一个关于 联合查询union 和排序order by 的例子

其实还有更简单的方法

Title
还可以把个别的select命令用圆括号，这样，可以对每次查询以及最终结果做出limit和order by 设置。

(select * from tbl1 order by colA limit 10) union (select * from tbl2 order by colA limit 10) order by

coltimestamp limit 5

//查询有效期内的计划任务节点

string sql = "select     g.grade_name,   "

+ " '' as degree_name,"

+ " '' as reason, "

+ " 'project_node' as node_type, "

+ " '' as fact_node_name, "

+ " p.create_time, "

+ "    p.id,    p.project_id,     p.node_name,        p.deptids,"

+ "    p.node_begin_time,         p.node_end_time,    p.worklevel"

+ "    from     hrp_project_node p "

+ "    left join hrp_cycle_grade g on g.id = p.worklevel "

+ "    where p.state = 0  and  p.node_state <> 3 "

+ " and UNIX_TIMESTAMP(p.node_begin_time) < UNIX_TIMESTAMP('" + DateTime.Now.ToString() + "') "

//优先查出所有

//+ " and UNIX_TIMESTAMP(p.node_end_time) > UNIX_TIMESTAMP('" + DateTime.Now.ToString("yyyy-MM-dd HH:mm:ss") + "') "

+ " and (p.exeids like '" + userId + "%' or p.exeids like '%," + userId + "%') ";

//合并还未执行的任务节点

sql += " union all "

+ " select     g.grade_name,   "

+ " '' as degree_name,"

+ " '' as reason, "

+ " 'fact_node' as node_type, "

+ " '' as fact_node_name, "

+ " f.create_time, "

+ "    f.id,    f.project_id,     f.node_name,        f.deptids,"

+ "    f.fact_begin_time,         f.fact_end_time,    f.worklevel"

+ "    from     hrp_fact_node_execute f "

+ "    left join hrp_cycle_grade g on g.id = f.worklevel "

+ "    where f.state = 0  and  f.fact_state_id = 0 "

+ " and (f.exeids like '" + userId + "%' or f.exeids like '%," + userId + "%') ";

sql = string.Format("select * from ({0}) as t1  order by create_time desc", sql);

ps: 我记得这样的一篇帖子好像发表过, 奇怪为什么不找不到了??

当时还参考了别人的博客,

迷糊ing.... 唉, 被 google.com搞得心神不宁啊~~~~