSource内部维持了一个map集合存储数据源名称和数据源对象的映射，而这个map映射是provider对象进行的，然后具体的类型的datasource又是由creator进行的
那么具体是如何通过注解实现的切换数据源的，众所周知，spring中注解拦截处理离不开AOP，这里介绍代码中如何使用AOP

再次返回最开始的自动配置类，从DynamicDataSourceAutoConfiguration入口配置类中dynamicDatasourceAnnotationAdvisor(DsProcessor dsProcessor)方法入手，该方法注入了一个DynamicDataSourceAnnotationAdvisor类型的bean对象
讲解之前，先简单介绍下spring中的advisor，几个概念：

advice: 具体进行的增强
pointcut：配置需要增强的规则，可以是切点表达式等
advisor: Advice 和 Pointcut 组成的独立的单元，并且能够传给 proxy factory 对象

我们看下DynamicDataSourceAnnotationAdvisor这个切面的代码：

public class DynamicDataSourceAnnotationAdvisor extends AbstractPointcutAdvisor implements BeanFactoryAware {// 通知private final Advice advice;// 切入点private final Pointcut pointcut;public DynamicDataSourceAnnotationAdvisor(@NonNull DynamicDataSourceAnnotationInterceptor dynamicDataSourceAnnotationInterceptor) {this.advice = dynamicDataSourceAnnotationInterceptor;this.pointcut = buildPointcut();}@Overridepublic Pointcut getPointcut() {return this.pointcut;}@Overridepublic Advice getAdvice() {return this.advice;}@Overridepublic void setBeanFactory(BeanFactory beanFactory) throws BeansException {if (this.advice instanceof BeanFactoryAware) {((BeanFactoryAware) this.advice).setBeanFactory(beanFactory);}}private Pointcut buildPointcut() {// 类级别Pointcut cpc = new AnnotationMatchingPointcut(DS.class, true);// 方法级别Pointcut mpc = new AnnotationMethodPoint(DS.class);// 合并类和方法上添加的注解，类上的注解会绑定到每个方法上。return new ComposablePointcut(cpc).union(mpc);}/*** In order to be compatible with the spring lower than 5.0*/private static class AnnotationMethodPoint implements Pointcut {private final Class<? extends Annotation> annotationType;public AnnotationMethodPoint(Class<? extends Annotation> annotationType) {Assert.notNull(annotationType, "Annotation type must not be null");this.annotationType = annotationType;}@Overridepublic ClassFilter getClassFilter() {return ClassFilter.TRUE;}@Overridepublic MethodMatcher getMethodMatcher() {return new AnnotationMethodMatcher(annotationType);}private static class AnnotationMethodMatcher extends StaticMethodMatcher {private final Class<? extends Annotation> annotationType;public AnnotationMethodMatcher(Class<? extends Annotation> annotationType) {this.annotationType = annotationType;}@Overridepublic boolean matches(Method method, Class<?> targetClass) {if (matchesMethod(method)) {return true;}// Proxy classes never have annotations on their redeclared methods.if (Proxy.isProxyClass(targetClass)) {return false;}// The method may be on an interface, so let's check on the target class as well.Method specificMethod = AopUtils.getMostSpecificMethod(method, targetClass);return (specificMethod != method && matchesMethod(specificMethod));}private boolean matchesMethod(Method method) {return AnnotatedElementUtils.hasAnnotation(method, this.annotationType);}}}
}

先现在看下@DS注解的advisor实现，在buildPointcut方法里拦截了被@DS注解的方法或类，并且使用ComposablePointcut组合切入点，可以实现方法优先级大于类优先级的特性。DynamicDataSourceAnnotationAdvisor通过构造方法传过来的参数类型是DynamicDataSourceAnnotationInterceptor类，跟进观察该类

public class DynamicDataSourceAnnotationInterceptor implements MethodInterceptor {/*** The identification of SPEL.*/private static final String DYNAMIC_PREFIX = "#";private final DataSourceClassResolver dataSourceClassResolver;private final DsProcessor dsProcessor;public DynamicDataSourceAnnotationInterceptor(Boolean allowedPublicOnly, DsProcessor dsProcessor) {dataSourceClassResolver = new DataSourceClassResolver(allowedPublicOnly);this.dsProcessor = dsProcessor;}@Overridepublic Object invoke(MethodInvocation invocation) throws Throwable {String dsKey = determineDatasourceKey(invocation);// 把获取到的数据源标识(如master)存入本地线程DynamicDataSourceContextHolder.push(dsKey);try {return invocation.proceed();} finally {DynamicDataSourceContextHolder.poll();}}private String determineDatasourceKey(MethodInvocation invocation) {String key = dataSourceClassResolver.findDSKey(invocation.getMethod(), invocation.getThis());// 如果DS注解内容是以#开头，则解析动态最终值；否则，直接返回。return (!key.isEmpty() && key.startsWith(DYNAMIC_PREFIX)) ? dsProcessor.determineDatasource(invocation, key) : key;}
}