一、Mybatis插件机制简介
Mybatis在四大组件(Executor,StatementHandler,ParameterHandler,ResultSetHandler)处都提供了插件机制。Mybatis通过对四大核心对象进行拦截,并增强对象的功能。本质上是通过动态代理来实现的,也就是说实际上,四大组件的对象最后都是代理对象。
1. Mybatis允许拦截的方法如下
-
执行器Executor的update, query, commit, rollback等方法
-
SQL语法构建器StatementHandler的prepare, parameterize, batch, update, query等方法
-
参数处理器ParameterHandler的getParameterObject, setParameterObject等方法
-
结果处理器ResultSetHandler的handleResultSets, handlerOutputPameters等方法
2. Mybatis插件原理
在四大组件创建时,每个对象创建之后不是直接返回,而是通过调用interceptorChain.pluginAll(handler)之后再返回。
public ParameterHandler newParameterHandler(MappedStatement mappedStatement, Object parameterObject, BoundSql boundSql) {
ParameterHandler parameterHandler = mappedStatement.getLang().createParameterHandler(mappedStatement, parameterObject, boundSql);
parameterHandler = (ParameterHandler) interceptorChain.pluginAll(parameterHandler);
return parameterHandler;
}
interceptorChain保存了所有的拦截器,这是在Mybatis初始化的时候创建的。
pluginAll方法就是遍历interceptorChain中的每一个inteceptor,执行它的plugin方法:
public Object pluginAll(Object target) {
for (Interceptor interceptor : interceptors) {
target = interceptor.plugin(target);
}
return target;
}
二、实现分页插件
如果我们要拦截四大组件的某个方法,可以通过@Intercepts, @Signature注解来实现,例如拦截Executor对象的query方法可以这么实现:
@Intercepts({
@Signature(
type=Executor.class,
method="query",
args={MappedStatement.class, Object.class, RowBounds.class, ResultHandler.class}
)
})
public class InterceptExecutorQuery implements Interceptor{
//......
}
然后在sqlMapConfig.xml中配置这个插件。
我们来实现一个Mybatis的分页插件,这个插件的用法是这样的:
public Page<GJAJEntity> findPage(GJAJEntity leaveEntity, int pageNum) {
PageHelper.startPage(pageNum, Constant.pageSize);
caseDao.queryList(leaveEntity);
return PageHelper.endPage();
}
- 调用PageHelper.startPage(pageNum, pageSize)方法,开始分页
- 执行查询操作
- 调用PageHelper.endPage()方法,结束分页,返回分页对象
1. 首先就是PageHelper这个类要实现Interceptor接口,并且标记拦截StatementHandler的prepare方法以及ResultSetHandler的handleResultSets方法
@Intercepts(
{
@Signature(type = StatementHandler.class, method = "prepare", args = {Connection.class, Integer.class}),
@Signature(type = ResultSetHandler.class, method = "handleResultSets", args = {Statement.class})
}
)
public class PageHelper implements Interceptor {
@Override
public Object intercept(Invocation invocation) throws Throwable {
return null;
}
@Override
public Object plugin(Object target) {
return null;
}
@Override
public void setProperties(Properties properties) {
}
}
2. 实现startPage和endPage方法
/**
* 开始分页
* @param pageNum
* @param pageSize
*/
public static void startPage(int pageNum, int pageSize) {
localPage.set(new Page(pageNum, pageSize));
}
/**
* 结束分页并返回结果,该方法必须被调用,否则localPage会一直保存下去,直到下一次startPage
* @return
*/
public static Page endPage() {
Page page = localPage.get();
localPage.remove();
return page;
}
这个localPage对象其实就是一个ThreadLocal对象,里面存储的是一个Page对象,Page对象里面保存了查询结果:
public static final ThreadLocal<Page> localPage = new ThreadLocal<Page>();
3. 实现Interceptor接口的plugin方法
/**
* 只拦截这两种类型的
* <br>StatementHandler
* <br>ResultSetHandler
* @param target
* @return
*/
@Override
public Object plugin(Object target) {
if (target instanceof StatementHandler || target instanceof ResultSetHandler) {
return Plugin.wrap(target, this);
} else {
return target;
}
}
4. 实现Interceptor接口的intercept方法
@Override
public Object intercept(Invocation invocation) throws Throwable {
if (localPage.get() == null) {
return invocation.proceed();
}
if (invocation.getTarget() instanceof StatementHandler) {
StatementHandler statementHandler = (StatementHandler) invocation.getTarget();
MetaObject metaStatementHandler = SystemMetaObject.forObject(statementHandler);
// 分离代理对象链(由于目标类可能被多个拦截器拦截,从而形成多次代理,通过下面的两次循环
// 可以分离出最原始的的目标类)
while (metaStatementHandler.hasGetter("h")) {
Object object = metaStatementHandler.getValue("h");
metaStatementHandler = SystemMetaObject.forObject(object);
}
// 分离最后一个代理对象的目标类
while (metaStatementHandler.hasGetter("target")) {
Object object = metaStatementHandler.getValue("target");
metaStatementHandler = SystemMetaObject.forObject(object);
}
MappedStatement mappedStatement = (MappedStatement) metaStatementHandler.getValue("delegate.mappedStatement");
//分页信息if (localPage.get() != null) {
Page page = localPage.get();
//Mybatis中最终执行的sql封装成了一个BoundSql对象,先获取到这个BoundSql对象
BoundSql boundSql = (BoundSql) metaStatementHandler.getValue("delegate.boundSql");
// 然后获取到sql语句。分页参数作为参数对象parameterObject的一个属性
String sql = boundSql.getSql();
// 重写sql
String pageSql = buildPageSqlMysql(sql, page);
//重写分页sql ,再把这个sql语句赋值给BoundSql对象中的sql属性
metaStatementHandler.setValue("delegate.boundSql.sql", pageSql);
Connection connection = (Connection) invocation.getArgs()[0];
// 重设分页参数里的总页数等
setPageParameter(sql, connection, mappedStatement, boundSql, page);
// 将执行权交给下一个拦截器
return invocation.proceed();
} else if (invocation.getTarget() instanceof ResultSetHandler) {
//拦截ResultSetHandler的handleResultSets方法,直接返回给Page
Object result = invocation.proceed();
Page page = localPage.get();
page.setResult((List) result);
return result;
}
return null;
}
/**
* 修改原SQL为分页SQL mysql
* @param sql
* @param page
* @return
*/
private String buildPageSqlMysql(String sql, Page page) {
StringBuilder pageSql = new StringBuilder(200);
pageSql.append(sql);
pageSql.append(" LIMIT ").append(page.getStartRow()).append(",").append(page.getEndRow()-page.getStartRow());
return pageSql.toString();
}
/**
* 获取总记录数
* @param sql
* @param connection
* @param mappedStatement
* @param boundSql
* @param page
*/
private void setPageParameter(String sql, Connection connection, MappedStatement mappedStatement,
BoundSql boundSql, Page page) {
// 记录总记录数
String countSql = "select count(0) from (" + sql + ")as total";
PreparedStatement countStmt = null;
ResultSet rs = null;
try {
countStmt = connection.prepareStatement(countSql);
BoundSql countBS = new BoundSql(mappedStatement.getConfiguration(), countSql,
boundSql.getParameterMappings(), boundSql.getParameterObject());
setParameters(countStmt, mappedStatement, countBS, boundSql.getParameterObject());
rs = countStmt.executeQuery();
int totalCount = 0;
if (rs.next()) {
totalCount = rs.getInt(1);
}
page.setTotal(totalCount);
int totalPage = totalCount / page.getPageSize() + ((totalCount % page.getPageSize() == 0) ? 0 : 1);
page.setPages(totalPage);
} catch (SQLException e) {
logger.error("Ignore this exception", e);
} finally {
try {
if(rs != null){
rs.close();
}
} catch (SQLException e) {
logger.error("Ignore this exception", e);
}
try {
countStmt.close();
} catch (SQLException e) {
logger.error("Ignore this exception", e);
}
}
}
/**
* 代入参数值
* @param ps
* @param mappedStatement
* @param boundSql
* @param parameterObject
* @throws SQLException
*/
private void setParameters(PreparedStatement ps, MappedStatement mappedStatement, BoundSql boundSql,
Object parameterObject) throws SQLException {
ParameterHandler parameterHandler = new DefaultParameterHandler(mappedStatement, parameterObject, boundSql);
parameterHandler.setParameters(ps);
}
5. 使用PageHelper
public Page<GJAJEntity> findPage(GJAJEntity leaveEntity, int pageNum) {
PageHelper.startPage(pageNum, Constant.pageSize);
caseDao.queryList(leaveEntity);
return PageHelper.endPage();
}
可以看到只要执行startPage和endPage两个方法,PageHelper本身不需要和数据发生耦合。