PageHelper是怎么分页的

1、前言

PageHelper是mybatis 提供的分页插件，通过PageHelper.startPage(pageNo,pageLimit)就可以帮我们实现分页，目前支持Oracle,Mysql,MariaDB,SQLite,Hsqldb,PostgreSQL六种数据库。

pom依赖：

<dependency>
    <groupId>com.github.pagehelper</groupId>
    <artifactId>pagehelper</artifactId>
    <version>5.3.1</version>
</dependency>

yml配置：

pagehelper:
  helperDialect: mysql #数据库类型，不指定的话会解析 datasource.url进行配置
  supportMethodsArguments: true
  params: count=countSql 

使用示例：

@Service
public class UserService {

    @Autowired
    private UserMapper userMapper;
 /**
     * 依据用户昵称进行模糊分页查询
     *
     * @param name
     * @param page
     * @param limit
     * @return
     */
    public PageInfo<User> findPageUsersByName(String name, int page, int limit) {
        PageHelper.startPage(page, limit);
        List<User> users = userMapper.selectByName(name);
        PageInfo<User> pageUsers = new PageInfo<>(users);
        return pageUsers;
    }
}

2、基本流程

1、PageHelper向 Mybatis 注册处理分页和 count 的拦截器 PageInterceptor

2、通过 PageHelper.startPage() 方法把分页相关的参数放到 ThreadLcoal 中

3、Mybatis 执行 SQL 过程中会调用拦截器

• 3.1、根据查询 SQL 构建 count SQL
• 3.2、从 ThreadLcoal 拿出分页信息，在查询 SQL 后面拼接 limit ?, ?
• 3.3、清空 ThreadLcoal

4、 使用 Page 创建 PageInfo 对象

3、源码分析

3.1、存储分页参数

直接看PageHelper.startPage，startPage() 方法会构建一个 Page 对象，存储分页相关的参数、设置，最后调用 setLocalPage(page)，将其放入ThreadLocal。由于ThreadLocal 每次查询后都会被remove掉，所以一次mapper查询对应一次PageHelper.startPage。

ThreadLocal原理可以看这篇ThreadLocal源码及其内存泄漏问题

    public static <E> Page<E> startPage(int pageNum, int pageSize, boolean count, Boolean reasonable, Boolean pageSizeZero) {
        Page<E> page = new Page(pageNum, pageSize, count);
        page.setReasonable(reasonable);
        page.setPageSizeZero(pageSizeZero);
        Page<E> oldPage = getLocalPage();
        if (oldPage != null && oldPage.isOrderByOnly()) {
            page.setOrderBy(oldPage.getOrderBy());
        }

        setLocalPage(page);
        return page;
    }

3.2、改造SQL

PageInterceptor 实现了 org.apache.ibatis.plugin.Interceptor 接口，MyBatis 底层查询其实就是借助SqlSession调用Executor#query，mybatis 在执行查询方法的时候（method = “query”）会调用本拦截器。

/**
 * Mybatis - 通用分页拦截器
 */
@SuppressWarnings({"rawtypes", "unchecked"})
@Intercepts(
        {
                @Signature(type = Executor.class, method = "query", args = {MappedStatement.class, Object.class, RowBounds.class, ResultHandler.class}),
                @Signature(type = Executor.class, method = "query", args = {MappedStatement.class, Object.class, RowBounds.class, ResultHandler.class, CacheKey.class, BoundSql.class}),
        }
)
public class PageInterceptor implements Interceptor {
    // 默认为 Pagehelper
    private volatile Dialect dialect;
    // count 方法的后缀
    private String countSuffix = "_COUNT";
    //  count 查询的缓存，只用于
    // 本例中 key 为 com.example.pagehelper.dao.UserMapper.selectUsers_COUNT
    protected Cache<String, MappedStatement> msCountMap = null;
    // 
    private String default_dialect_class = "com.github.pagehelper.PageHelper";
    ...
}    

