Spring Mybatis-分页插件使用

Mybatis分页切入点

Mybatis内部有个plugins(插件)概念，本质上属于拦截器的思想。具体的解析可见他文MyBatis拦截器原理探究。本文将在此基础上直接展示实际项目的实现代码和其他的相关解析

分页具体代码实现

首先我们可以定义方言抽象类，用于实现分页AbstractDialect.java

public abstract class AbstractDialect{
    /**
     * 是否支持limit和偏移量
     * @return
     */
    public abstract boolean supportsLimitOffset();

    /**
     * 是否支持limit
     * @return
     */
    public abstract boolean supportsLimit();

    /**
     * 获取增加了分页属性之后的SQL
     * @param sql
     * @param offset
     * @param limit
     * @return
     */
    public abstract String getLimitString(String sql, int offset, int limit);
}

再而我们就以Oracle与Mysql数据库的分页技术作下分别的实现

---

MySQLDialect.java-Mysql分页方言

public class MySQLDialect extends AbstractDialect {

    public boolean supportsLimitOffset() {
        return true;
    }

    public boolean supportsLimit() {
        return true;
    }

    public String getLimitString(String sql, int offset, int limit) {
        if (offset > 0) {
            return sql + " limit " + offset + "," + limit;
        } else {
            return sql + " limit " + limit;
        }
    }

}

OracleDialect.java-Oracle方言实现

public class OracleDialect extends ADialect{
    @Override
    public boolean supportsLimitOffset() {
        return false;
    }

    @Override
    public boolean supportsLimit() {
        return false;
    }

    @Override
    public String getLimitString(String sql, int start, int limit) {
        if(start < 0){
            start = 0;
        }
        if(limit < 0){
            limit = 10;
        }
        StringBuilder pageSql = new StringBuilder(100);
        pageSql.append("select * from ( select temp.*, rownum row_id from ( ");
        pageSql.append(sql);
        pageSql.append(" ) temp where rownum <= ").append(start+limit);
        pageSql.append(") where row_id > ").append(start);
        return pageSql.toString();
    }
}

---

对应的Mybatis插件拦截器实现如下，拦截StatementHandler#prepare(Connection con)创建SQL语句对象方法

PaginationInterceptor.java

@Intercepts({ @Signature(type = StatementHandler.class, method = "prepare", args = { Connection.class }) })
public final class PaginationInterceptor implements Interceptor {

    private final static Logger log = LoggerFactory
        .getLogger(PaginationInterceptor.class);

    private ADialect dialect;

    public void setDialect(ADialect dialect) {
        this.dialect = dialect;
    }

    @Override
    public Object intercept(Invocation invocation) throws Throwable {
		// 直接获取拦截的对象，其实现类RoutingStatementHandler
        StatementHandler statementHandler = (StatementHandler) invocation
            .getTarget();
		
		// 获取元对象，主要用于获取statementHandler所关联的对象及属性
        MetaObject metaStatementHandler = MetaObject.forObject(
            statementHandler, new DefaultObjectFactory(),
            new DefaultObjectWrapperFactory());

        MappedStatement mappedStmt= (MappedStatement) metaStatementHandler
            .getValue("delegate.mappedStatement".intern());
	   // 只对queryPagination()方法进行分页操作
	   if(mappedStmt.getId().indexOf("queryPagination")==-1){
			return invocation.proceed();
		}
        
		// 重新构造分页的sql
        String originalSql = (String) metaStatementHandler
            .getValue("delegate.boundSql.sql".intern());
        
        // 注意RowBounds是必须被提前指定的，用于分页语句的拼装
        RowBounds rowBounds = metaStatementHandler.getValue("delegate.rowBounds") ;

        metaStatementHandler.setValue("delegate.boundSql.sql".intern(), dialect
            .getLimitString(originalSql, rowBounds.getOffset(),
                rowBounds.getLimit()));

        metaStatementHandler.setValue("delegate.rowBounds.offset".intern(),
            RowBounds.NO_ROW_OFFSET);

        metaStatementHandler.setValue("delegate.rowBounds.limit".intern(),
            RowBounds.NO_ROW_LIMIT);

        log.debug("page sql : " + boundSql.getSql());

        return invocation.proceed();

    }

    // 拦截对象
    @Override
    public Object plugin(Object target) {
        return Plugin.wrap(target, this);

    }

    @Override
    public void setProperties(Properties properties) {

    }

}

---

Spring对应的xml配置可如下，以oracle分页为例子

	<!-- oracle方言配置，用于oracle的分页 -->
	<bean id="paginationInterceptor"		class="com.hsnet.winner.cas.admin.core.dao.mybatis.interceptor.PaginationInterceptor">
		<property name="dialect">
			<bean class="cn.cloud.winner.oss.manager.mybatis.page.OracleDialect" />
		</property>
	</bean>
	
