如果没有找到行，则为 CASE 语句 |甲骨文

如何解决如果没有找到行，则为 CASE 语句 |甲骨文

我有一个包含以下数据的示例表。我正在使用 Oracle。

ID  Sales  STATUS
1    100   XX
1    50    XX
1    40    YY

我必须对具有不同状态的 ID 列上的数据进行分组。我正在使用以下 CASE 查询

SELECT 
ID,sum(sales) SALES,CASE 
WHEN STATUS='XX' THEN 'POSITIVE'
WHEN STATUS='YY' THEN 'NEGATIVE'
END INDICATOR
from 
sample_table
group by 
ID,CASE 
WHEN STATUS='XX' THEN 'POSITIVE'
WHEN STATUS='YY' THEN 'NEGATIVE'
END;

上面的语句给了我正确的输出

ID    SALES    INDICATOR
1     150      POSITIVE
1     40       NEGATIVE

现在如果我将数据更改为以下

ID  Sales  STATUS
1    100   XX
1    50    XX
1    40    XX

上面使用的查询仅给出 POSITIVE 指标行，因为表中没有 YY STATUS。即使同一个 ID 没有状态 YY，我也想显示 0 SALES。

OUTPUT
ID   SALES   STATUS
1    190     POSITIVE

Desired output
ID   SALES   STATUS
1     190     POSITIVE
1     0       NEGATIVE

是否有可能获得所需的输出？对于我的数据来说，这是一种罕见的情况，但我仍然想处理它。提前感谢您的帮助。

解决方法

您可以先将 PIVOT，然后将 UNPIVOT 与 NVL 结合使用，以便您始终为每个 ID 获得一个正面和负面的指示符。

案例 1

WITH
    sample_table (ID,Sales,STATUS)
    AS
        (SELECT 1,100,'XX' FROM DUAL
         UNION ALL
         SELECT 1,50,40,'YY' FROM DUAL)
SELECT *
  FROM (SELECT id,NVL (indicator_pos,0) AS indicator_pos,NVL (indicator_neg,0) AS indicator_neg
          FROM (SELECT ID,SALES,CASE
                           WHEN STATUS = 'XX' THEN 'POSITIVE'
                           WHEN STATUS = 'YY' THEN 'NEGATIVE'
                       END    INDICATOR
                  FROM sample_table)
               PIVOT (SUM (sales)
                     FOR INDICATOR
                     IN ('POSITIVE' AS indicator_pos,'NEGATIVE' AS indicator_neg)))
       UNPIVOT (sales FOR INDICATOR IN (indicator_pos AS 'POSITIVE',indicator_neg AS 'NEGATIVE'));

   ID    INDICATOR    SALES
_____ ____________ ________
    1 POSITIVE          150
    1 NEGATIVE           40

案例 2

WITH
    sample_table (ID,'XX' FROM DUAL)
SELECT *
  FROM (SELECT id,indicator_neg AS 'NEGATIVE'));

   ID    INDICATOR    SALES
_____ ____________ ________
    1 POSITIVE          190
    1 NEGATIVE            0

,

您可以添加虚拟数据来处理这种情况-

SELECT ID,SUM(sales) SALES,CASE WHEN STATUS='XX' THEN 'POSITIVE'
            WHEN STATUS='YY' THEN 'NEGATIVE'
       END INDICATOR
  FROM (SELECT ID,sales,STATUS
          FROM sample_table
        -- ADDING A DUMMY ROW
         UNION ALL
        SELECT DISTINCT ID,'YY'
          FROM sample_table)
 GROUP BY ID,CASE WHEN STATUS='XX' THEN 'POSITIVE'
               WHEN STATUS='YY' THEN 'NEGATIVE'
          END;

,

有很多方法可以解决这个问题。我喜欢首先定义我希望始终出现的列，然后外部连接表以用于聚合查询。

WITH
    sample_table (ID,'XX' FROM DUAL
         UNION ALL
         SELECT 2,'YY' FROM DUAL)
SELECT 
ID2,sum(nvl(sales,0)) SALES,CASE 
WHEN STATUS2='XX' THEN 'POSITIVE'
WHEN STATUS2='YY' THEN 'NEGATIVE'
END INDICATOR
from 
(select distinct t1.id id2,t2.status status2 
    from sample_table t1 cross join sample_table t2) base_table
left outer join sample_table on id=id2 and status=status2  
group by 
ID2,CASE 
WHEN STATUS2='XX' THEN 'POSITIVE'
WHEN STATUS2='YY' THEN 'NEGATIVE'
END
order by 1;

输出：

       ID2      SALES INDICATOR
---------- ---------- ---------
         1          0 NEGATIVE 
         1        190 POSITIVE 
         2         50 NEGATIVE 
         2          0 POSITIVE