如何從具有使用SQL

Question

從其他列中的多個值的列只顯示了更高的價值

我的數據是這樣

表：customer_error

我只是想希望得到結果作爲第一個出現的錯誤ID，而不是前面的錯誤ID。

Answer 1

我相信你正在尋找這樣的：

這裏是1路：

select distinct first_value(customer_id) over (partition by customer_id 
             order by error_id) customer_id, 
        first_value(error_id) over (partition by customer_id 
             order by error_id) error_id, 
        first_value(error_description) over (partition by customer_id 
             order by error_id) error_description 
    from customer_error 
/

和方式略有不同：

select customer_id, error_id, error_description 
    from (
     select row_number() over (partition by customer_id 
            order by error_id) rnum, 
       customer_id, error_id, error_description 
      from customer_error 
     ) 
    where rnum = 1 
/

兩者都使用Analytics（分析），一個非常有用的工具來做這種事情，我建議讀o因爲它非常有用，所以學習它。

Answer 2

您可以使用CTE和Row_Number（）函數。

with tmp1 as 
(select *, 
ROW_NUMBER() over (partition by customer_id order by error_id) as RowNum 
from YourTable 
) 

select customer_id, error_id, error_description 
from tmp1 
where RowNum = 1 

文檔上ROW_NUMBER（）：https://msdn.microsoft.com/en-us/library/ms186734.aspx和https://docs.oracle.com/cd/B19306_01/server.102/b14200/functions137.htm

Answer 3

該解決方案假設錯誤是由ERROR_ID訂購和使用聚合函數，而不是解析函數：

select customer_id 
     , max(error_id) KEEP (DENSE_RANK FIRST ORDER BY error_id) error_id 
     , max(error_description) KEEP (DENSE_RANK FIRST ORDER BY error_id) error_description 
    from customer_error 
    group by customer_id; 

Answer 4

如果我們假設ERROR_NUMBER最小值是首先出現的值，我們可以用普通的sql來做到這一點。我們的目標是獲得每個客戶的最小錯誤數量，然後將錯誤的相關錯誤描述粘貼到其上。

SELECT a.customer_id, a.error_id, b.error_description FROM 
    (SELECT customer_id, MIN(error_id) FROM customer_error 
     GROUP BY customer_id) a 
    LEFT JOIN customer_error b on a.error_id=b.error_id;