Hibernate与Jdbc的效率测试(ZT)

首先，建数据库表，采用MySQL数据库。

CREATE TABLE `t_jobstat` (
  `id` bigint(20) unsigned NOT NULL AUTO_INCREMENT,
  `jobNo` varchar(45) DEFAULT NULL,
  `clickSum` int(10) unsigned DEFAULT NULL,
  `showSum` int(10) unsigned DEFAULT NULL,
  `mailSum` int(10) unsigned DEFAULT NULL,
  `locationid` varchar(10) DEFAULT NULL,
  `logdate` char(8) DEFAULT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=7320077 DEFAULT CHARSET=gbk MAX_ROWS=500000000;

其次，插入大批量数据，本次测试，插入730万条数据。

 

一.Jdbc测试

主要方法如下：

1.数据访问层

   /**
 * 检索前1000条记录
 * @return
 * @throws SQLException
 */
public List<Jobstat> findAll() throws SQLException {
    String sql = "select * from t_jobstat where locationid like '%530%' limit 0,1000 ";

    // 执行查询
    ps = conn.prepareStatement(sql);
    ResultSet rs = ps.executeQuery();

    // 将查询结果组装至Asset实体对象
    Jobstat jobstat = null;
    List<Jobstat> list = new ArrayList<Jobstat>();

    while (rs.next()) {
        jobstat = new Jobstat();
        jobstat.setId(rs.getInt("id"));
        jobstat.setJobNo(rs.getString("jobNo"));
        jobstat.setClickSum(rs.getInt("clickSum"));
        jobstat.setShowSum(rs.getInt("showSum"));
        jobstat.setMailSum(rs.getInt("mailSum"));
        jobstat.setLocationId(rs.getString("locationId"));
        jobstat.setLogdate(rs.getString("logdate"));
        list.add(jobstat);
        jobstat = null;
    }

    return list;
}

    /**
	 * 检索前1000条记录
	 * @return
	 * @throws SQLException
	 */
	public List<Jobstat> findAll() throws SQLException {
		String sql = "select * from t_jobstat where locationid like '%530%' limit 0,1000 ";

		// 执行查询
		ps = conn.prepareStatement(sql);
		ResultSet rs = ps.executeQuery();

		// 将查询结果组装至Asset实体对象
		Jobstat jobstat = null;
		List<Jobstat> list = new ArrayList<Jobstat>();
		
		while (rs.next()) {
			jobstat = new Jobstat();
			jobstat.setId(rs.getInt("id"));
			jobstat.setJobNo(rs.getString("jobNo"));
			jobstat.setClickSum(rs.getInt("clickSum"));
			jobstat.setShowSum(rs.getInt("showSum"));
			jobstat.setMailSum(rs.getInt("mailSum"));
			jobstat.setLocationId(rs.getString("locationId"));
			jobstat.setLogdate(rs.getString("logdate"));
			list.add(jobstat);
			jobstat = null;
		}
		
		return list;
	}

2.Action层

  

             /**
     * 检索前1000条记录
     * @param mapping
     * @param form
     * @param request
     * @param response
     * @return
     * @throws Exception
     */
    public ActionForward findTop(ActionMapping mapping, ActionForm form,
            HttpServletRequest request, HttpServletResponse response) throws Exception{

        JobstatDAO dao = new JobstatDAO();
        long start = System.currentTimeMillis();
        List<Jobstat>list = dao.findAll();
        request.setAttribute("results", list);
        long end = System.currentTimeMillis();
        System.out.println("取出前1000条数据组装为对象,花费时间为"+(end-start)+"ms");
        return mapping.findForward("top");
}

             /**
	 * 检索前1000条记录
	 * @param mapping
	 * @param form
	 * @param request
	 * @param response
	 * @return
	 * @throws Exception
	 */
	public ActionForward findTop(ActionMapping mapping, ActionForm form,
			HttpServletRequest request, HttpServletResponse response) throws Exception{

    	JobstatDAO dao = new JobstatDAO();
    	long start = System.currentTimeMillis();
        List<Jobstat>list = dao.findAll();
        request.setAttribute("results", list);
        long end = System.currentTimeMillis();
        System.out.println("取出前1000条数据组装为对象,花费时间为"+(end-start)+"ms");
        return mapping.findForward("top");
}

 3.我们打开一个测试页面，http://localhost:8080/jdbctest/jobstat.do?p=findTop

 然后将默认浏览器设为能打开多Tab页的遨游，狂点测试页，最后数据取出速度稳定在16ms左右。

 

二.Hibernate  Test

1.DAO访问层

public static void main(String[] args) throws UnsupportedEncodingException{
    session = HibernateUtil.currentSession(); // 开启连接
    long start = System.currentTimeMillis();
    String queryString = "from Jobstat as js";
    Query queryObject = session.createQuery(queryString);
    int pageNo =0;
    pageNo = new Random().nextInt(10);
    queryObject.setFirstResult(1000*pageNo);
    queryObject.setMaxResults(1000*(pageNo+1));
    List list = queryObject.list();
    long end = System.currentTimeMillis();
    System.out.println("随机取出前1000条数据组装为对象,花费时间为"+(end-start)+"ms");
}

	public static void main(String[] args) throws UnsupportedEncodingException{
		session = HibernateUtil.currentSession(); // 开启连接
		long start = System.currentTimeMillis();
		String queryString = "from Jobstat as js";
		Query queryObject = session.createQuery(queryString);
		int pageNo =0;
		pageNo = new Random().nextInt(10);
     queryObject.setFirstResult(1000*pageNo);
     queryObject.setMaxResults(1000*(pageNo+1));
     List list = queryObject.list();
     long end = System.currentTimeMillis();
     System.out.println("随机取出前1000条数据组装为对象,花费时间为"+(end-start)+"ms");
	}

2进行数据量测试，当数据库中未建主键测试中，取出1000条数据花费时间为680ms，当主键索引存在后，取出1000条数据花费时间为530ms

 

综上测试，

 Hibernate作为一个O/R Mapping,比JDBC具备的优势有：

1.编程思想上，更加符合人的逻辑思维习惯，面向对象比面向过程更加容易理解，测试和维护

2.开发维护速度上，Hibernate显著的快，代码量显著小

3.通过Annotation进行数据库的字段加密

4.对Sql不熟的菜鸟来说可以自动调优

5.结合Spring，通过声明式事务可以省略事务的控制，事务以横切面形式出现

 

Jdbc比Hibernate具备的优势有：

1.大数据量访问时，Jdbc的效率显著快

2.直接操作数据库比较灵活

 

Hibernate比Jdbc慢的地方：

1.将Hql语句转化为标准Sql语句，花费时间可以忽略不急

2.将ResultSet中数据组装为实体对象，如果即所得数据量比较大(比如一次检索1000条数据)，这是效率就会降低，在组装实体对象花费时间较长

 

检索单表的1000条数据，Hibernate与Jdbc均不会超过1S钟.

 

from: http://www.iteye.com/topic/236784