	<!--sqlSessionFactoryBean配置plugins插件用于拦截-->
	<bean id="sqlSessionFactory" class="org.mybatis.spring.SqlSessionFactoryBean">
		<property name="dataSource" ref="dataSource" />
		<property name="mapperLocations" value="classpath*:mybatis/*DAO.xml" />
		<property name="plugins">
			<array>
				<ref bean="paginationInterceptor" />
			</array>
		</property>
	</bean>

使用以上的代码以及配置即可完成对oracle数据库以及mysql数据库的分页操作。并且博主对其中的某个点作下解析

Mybatis#MetaObject-元数据对象解析

以上的代码博主之前在使用的时候，对其中的MetaObject这个类很费解，其直接通过getValue()方法便可以将所代理的对象的所关联的属性全都拿取到。我们可以跟随源码深入了解下

MetaObject#forObject()

代理对象均通过此静态方法进入

public static MetaObject forObject(Object object, ObjectFactory objectFactory, ObjectWrapperFactory objectWrapperFactory) {
    if (object == null) {
      return SystemMetaObject.NULL_META_OBJECT;
    } else {
      return new MetaObject(object, objectFactory, objectWrapperFactory);
    }
  }

我们可以直接观察其中的构造函数，玄机就在此处

private MetaObject(Object object, ObjectFactory objectFactory, ObjectWrapperFactory objectWrapperFactory) {
    this.originalObject = object;
    this.objectFactory = objectFactory;
    this.objectWrapperFactory = objectWrapperFactory;
    // 所有的属性获取均通过objectWrapper类来获取，此处主要对所代理的object对象类型进行判断
    if (object instanceof ObjectWrapper) {
      this.objectWrapper = (ObjectWrapper) object;
    } else if (objectWrapperFactory.hasWrapperFor(object)) {
      this.objectWrapper = objectWrapperFactory.getWrapperFor(this, object);
    } else if (object instanceof Map) {
      this.objectWrapper = new MapWrapper(this, (Map) object);
    } else if (object instanceof Collection) {
      this.objectWrapper = new CollectionWrapper(this, (Collection) object);
    } else {
      // 我们常用的便是BeanWrapper
      this.objectWrapper = new BeanWrapper(this, object);
    }
  }

为了理解的更为渗透，我们继续跟进，最后我们得知其会调用Reflector类的构造函数

private Reflector(Class<?> clazz) {
    type = clazz;
    // 获取构造类
    addDefaultConstructor(clazz);
    // 获取get方法集合
    addGetMethods(clazz);
    // 获取set方法集合
    addSetMethods(clazz);
    // 获取内部属性集合
    addFields(clazz);
    readablePropertyNames = getMethods.keySet().toArray(new String[getMethods.keySet().size()]);
    writeablePropertyNames = setMethods.keySet().toArray(new String[setMethods.keySet().size()]);
    for (String propName : readablePropertyNames) {
      caseInsensitivePropertyMap.put(propName.toUpperCase(Locale.ENGLISH), propName);
    }
    for (String propName : writeablePropertyNames) {
      caseInsensitivePropertyMap.put(propName.toUpperCase(Locale.ENGLISH), propName);
    }
  }

由此我们便可知使用Reflector代理类以及MetaObject便可以遍历代理被代理类的所关联的所有属性，就拿RoutingStatementHandler类来说，经过上述操作后其便可以访问内部属性delegate以及delegate的内部属性configuration/objectFactory/typeHandlerRegistry/resultSetHandler/parameterHandler/mappedStatement等属性

MetaObject#getValue()

上述阐述的是如何代理被代理类的内部属性，我们也简单的看下是如何正确的调用

public Object getValue(String name) {
	// PropertyTokenizer与StringTokenizer类似，只是前者写死以.为分隔符
    PropertyTokenizer prop = new PropertyTokenizer(name);
    if (prop.hasNext()) {
      MetaObject metaValue = metaObjectForProperty(prop.getIndexedName());
      if (metaValue == SystemMetaObject.NULL_META_OBJECT) {
        return null;
      } else {
        return metaValue.getValue(prop.getChildren());
      }
    } else {
      return objectWrapper.get(prop);
    }
  }

具体的解析就不在此阐述了，如何用户想获取StatementHandler所拥有的sql字符串，可通过getValue("delegate.boundSql.sql")其中以.为分隔符并其中的属性必须是内部属性(区分大小写)。

MetaObject#setValue()

原理同MetaObject#getValue()方法

总结

结合实例以及源码帮助大家更透彻的了解Mybatis是如何进行分页操作的，欢迎指正