Question

我的数据如下：

App ID  Ref Id  App Type  Reg Date
1       1       Main      2017-05-13
2       2       Sub       2017-05-14
3       1       Sub       2017-05-16
4       2       Main      2017-05-15
5       3       Main      2017-05-14
6       1       sub       2017-05-17

我想更改此表，如下所示。

App ID  Ref Id  App Type Reg Date
4       2       Main     2017-05-15
2       2       Sub      2017-05-14
5       3       Main     2017-05-14
1       1       Main     2017-05-13
6       1       sub      2017-05-17
3       1       Sub      2017-05-16

显示具有相同参考ID的内容，并且具有Main的内容位于顶部。最近注册的内容必须位于顶部。也就是说，我想创建一个层次结构。

Select * 
from table 
order by refId desc, App Type asc, Reg Date desc

我应该像上面那样配置查询吗？

Answer 1

我想你想要：

select t.*
from t
order by (select max(t2.regdate) from t t2 where t2.refid = t.refid) desc,
         t.refid,
         (t.apptype = 'Main') desc,
         t.regdate desc;

Answer 2

测试了它：

SELECT * 
FROM t 
JOIN (SELECT `Ref Id`, MAX(`Reg Date`) AS maxdate FROM t WHERE `App Type` = 'Main' GROUP BY 1) md USING(`Ref Id`) 
ORDER BY maxdate DESC, `Ref Id`, (`App Type` = 'Main') DESC;

+--------+--------+----------+------------+------------+
| Ref Id | App ID | App Type | Reg Date   | maxdate    |
+--------+--------+----------+------------+------------+
|      2 |      4 | Main     | 2017-05-15 | 2017-05-15 |
|      2 |      2 | Sub      | 2017-05-14 | 2017-05-15 |
|      3 |      5 | Main     | 2017-05-14 | 2017-05-14 |
|      1 |      1 | Main     | 2017-05-13 | 2017-05-13 |
|      1 |      6 | sub      | 2017-05-17 | 2017-05-13 |
|      1 |      3 | Sub      | 2017-05-16 | 2017-05-13 |
+--------+--------+----------+------------+------------+

我想使用orderby子句构造一个查询。请指教

2 个答案: