by子句

Question

排除SELECT語句的列組我有一個生成一個報告

select 
    cust.firstname as "FIRST NAME", 
    cust.lastname as "LAST NAME", 
    dept.name as "DEPARTMENT", 
    subdept.name as "SUB DEPARTMENT", 
    count(*) "cust type" 
from cust 
join dept on dept.deptid = cust.deptid 
join subdept on subdept.deptid = cust.subdeptid 
join custactivity on custactivity.custid = cust.custid 
where custactivity.swipetime > (trunc(sysdate - 1)) 
group by cust.firstname,cust.lastname,dept.name,subdept.name 
having count(*) > 5 
order by "DEPARTMENT" ; 

現在下面的查詢，我想包括列「custactivity.swipetime」只有在SELECT語句，但沒有組by子句，因爲它是一個時間字段，如果將它包含在group by子句中，我會得到不正確的結果。這可能怎麼樣？

我試過在ORACLE Select and group by excluding one field 的解決方案，但它沒有解決。

任何人都可以幫忙嗎？

Answer 1

每個顧客可能有很多的custactivities。所以你必須告訴DBMS選擇哪個swipetime。例如最新的：

, max(custactivity.swipetime) as "last swipe time" 

---

在下面你說你要每cust.firstname，cust.lastname，dept.name，subdept.name，和 custactivity.swipetime一個結果行的註釋。所以swipetime必須添加到group by子句中。

這是您的swipetime添加到group by子句的查詢。您需要在原始組上使用SUM分析函數。由於您無法在having子句中使用分析函數，因此您需要周圍的查詢。

select * 
from 
(
    select 
    cust.firstname as "FIRST NAME", 
    cust.lastname as "LAST NAME", 
    dept.name as "DEPARTMENT", 
    subdept.name as "SUB DEPARTMENT", 
    custactivity.swipetime as "SWIPE TIME", 
    sum(count(*)) over (partition by 
         cust.firstname, 
         cust.lastname, 
         dept.name, 
         subdept.name) as "cust type" 
    from cust 
    join dept on dept.deptid = cust.deptid 
    join subdept on subdept.deptid = cust.subdeptid 
    join custactivity on custactivity.custid = cust.custid 
    where custactivity.swipetime > (trunc(sysdate - 1)) 
    group by 
    cust.firstname, 
    cust.lastname, 
    dept.name, 
    subdept.name, 
    custactivity.swipetime 
) 
where "cust type" > 5 
order by "DEPARTMENT";