Question

我正在编写一个sql报告，对我的要求之一是为CASE表达式的两个选项都包含行。

该报告将显示数据库中所有客户与3个所选客户中的每个客户之间的“费用”和“非成本”交易的计数。

请参阅所需的示例布局。

CUSTOMER_ID | CUSTOMER_NAME | COSTING     | CUST_1 | CUST_2 | CUST_3
12          | Red Corp      | Non-costing | 1245   | 734    | 84
12          | Red Corp      | Costing     | 376    | 23     | 2
17          | Blue Corp     | Non-costing | 4538   | 36     | 3256
17          | Blue Corp     | Costing     | 0      | 0      | 0
...

我尝试将COSTING列放在case语句中。但是，由于空白，我无法在输出中显示示例“ Blue Corp”“ Costing”行。

select  CUSTOMER_ID
        ,CUSTOMER_NAME
        ,case 
            when PURCHASE_VALUE = 0 or PURCHASE_VALUE is null then 'Non-costing'
            when PURCHASE_VALUE != 0 or PURCHASE_VALUE is not null then 'Costing'
        end AS COSTING
,count(CASE WHEN CUSTOMER_ID = 11 then 1 ELSE NULL END) as "CUST_1"
,count(CASE WHEN CUSTOMER_ID = 22 then 1 ELSE NULL END) as "CUST_2"
,count(CASE WHEN CUSTOMER_ID = 33 then 1 ELSE NULL END) as "CUST_3"
from schema.CUSTOMER_TRANSACTIONS
group by case 
            when PURCHASE_VALUE = 0 or PURCHASE_VALUE is null then 'Non-costing'
            when PURCHASE_VALUE != 0 or PURCHASE_VALUE is not null then 'Costing'
        end 
        ,CUSTOMER_ID
        ,CUSTOMER_NAME

我目前与示例相同，但是缺少0计数行。请在下面查看我当前得到的输出。

CUSTOMER_ID | CUSTOMER_NAME | COSTING     | CUST_1 | CUST_2 | CUST_3
12          | Red Corp      | Non-costing | 1245   | 734    | 84
12          | Red Corp      | Costing     | 376    | 23     | 2
17          | Blue Corp     | Non-costing | 4538   | 36     | 3256
...

Answer 1

使用cross join生成行，并使用left join添加列：

select c.CUSTOMER_ID, c.CUSTOMER_NAME, t.costing,
       sum(CASE WHEN CUSTOMER_ID = 11 then 1 ELSE 0 END) as CUST_1,
       sum(CASE WHEN CUSTOMER_ID = 22 then 1 ELSE 0 END) as CUST_2,
       sum(CASE WHEN CUSTOMER_ID = 33 then 1 ELSE 0 END) as CUST_3
from (select distinct ct.customer_id, ct.customer_name
      from schema.CUSTOMER_TRANSACTIONS ct
     ) c cross join
     (select 'Non-costing' as costing from dual union all
      select 'Costing' from dual
     ) t left join
     schema.customer_transactions ct
     on c.customer_id = ct.customer_id,
        t.costing = (case when PURCHASE_VALUE = 0 or PURCHASE_VALUE is null
                         then 'Non-costing'
                         else 'Costing'
                     end)
group by c.CUSTOMER_ID, c.CUSTOMER_NAME, t.type;

Answer 2

根据您的评论

Alex Poole，您是正确的，没有用于“ Costing”的数据 “ Blue-corp”。但是我希望该行显示为0而不是完全出现。

我认为您需要生成一个记录，其中没有成本计算之一。

我试图创建查询，如果“成本”或“非成本”记录不适用于任何客户，则将生成记录。

我无法对其进行测试，因此如果您遇到任何问题，请分享。

-- Your original query starts from here
WITH DATAA AS (select  CUSTOMER_ID
        ,CUSTOMER_NAME
        ,case 
            when PURCHASE_VALUE = 0 or PURCHASE_VALUE is null then 'Non-costing'
            when PURCHASE_VALUE != 0 or PURCHASE_VALUE is not null then 'Costng'
        end AS COSTING
,count(CASE WHEN CUSTOMER_ID = 11 then 1 ELSE NULL END) as "CUST_1"
,count(CASE WHEN CUSTOMER_ID = 22 then 1 ELSE NULL END) as "CUST_2"
,count(CASE WHEN CUSTOMER_ID = 33 then 1 ELSE NULL END) as "CUST_3"
from schema.CUSTOMER_TRANSACTIONS
group by case 
            when PURCHASE_VALUE = 0 or PURCHASE_VALUE is null then 'Non-costing'
            when PURCHASE_VALUE != 0 or PURCHASE_VALUE is not null then 'Costing'
        end 
        ,CUSTOMER_ID
        ,CUSTOMER_NAME)
-- Your original query ends here


-- Solution
SELECT
    CUSTOMER_ID,
    CUSTOMER_NAME,
    COSTING,
    CUST_1,
    CUST_2,
    CUST_3
FROM
    DATAA
UNION ALL
( SELECT
    CUSTOMER_ID,
    CUSTOMER_NAME,
    CASE
        WHEN COUNT(CASE
            WHEN COSTING = 'Non-costing' THEN 1
        END) = 0 THEN 'Non-costing'
        WHEN COUNT(CASE
            WHEN COSTING = 'Costing' THEN 1
        END) = 0 THEN 'Costing'
    END AS COSTING,
    0,
    0,
    0
FROM
    DATAA D1
GROUP BY
    CUSTOMER_ID,
    CUSTOMER_NAME
HAVING ( COUNT(CASE
    WHEN COSTING = 'Non-costing' THEN 1
END) = 0
         OR COUNT(CASE
    WHEN COSTING = 'Costng' THEN 1
END) = 0 )
);

干杯！

Answer 3

您可以尝试一下。没有时它将给您0。

select  CUSTOMER_ID
        ,CUSTOMER_NAME
, CASE WHEN ISNULL(PURCHASE_Value,0) = 0 THEN 'Non-costing' ELSE 'Costing' END        
,SUM(CASE WHEN CUSTOMER_ID = 11 then 1 ELSE 0 END) as "CUST_1"
,SUM(CASE WHEN CUSTOMER_ID = 22 then 1 ELSE 0 END) as "CUST_2"
,SUM(CASE WHEN CUSTOMER_ID = 33 then 1 ELSE 0 END) as "CUST_3"
from schema.CUSTOMER_TRANSACTIONS
group by CUSTOMER_ID
        ,CUSTOMER_NAME
,PURCHASE_Value

显示案例表达式中的空白行

3 个答案: