jeesite使用心得（二）

按照上一篇的内容，有一些缺陷的地方。
分页对象传什么都可以，但是返回的是list>集合的话，分页方法就是无效了。

这里把我写的贴一下:

Controller:

// 带条件分页
		Integer pageNo = null;
		Integer pageSize = null;
		if (request.getParameter("pageNo") != null) {
			pageNo = Integer.parseInt((String) request.getParameter("pageNo"));
		} else {
			page.setPageNo(1);
			pageNo = 1;
		}
		if (request.getParameter("pageSize") != null) {
			pageSize = Integer.parseInt((String) request.getParameter("pageSize"));
		} else {
			pageSize = 30;
		}
		page.setPageNo(pageNo);
		page.setPageSize(pageSize);
		//根据条件查到的集合
		List> list = busMddkApplicationService.findPage(page, params);
		page.setList(list);
		model.addAttribute("page", page);

获取前台传过来的页码和分页数，这样再set到page对象中，前台就可以显示正常的分页和页码了。

jsp页面：

jsp页面，没改过，就用他生成的就可以。


再说一下shiro的自定义标签：

这个标签的意思是有任一权限就可以

这个标签的意思是有某个权限

Controller:

@RequiresPermissions(value= {"contract:busMddkApplication:active","contract:busMddkApplication:edit","contract:busMddkApplication:add","contract:busMddkApplication:info"})
@RequestMapping(value = "form")
	protected String form(HttpServletRequest request, BusMddkApplicationVo busMddkApplicationVo2, Model model,
			HttpServletResponse response, RedirectAttributes redirectAttributes) { }

上面的意思是有任一权限就可以进入这个方法。
这是说一下分页的问题,可以用他的page对象，会进入pageInterceptor的拦截器。感觉他查询总数的方法有点问题。贴一下原来的和我改动的代码。


//得到总记录数

common.persistence.interceptor.SQLHelper包下面，在PaginationInterceptor里边先调用 

page.setCount(SQLHelper.getCount(originalSql, null, mappedStatement, parameterObject, boundSql, log));

贴一下他原来的getCount()方法。

public static int getCount(final String sql, final Connection connection,
    							final MappedStatement mappedStatement, final Object parameterObject,
    							final BoundSql boundSql, Log log) throws SQLException {
    	String dbName = Global.getConfig("jdbc.type");
		final String countSql;
		if("oracle".equals(dbName)){
			countSql = "select count(1) from (" + sql + ") tmp_count";
			//需要优化的sql
		}else{
			countSql = "select count(1) from (" + removeOrders(sql) + ") tmp_count";
		}
        Connection conn = connection;
        PreparedStatement ps = null;
        ResultSet rs = null;
        try {
        	if (log.isDebugEnabled()) {
                log.debug("COUNT SQL: " + StringUtils.replaceEach(countSql, new String[]{"\n","\t"}, new String[]{" "," "}));
            }
        	if (conn == null){
        		conn = mappedStatement.getConfiguration().getEnvironment().getDataSource().getConnection();
            }
        	ps = conn.prepareStatement(countSql);
            BoundSql countBS = new BoundSql(mappedStatement.getConfiguration(), countSql,
                    boundSql.getParameterMappings(), parameterObject);
            //解决MyBatis 分页foreach 参数失效 start
			if (Reflections.getFieldValue(boundSql, "metaParameters") != null) {
				MetaObject mo = (MetaObject) Reflections.getFieldValue(boundSql, "metaParameters");
				Reflections.setFieldValue(countBS, "metaParameters", mo);
			}
			//解决MyBatis 分页foreach 参数失效 end 
            SQLHelper.setParameters(ps, mappedStatement, countBS, parameterObject);
            rs = ps.executeQuery();
            int count = 0;
            if (rs.next()) {
                count = rs.getInt(1);
            }
            return count;
        } finally {
            if (rs != null) {
                rs.close();
            }
            if (ps != null) {
            	ps.close();
            }
            if (conn != null) {
            	conn.close();
            }
        }
    }

      /** 
     * 去除hql的orderBy子句。 
     * @param hql 
     * @return 
     */  
    @SuppressWarnings("unused")
	private static String removeOrders(String qlString) {  
        Pattern p = Pattern.compile("order\\s*by[\\w|\\W|\\s|\\S]*", Pattern.CASE_INSENSITIVE);  
        Matcher m = p.matcher(qlString);  
        StringBuffer sb = new StringBuffer();  
        while (m.find()) {  
            m.appendReplacement(sb, "");  
        }
        m.appendTail(sb);
        return sb.toString();  
    }

 这样生成的sql就是正常的条件查询语句外面包了一层select count(1) from(原sql)。
    但是在实际使用过程中，数据量比较大(百万级别),并且查询条件和参数很多的时候,这样会很慢。不如直接select count(1) from (***) 这样的快。
    所以改写了他的getCount()方法,在需要优化的sql的where后边加上'needOptimization' = 'needOptimization'。

public static int getCount(final String sql, final Connection connection,
    							final MappedStatement mappedStatement, final Object parameterObject,
    							final BoundSql boundSql, Log log) throws SQLException {
    	String dbName = Global.getConfig("jdbc.type");
		final String countSql;
		if("oracle".equals(dbName)){
			countSql = "select count(1) from (" + sql + ") tmp_count";
			//需要优化的sql
		}else if(sql.contains("needOptimization")){
			countSql = "select count(1) from "+ removeContent(sql);
		}else{
			countSql = "select count(1) from (" + removeOrders(sql) + ") tmp_count";
		}
        Connection conn = connection;
        PreparedStatement ps = null;
        ResultSet rs = null;
        try {
        	if (log.isDebugEnabled()) {
                log.debug("COUNT SQL: " + StringUtils.replaceEach(countSql, new String[]{"\n","\t"}, new String[]{" "," "}));
            }
        	if (conn == null){
        		conn = mappedStatement.getConfiguration().getEnvironment().getDataSource().getConnection();
            }
        	ps = conn.prepareStatement(countSql);
            BoundSql countBS = new BoundSql(mappedStatement.getConfiguration(), countSql,
                    boundSql.getParameterMappings(), parameterObject);
            //解决MyBatis 分页foreach 参数失效 start
			if (Reflections.getFieldValue(boundSql, "metaParameters") != null) {
				MetaObject mo = (MetaObject) Reflections.getFieldValue(boundSql, "metaParameters");
				Reflections.setFieldValue(countBS, "metaParameters", mo);
			}
			//解决MyBatis 分页foreach 参数失效 end 
            SQLHelper.setParameters(ps, mappedStatement, countBS, parameterObject);
            rs = ps.executeQuery();
            int count = 0;
            if (rs.next()) {
                count = rs.getInt(1);
            }
            return count;
        } finally {
            if (rs != null) {
                rs.close();
            }
            if (ps != null) {
            	ps.close();
            }
            if (conn != null) {
            	conn.close();
            }
        }
    }

     /**
     * 将第一个from前的内容删除
     * @Description:TODO
     * @author:gmwang
     * @time:2017年8月29日 上午10:16:28
     */
    private static String removeContent(String sql) {
    	String sqlAfter = StringUtils.substringAfter(sql,"FROM");
		return sqlAfter;
	}

如果原来是select count(1) from (
select a.*,b.* from aa a 
left join bb b on a.id = b.id
)
现在就改成了 select count(1) from aa a left join bb b on a.id = b.id
原来的sql执行24秒,现在大约是6秒。