hiber nate批量提交数据

Transaction transaction = null;
	try {
		transaction = session.beginTransaction();
		int p = 0;
		for(int t=0;t<cardnum;t++) {
		Date dates = new Date();
		SimpleDateFormat df = new SimpleDateFormat("yyMMdd");
							
		String pwd="";
		String money = money1;
		String out = "";
		String str1 = df.format(dates);
												
		String str3 = String.valueOf(start);
		for(int i=1;i<6;i++){
			out+=str1.substring(i-1, i)+str2.substring(i-1, i);
				}
								
				out=out+str1.substring(5, 6)+str3;		
								
				MCard mcard = new MCard();
				mcard.setCardNum(out);
				mcard.setCardSec(pwd);
				mcard.setCardMoney(money);
				mcard.setRegDate(DateUtils.stringToDate(stime));
				mcard.setBecomeDate(DateUtils.stringToDate(otime));
				                session.save(mcard);
				                p++;
				                if (p == 1000) {
				                    transaction.commit();
				                    log.debug("save " + 100 + "data successful");
				                    p = 0;
				                    session.flush(); session.clear(); 
				                    transaction = session.beginTransaction();
				               }
				            }
				            if (p != 0) {
				                transaction.commit();
				                session.flush(); session.clear(); 
				            }
				            // session.
				            log.debug("save larger successful");
				           
				        } catch (Exception re) {
				            log.error("save failed", re);
				            if (transaction != null && transaction.isActive())
				                transaction.rollback();
				           
				        }
				        transaction.commit();

 今天由于项目的需要，需要一次性插入上万条数据。开始用每生成一条数据就插一次，速度慢而且到最后是内存益出。然后有用了JDBC来，不过后面觉的还的用HIBERNATE，最后尝试1000条数据提交一次事务。

最后测试，插入十万条数据用了61秒时间。