进入PageInterceptor.intercept()方法，

@Override
public Object intercept(Invocation invocation) throws Throwable {
    try {
        // 获取方法参数
        Object[] args = invocation.getArgs();
        MappedStatement ms = (MappedStatement) args[0];
        Object parameter = args[1];
        RowBounds rowBounds = (RowBounds) args[2];
        ResultHandler resultHandler = (ResultHandler) args[3];
        Executor executor = (Executor) invocation.getTarget();
        CacheKey cacheKey;
        BoundSql boundSql;
        //由于逻辑关系，只会进入一次
        if (args.length == 4) {
            //4 个参数时
            
            // 拿到原始的查询 SQL
            boundSql = ms.getBoundSql(parameter);
            cacheKey = executor.createCacheKey(ms, parameter, rowBounds, boundSql);
        } else {
            //6 个参数时
            cacheKey = (CacheKey) args[4];
            boundSql = (BoundSql) args[5];
        }
        checkDialectExists();
        //对 boundSql 的拦截处理
        // 实际什么都没做，原样返回了
        if (dialect instanceof BoundSqlInterceptor.Chain) {
            boundSql = ((BoundSqlInterceptor.Chain) dialect).doBoundSql(BoundSqlInterceptor.Type.ORIGINAL, boundSql, cacheKey);
        }
        List resultList;
        //调用方法判断是否需要进行分页，如果不需要，直接返回结果
        if (!dialect.skip(ms, parameter, rowBounds)) {
            //判断是否需要进行 count 查询
            if (dialect.beforeCount(ms, parameter, rowBounds)) {
                // 查询总数
                // 见 PageInterceptor.count()
                Long count = count(executor, ms, parameter, rowBounds, null, boundSql);
                //处理查询总数，返回 true 时继续分页查询，false 时直接返回
                if (!dialect.afterCount(count, parameter, rowBounds)) {
                    //当查询总数为 0 时，直接返回空的结果
                    return dialect.afterPage(new ArrayList(), parameter, rowBounds);
                }
            }
            // 执行分页查询
            resultList = ExecutorUtil.pageQuery(dialect, executor,
                    ms, parameter, rowBounds, resultHandler, boundSql, cacheKey);
        } else {
            //rowBounds用参数值，不使用分页插件处理时，仍然支持默认的内存分页
            resultList = executor.query(ms, parameter, rowBounds, resultHandler, cacheKey, boundSql);
        }
        // 将count、分页 信息放入 ThreadLocal
        return dialect.afterPage(resultList, parameter, rowBounds);
    } finally {
        if(dialect != null){
            dialect.afterAll();
        }
    }
}

count
PageInterceptor.count()

private Long count(Executor executor, MappedStatement ms, Object parameter,
                    RowBounds rowBounds, ResultHandler resultHandler,
                    BoundSql boundSql) throws SQLException {
    // countMsId = "com.example.pagehelper.dao.UserMapper.selectUsers_COUNT"
    String countMsId = ms.getId() + countSuffix;
    Long count;
    //先判断是否存在手写的 count 查询
    MappedStatement countMs = ExecutorUtil.getExistedMappedStatement(ms.getConfiguration(), countMsId);
    if (countMs != null) {
        // 直接执行手写的 count 查询
        count = ExecutorUtil.executeManualCount(executor, countMs, parameter, boundSql, resultHandler);
    } else {
        // 先从缓存中查
        if (msCountMap != null) {
            countMs = msCountMap.get(countMsId);
        }
        // 缓存中没有，然后自动创建，并放入缓存
        if (countMs == null) {
            //根据当前的 ms 创建一个返回值为 Long 类型的 ms
            countMs = MSUtils.newCountMappedStatement(ms, countMsId);
            if (msCountMap != null) {
                // 放入缓存
                msCountMap.put(countMsId, countMs);
            }
        }
        // 执行 count 查询
        count = ExecutorUtil.executeAutoCount(this.dialect, executor, countMs, parameter, 
                                              boundSql, rowBounds, resultHandler);
    }
    return count;
}

