Question

+------+------+
|DealId|Failcode|
+------+------+
|1     |A     |
+------+------+
|2     |A     |
+------+------+
|2     |A     |
+------+------+
|1     |B     |
+------+------+
|2     |B     |
+------+------+
|1     |A     |
+------+------+
|1     |A     |
+------+------+

我想选择使用不同的失败代码及其失败率多次出现的DealId

输出应类似于：

+------+------+-----+
|DealId|Fail  |count|
+------+------+-----+
|1     |A     |3    |
+------+------+-----+
|1     |B     |1    |
+------+------+-----+
|2     |A     |2    |
+------+------+-----+
|2     |B     |1    |
+------+------+-----+

Answer 1

使用聚合和group by

select DealId,Fail,count(*) as cnt
from tablename
group by DealId,Fail
having count(*) > 1;

Answer 2

您可以像下面一样使用WHERE EXIST来检查DealId在不同的Failcode中多次出现的情况。

要显示计数，您只需使用GROUP BY dealid,failcode

SELECT T1.dealid, 
       T1.failcode Fail, 
       Count(*)    AS CT 
FROM   YourTableName T1 
WHERE  EXISTS (SELECT 1 
               FROM   YourTableName T2 
               WHERE  T2.dealid = T1.dealid 
                      AND T2.failcode <> T1.failcode) 
GROUP  BY T1.dealid, 
          T1.failcode

注意：将YourTableName替换为实际的表名。

Answer 3

我想选择使用不同的失败代码及其失败率多次出现的DealId

我的解释是，您想要一种以上的失败率。您的样本数据未提供第一个条件过滤掉的交易ID的任何示例。但是，这个问题的答案似乎是：

select dealid, fail, cnt
from (select DealId, Fail, count(*) as cnt,
             count(*) over (partition by DealId) as cnt_fails
      from tablename
      group by DealId, Fail
     ) t
having cnt_fails > 1;

用于计算故障率的SQL编程

3 个答案: