Sql server 2008 SQL将公式应用于平均值

给定伪表：

+-----+---------------------+------+
| tag | data                | read |
+-----+---------------------+------+
| A   | 2013-10-10 15:00:00 | 1345 |
+-----+---------------------+------+
| A   | 2013-10-10 15:15:00 | 3454 |
+-----+---------------------+------+
| A   | 2013-10-10 15:30:00 | 2345 |
+-----+---------------------+------+
| A   | 2013-10-10 15:45:00 | 1132 |
+-----+---------------------+------+
| B   | 2013-10-10 15:00:00 | 6234 |
+-----+---------------------+------+
| B   | 2013-10-10 15:15:00 | 5432 |
+-----+---------------------+------+
| B   | 2013-10-10 15:30:00 | 4563 |
+-----+---------------------+------+
| B   | 2013-10-10 15:45:00 | 5432 |
+-----+---------------------+------+

是否可以仅使用SQL应用以下等式

例如：

result=AVG(A)-(AVG(B)+AVG(C))

或

---

按日期分组？

它应该计算结果

这里有一个演示

---

您可以分别为每个标记选择一个sum或avg，然后在每个查询上使用所需的操作进行选择

select (select SUM([read]) from table where tag = 'A') + 
  (select SUM([read]) from table where tag = 'B') 

---

它很好用。但必须删除“else 0”，因为它会影响平均值计算，所以

。。。平均值（标记为“A”然后读取结束时的情况）.

select (AVG(case when tag = 'A' then read end) + AVG(case when tag = 'B' then read end)) 'result', data
from TBL
group by data

select (select SUM([read]) from table where tag = 'A') + 
  (select SUM([read]) from table where tag = 'B')