ExecutorUtil.executeAutoCount()

public static Long executeAutoCount(Dialect dialect, Executor executor, MappedStatement countMs,
                                    Object parameter, BoundSql boundSql,
                                    RowBounds rowBounds, ResultHandler resultHandler) throws SQLException {
    Map<String, Object> additionalParameters = getAdditionalParameter(boundSql);
    //创建 count 查询的缓存 key
    CacheKey countKey = executor.createCacheKey(countMs, parameter, RowBounds.DEFAULT, boundSql);
    //调用方言获取 count sql：SELECT count(0) FROM user
    // 见 PageHelper.getCountSql()
    String countSql = dialect.getCountSql(countMs, boundSql, parameter, rowBounds, countKey);
    //countKey.update(countSql);
    BoundSql countBoundSql = new BoundSql(countMs.getConfiguration(), countSql, boundSql.getParameterMappings(), parameter);
    //当使用动态 SQL 时，可能会产生临时的参数，这些参数需要手动设置到新的 BoundSql 中
    for (String key : additionalParameters.keySet()) {
        countBoundSql.setAdditionalParameter(key, additionalParameters.get(key));
    }
    //对 boundSql 的拦截处理
    if (dialect instanceof BoundSqlInterceptor.Chain) {
        countBoundSql = ((BoundSqlInterceptor.Chain) dialect).doBoundSql(BoundSqlInterceptor.Type.COUNT_SQL, countBoundSql, countKey);
    }
    //执行 count 查询
    Object countResultList = executor.query(countMs, parameter, RowBounds.DEFAULT, resultHandler, countKey, countBoundSql);
    Long count = (Long) ((List) countResultList).get(0);
    return count;
}

3.3、分页查询

ExecutorUtil.pageQuery

public static <E> List<E> pageQuery(Dialect dialect, Executor executor, MappedStatement ms, Object parameter,
                                    RowBounds rowBounds, ResultHandler resultHandler,
                                    BoundSql boundSql, CacheKey cacheKey) throws SQLException {
    //判断是否需要进行分页查询
    if (dialect.beforePage(ms, parameter, rowBounds)) {
        //生成分页的缓存 key
        CacheKey pageKey = cacheKey;
        //处理参数对象
        parameter = dialect.processParameterObject(ms, parameter, boundSql, pageKey);
        //调用方言获取分页 sql，这里是重点，是添加 limit 的地方
        // pageSql = select id, name from user LIMIT ?, ? 
        String pageSql = dialect.getPageSql(ms, boundSql, parameter, rowBounds, pageKey);
        BoundSql pageBoundSql = new BoundSql(ms.getConfiguration(), pageSql, boundSql.getParameterMappings(), parameter);

        Map<String, Object> additionalParameters = getAdditionalParameter(boundSql);
        //设置动态参数
        for (String key : additionalParameters.keySet()) {
            pageBoundSql.setAdditionalParameter(key, additionalParameters.get(key));
        }
        //对 boundSql 的拦截处理
        if (dialect instanceof BoundSqlInterceptor.Chain) {
            pageBoundSql = ((BoundSqlInterceptor.Chain) dialect).doBoundSql(BoundSqlInterceptor.Type.PAGE_SQL, pageBoundSql, pageKey);
        }
        //执行分页查询
        return executor.query(ms, parameter, RowBounds.DEFAULT, resultHandler, pageKey, pageBoundSql);
    } else {
        //不执行分页的情况下，也不执行内存分页
        return executor.query(ms, parameter, RowBounds.DEFAULT, resultHandler, cacheKey, boundSql);
    }
}

3.4、使用 Page 创建 PageInfo 对象

最后使用PageInfo<User> pageUsers = new PageInfo<>(users)获取到分页信息