2014-09-26 41 views
0

下面是在我計算值四個屬性查詢..如何把我的查詢合併到一個單一的查詢(或可能是一個存儲過程。)

1. 
Here I am calculating the values for chrg_orig 


    with ep as 
     (select emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm') as pay_month, 
     max(code), 
     max(bill) as bill, 
     max(chrg) as charge, 
     sum(nvl(pay_1,0)) sum_pay1, 
     sum(nvl(pay_2, 0)) sum_pay2, 
     (chrg_mon*22)+ (chrg_week*5)+ chrg_day as days, 
     from emp_payments 
     where emp_cd in ('HP','2000') 
     and code  in ('X','Y','Z') 
     group by emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm'), 
     code 
    ) 
    select emp_cd, 
     emp_num, 
     pay_month, 
     max(code), 
     sum(bill) 
     case when sum(days)=22 then sum(chrg) else round((round(sum(chrg)/sum(days),4)*22),2) end as chrg_orig 
    from ep 
    where chrg <>0 
    group by 
    emp_cd, 
    emp_num, 
    paymonth 

---------------------------------------------------------------------------------------- 
2. 
Here I am calculating the values for rate_chrg 

    with ep as 
     (select emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm') as pay_month, 
     code, 
     max(bill) as bill, 
     max(chrg) as charge, 
     sum(nvl(pay_1,0)) sum_pay1, 
     sum(nvl(pay_2, 0)) sum_pay2, 
     (chrg_mon*22)+ (chrg_week*5)+ chrg_day as days, 
     from emp_payments 
     where emp_cd in ('HP','2000') 
     and code  in ('X','Y','Z') 
     group by emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm'), 
     code 
    ) 
    (
    select a.emp_cd,a.emp_num,a.key,b.rate as rate_chrg from 
     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,max(invc_dt) as invc_dt from ep 
      where code in ('X','Y') and rate <> 0 
      group by emp_cd,emp_num,to_char(invc_dt,'yyyymm')) a, 

     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,invc_dt,rate from ep 
      where code in ('X','Y') and rate <> 0) b 
    where a.emp_cd = b.emp_cd 
    and a.emp_num = b.emp_num 
    and a.key = b.key 
    and a.invc_dt = b.invc_dt 
    ) 



----------------------------------------------------------------------------------------- 
3. 
Here I am calculating the values for bonus_chrg 




    with ep as 
     (select emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm') as pay_month, 
     code, 
     max(bill) as bill, 
     max(chrg) as charge, 
     sum(nvl(pay_1,0)) sum_pay1, 
     sum(nvl(pay_2, 0)) sum_pay2, 
     (chrg_mon*22)+ (chrg_week*5)+ chrg_day as days, 
     from emp_payments 
     where emp_cd in ('HP','2000') 
     and code  in ('X','Y','Z') 
     group by emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm'), 
     code 
    ) 
    (
    select a.emp_cd,a.emp_num,a.key,b.rate as bonus_chrg from 
     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,max(invc_dt) as invc_dt from ep 
      where code in ('Z') and rate <> 0 
      group by emp_cd,emp_num,to_char(invc_dt,'yyyymm')) a, 

     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,invc_dt,rate from ep 
      where code in ('Z') and rate <> 0) b 
    where a.emp_cd = b.emp_cd 
    and a.emp_num = b.emp_num 
    and a.key = b.key 
    and a.invc_dt = b.invc_dt 
    ) 

------------------------------------------------------------------------------------------ 
4. 
Here I am calculating the values for comp_days 



    with ep as 
     (select emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm') as pay_month, 
     code, 
     max(bill) as bill, 
     max(chrg) as charge, 
     sum(nvl(pay_1,0)) sum_pay1, 
     sum(nvl(pay_2, 0)) sum_pay2, 
     (chrg_mon*22)+ (chrg_week*5)+ chrg_day as days, 
     from emp_payments 
     where emp_cd in ('HP','2000') 
     and code  in ('X','Y','Z') 
     group by emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm'), 
     code 
    ) 
    (
    select emp_cd,emp_num,paymonth as key,sum(days) as comp_days from ep 
    where code in ('X','Y') 
    group by emp_cd,emp_num,key 
    ) 

-------------------------------------------------- 

我迄今所做是,我已將所有這些單獨的查詢放入ETL工具中,並使用chrg_orig作爲驅動表執行左外連接,併爲非匹配列指定零。 但我想我需要一個完整的外連接(我無法在工具中實現它)併爲所有不匹配的值分配零。

我想將這些查詢分組到一個查詢中。什麼是解決它的最佳方法? 所有的輸入和建議都是有價值的。感謝..

回答

1

您可以通過用逗號分隔CTE定義來創建帶有CTE的more than one virtual table。而且,CTE可以指其他CTE。

假設ep在所有這些查詢一樣,你可以做這樣的事情:

with ep as 
    (select emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm') as pay_month, 
     max(code), 
     max(bill) as bill, 
     max(chrg) as charge, 
     sum(nvl(pay_1,0)) sum_pay1, 
     sum(nvl(pay_2, 0)) sum_pay2, 
     (chrg_mon*22)+ (chrg_week*5)+ chrg_day as days, 
    from emp_payments 
    where emp_cd in ('HP','2000') 
    and code  in ('X','Y','Z') 
    group by emp_cd, 
     emp_num, 
     to_char(pay_dt,'yyyymm'), 
     code 
    ), 
chrg_orig (<field names here>) as (
    select emp_cd, 
    emp_num, 
    pay_month, 
    max(code), 
    sum(bill) 
    case when sum(days)=22 then sum(chrg) else round((round(sum(chrg)/sum(days),4)*22),2) end as chrg_orig 
    from ep 
    where chrg <>0 
    group by 
    emp_cd, 
    emp_num, 
    paymonth 
), 
rate_chrg (<field names here>) as (
    select a.emp_cd,a.emp_num,a.key,b.rate as rate_chrg from 
     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,max(invc_dt) as invc_dt from ep 
      where code in ('X','Y') and rate <> 0 
      group by emp_cd,emp_num,to_char(invc_dt,'yyyymm')) a, 

     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,invc_dt,rate from ep 
      where code in ('X','Y') and rate <> 0) b 
    where a.emp_cd = b.emp_cd 
    and a.emp_num = b.emp_num 
    and a.key = b.key 
    and a.invc_dt = b.invc_dt 
), 
bonus_chrg (<field names here>) as (
    select a.emp_cd,a.emp_num,a.key,b.rate as bonus_chrg from 
     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,max(invc_dt) as invc_dt from ep 
      where code in ('Z') and rate <> 0 
      group by emp_cd,emp_num,to_char(invc_dt,'yyyymm')) a, 

     (select emp_cd,emp_num,to_char(pay_dt,'yyyymm') as key,invc_dt,rate from ep 
      where code in ('Z') and rate <> 0) b 
    where a.emp_cd = b.emp_cd 
    and a.emp_num = b.emp_num 
    and a.key = b.key 
    and a.invc_dt = b.invc_dt 
), 
comp_days (<field names here>) as (
    select emp_cd,emp_num,paymonth as key,sum(days) as comp_days from ep 
    where code in ('X','Y') 
    group by emp_cd,emp_num,key 
) 
SELECT * 
FROM ep 
LEFT OUTER JOIN chrg_orig 
    ON <JOIN CONDITION> 
LEFT OUTER JOIN rate_chrg 
    ON <JOIN CONDITION> 
LEFT OUTER JOIN bonus_chrg 
    ON <JOIN CONDITION> 
LEFT OUTER JOIN comp_days 
    ON <JOIN CONDITION> 
+0

明白了,感謝..是不是最好的辦法? – Spider 2014-09-26 15:56:54

+0

@ user2942261這很難回答,因爲性能會因查詢,數據等而異。可能有更好的方法來拉動數據週期,比如使用'OVER()'子句來聚合數據,這可能允許多個聚合函數在具有較少CTE或子查詢的同一語句中。這取決於你的情況。 – 2014-09-26 16:02:44

相關問題