Question

有没有一种方法可以执行group by并将非值用于列（如果有）。即

a  | b | c | d    | e       | f                   |
---------------------------------------------------
1  | 2 | 3 | x    | test1   | 2019-07-01 07:17:01 |
1  | 2 | 3 | NULL | test2   | 2019-07-01 10:23:11 |
1  | 2 | 3 | NULL | test3   | 2019-07-01 22:00:51 |
1  | 2 | 7 | NULL | testTet | 2019-07-01 23:00:00 |

在上面的例子中，如果d出现a=1,b=2,c=3，它将始终为x，否则可以为null。所以我的查询就像

select a,
       b,
       c,
       d,
       count(distinct e) as something
from tableX
where f between '2019-07-01 00:00:00' and '2019-07-01 23:59:59.999'
group by a,
         b,
         c,
         d

结果将是：

a  | b | c | d    | something |
------------------------------|
1  | 2 | 3 | x    | 1         |
1  | 2 | 3 | NULL | 2         |
1  |2  | 7 | NULL | 1         |

如果可以的话，这会很棒（因为对于每个group by组合，我都知道它为null或存在的唯一值）：

a  | b | c | d    | something |
------------------------------|
1  | 2 | 3 | x    | 3         |
1  | 2 | 7 | NULL | 1         |

Answer 1

从您的样本数据中，我认为您不需要d子句中的group by。
因此获得其最大值：

select 
  a, b, c, 
  max(d) d,
  count(distinct e) as something
from tableX
where f between '2019-07-01 00:00:00' and '2019-07-01 23:59:59.999'
group by a, b, c

Answer 2

尝试如下

with cte as (select a,
       b,
       c,
       d,
       count(distinct e) as something
from tableX
where f between '2019-07-01 00:00:00' and '2019-07-01 23:59:59.999'
group by a,
         b,
         c,
         d) select a,b,c,max(d) as d,sum(something) from cte group by a,b,c

分组并选择非null值（如果存在）

2 个答案: