另一個帶日期的NVL

Question

下面是一個關於NVL的示例SQL問題。

檢查PROGRAM表的結構。

CREATE TABLE programs (prog_id NOT NULL NUMBER(3), 
         prog_cost NUMBER(8, 2) 
         start_date NOT NULL DATE, 
         end_date DATE 
         ); 

哪兩個SQL語句會成功執行？ （選擇兩項）。

A) SELECT nvl(add_months(end_date, 1), SYSDATE) 
from programs; 

B) SELECT to_date(nvl(SYSDATE - end_date, SYSDATE)) 
from programs; 

C) SELECT nvl(months_between(start_date, end_date), 'ongoing') 
from programs; 

D) SELECT nvl(to_char(months_between(start_date, end_date)), 'ongoing') 
from programs; 

對我來說，只要在節目表中的數據沒有給出，所有上述四個語句可以成功執行。

正確答案是A & D.答案沒有解釋。有誰知道作者是如何到達A & D.

Answer 1

B和C在邏輯上有一些問題。還有臭名昭着的隱式轉換。

B) SELECT to_date(nvl(SYSDATE - end_date, SYSDATE)) 
from programs; 

表達式SYSDATE - end_date會給你兩個日期之間的間隔。由於它是數值，因此結果將有NUMBER數據類型。第二個參數（SYSDATE）是DATE。雖然隱式轉換可能實際上有助於避免異常，並給你一些數字結果，但我沒有真正理解，爲什麼有人需要將當前日期轉換爲數字。似乎很無用。

C) SELECT nvl(months_between(start_date, end_date), 'ongoing') 
from programs; 

好吧，我可以想象它的某種情況，雖然我可能會做它不同，但隱式轉換再次。

MONTHS_BETWEEN （doc） 返回日期之間的月數。

所以第一個參數有一個NUMBER數據類型。通過'ongoing'字符串作爲第二個參數應該會提高無效號碼例外，除非您有一些非常奇怪的NLS參數。

Answer 2

A和D作爲答案的原因歸結爲nvl函數中的數據類型不匹配。兩個參數必須是相同的類型。

with test_data as 
(
    select 1 as prog_id, 1 as prog_cost, to_date('10-12-2015', 'MM-dd-yyyy') as start_date, sysdate as end_date from dual 
    union all 
    select 2, null, to_date('10-12-2015', 'MM-dd-yyyy'), sysdate from dual 
    union all 
    select 3, 1, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
    union all 
    select 4, null, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
) 
select nvl(add_months(end_date, 1), sysdate) 
from test_data; 

上述執行並返回結果的所有記錄...

with test_data as 
(
    select 1 as prog_id, 1 as prog_cost, to_date('10-12-2015', 'MM-dd-yyyy') as start_date, sysdate as end_date from dual 
    union all 
    select 2, null, to_date('10-12-2015', 'MM-dd-yyyy'), sysdate from dual 
    union all 
    select 3, 1, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
    union all 
    select 4, null, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
) 
select to_date(nvl(SYSDATE - end_date, SYSDATE)) 
from test_data; 

因爲nvl的兩個參數是不相同的數據類型的上述引發錯誤ORA-01847: day of month must be between 1 and last day of month ...

with test_data as 
(
    select 1 as prog_id, 1 as prog_cost, to_date('10-12-2015', 'MM-dd-yyyy') as start_date, sysdate as end_date from dual 
    union all 
    select 2, null, to_date('10-12-2015', 'MM-dd-yyyy'), sysdate from dual 
    union all 
    select 3, 1, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
    union all 
    select 4, null, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
) 
select nvl(months_between(start_date, end_date), 'ongoing') 
from test_data; 

和上一個一樣，上面會拋出錯誤ORA-01722: invalid number，因爲nvl的兩個參數不是t他是相同的數據類型。 months_between返回一個數字，'ongoing'是一個varchar ...

最後一個作品，因爲它轉換的months_between返回到char ...

with test_data as 
(
    select 1 as prog_id, 1 as prog_cost, to_date('10-12-2015', 'MM-dd-yyyy') as start_date, sysdate as end_date from dual 
    union all 
    select 2, null, to_date('10-12-2015', 'MM-dd-yyyy'), sysdate from dual 
    union all 
    select 3, 1, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
    union all 
    select 4, null, to_date('10-12-2015', 'MM-dd-yyyy'), null from dual 
) 
select nvl(to_char(months_between(start_date, end_date)), 'ongoing') 
from test_data;