Question

假设我的condition表格中包含以下列：

+----+----------+-----------+
| ID |   Name   | Condition |
+----+----------+-----------+
|  1 | Value 1  | NULL      |
|  4 | Value 2  | NULL      |
| 13 | Value 3  | NULL      |
|  8 | Value 4  | 1         |
| 10 | Value 5  | 1         |
| 12 | Value 6  | 1         |
| 17 | Value 7  | 1         |
| 18 | Value 8  | 1         |
| 20 | Value 9  | 1         |
| 21 | Value 10 | 1         |
|  6 | Value 11 | 4         |
| 11 | Value 12 | 4         |
| 16 | Value 13 | 4         |
| 14 | Value 14 | 13        |
| 19 | Value 15 | 13        |
+----+----------+-----------+

现在，我希望他们订购ID和Condition按顺序逐个排序：

+----+----------+-----------+
| ID |   Name   | Condition |
+----+----------+-----------+
|  1 | Value 1  | NULL      |
|  8 | Value 4  | 1         |
| 10 | Value 5  | 1         |
| 12 | Value 6  | 1         |
| 17 | Value 7  | 1         |
| 18 | Value 8  | 1         |
| 20 | Value 9  | 1         |
| 21 | Value 10 | 1         |
|  4 | Value 2  | NULL      |
|  6 | Value 11 | 4         |
| 11 | Value 12 | 4         |
| 16 | Value 13 | 4         |
| 13 | Value 13 | NULL      |
| 14 | Value 14 | 13        |
| 19 | Value 15 | 13        |
+----+----------+-----------+

我该怎么做？

Answer 1

这是一个奇怪的排序顺序。因此，Condition中不存在ID的记录应先由Condition然后ID排序，而Condition中包含ID的记录应按ID排序在Condition中具有该ID的记录之前。一种方法是使用上面提到的sortkeys选择这两个集合，并在两个集合上使用union all。

select id, name, condition
from
(
  select 
    id, 
    name, 
    condition, 
    id as sortkey1, 
    1 as sortkey2, 
    null as sortkey3
  from mytable 
  where id in (select condition from mytable)
  union all
  select 
    id, 
    name, 
    condition, 
    condition as sortkey1, 
    2 as sortkey2, 
    id as sortkey3
  from mytable 
  where id not in (select condition from mytable)
)
order by sortkey1, sortkey2, sortkey3;

逐列ORDER BY具有序列

1 个答案: