另一个coalesce()不停止对第一个非空值求值的证明:

SELECT COALESCE(1, my_sequence.nextval) AS answer FROM dual;

运行这个，然后检查my_sequence.currval;

NVL将隐式转换为第一个参数的数据类型，因此下面的操作不会出错

select nvl('a',sysdate) from dual;

COALESCE期望一致的数据类型。

select coalesce('a',sysdate) from dual;

将抛出'不一致的数据类型错误'

另一个coalesce()不停止对第一个非空值求值的证明:

SELECT COALESCE(1, my_sequence.nextval) AS answer FROM dual;

运行这个，然后检查my_sequence.currval;

在计划处理方面也有不同。

当搜索包含nvl结果与索引列的比较时，Oracle可以通过串联分支过滤器形成优化计划。

create table tt(a, b) as
select level, mod(level,10)
from dual
connect by level<=1e4;

alter table tt add constraint ix_tt_a primary key(a);
create index ix_tt_b on tt(b);

explain plan for
select * from tt
where a=nvl(:1,a)
  and b=:2;

explain plan for
select * from tt
where a=coalesce(:1,a)
  and b=:2;

nvl:

-----------------------------------------------------------------------------------------
| Id  | Operation                     | Name    | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |         |     2 |    52 |     2   (0)| 00:00:01 |
|   1 |  CONCATENATION                |         |       |       |            |          |
|*  2 |   FILTER                      |         |       |       |            |          |
|*  3 |    TABLE ACCESS BY INDEX ROWID| TT      |     1 |    26 |     1   (0)| 00:00:01 |
|*  4 |     INDEX RANGE SCAN          | IX_TT_B |     7 |       |     1   (0)| 00:00:01 |
|*  5 |   FILTER                      |         |       |       |            |          |
|*  6 |    TABLE ACCESS BY INDEX ROWID| TT      |     1 |    26 |     1   (0)| 00:00:01 |
|*  7 |     INDEX UNIQUE SCAN         | IX_TT_A |     1 |       |     1   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------
   2 - filter(:1 IS NULL)
   3 - filter("A" IS NOT NULL)
   4 - access("B"=TO_NUMBER(:2))
   5 - filter(:1 IS NOT NULL)
   6 - filter("B"=TO_NUMBER(:2))
   7 - access("A"=:1)

合并:

---------------------------------------------------------------------------------------
| Id  | Operation                   | Name    | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT            |         |     1 |    26 |     1   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS BY INDEX ROWID| TT      |     1 |    26 |     1   (0)| 00:00:01 |
|*  2 |   INDEX RANGE SCAN          | IX_TT_B |    40 |       |     1   (0)| 00:00:01 |
---------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - filter("A"=COALESCE(:1,"A"))
   2 - access("B"=TO_NUMBER(:2))

请访问http://www.xt-r.com/2012/03/nvl-coalesce-concatenation.html。

事实上，我不能同意每一种说法。

COALESCE期望所有参数具有相同的数据类型。

This is wrong, see below. Arguments can be different data types, that is also documented: If all occurrences of expr are numeric data type or any nonnumeric data type that can be implicitly converted to a numeric data type, then Oracle Database determines the argument with the highest numeric precedence, implicitly converts the remaining arguments to that data type, and returns that data type.. Actually this is even in contradiction to common expression "COALESCE stops at first occurrence of a non-Null value", otherwise test case No. 4 should not raise an error.

根据测试用例5,COALESCE对参数进行隐式转换。

DECLARE
    int_val INTEGER := 1;
    string_val VARCHAR2(10) := 'foo';
BEGIN

    BEGIN
    DBMS_OUTPUT.PUT_LINE( '1. NVL(int_val,string_val) -> '|| NVL(int_val,string_val) );
    EXCEPTION WHEN OTHERS THEN DBMS_OUTPUT.PUT_LINE('1. NVL(int_val,string_val) -> '||SQLERRM ); 
    END;

    BEGIN
    DBMS_OUTPUT.PUT_LINE( '2. NVL(string_val, int_val) -> '|| NVL(string_val, int_val) );
    EXCEPTION WHEN OTHERS THEN DBMS_OUTPUT.PUT_LINE('2. NVL(string_val, int_val) -> '||SQLERRM ); 
    END;

    BEGIN
    DBMS_OUTPUT.PUT_LINE( '3. COALESCE(int_val,string_val) -> '|| COALESCE(int_val,string_val) );
    EXCEPTION WHEN OTHERS THEN DBMS_OUTPUT.PUT_LINE('3. COALESCE(int_val,string_val) -> '||SQLERRM ); 
    END;

    BEGIN
    DBMS_OUTPUT.PUT_LINE( '4. COALESCE(string_val, int_val) -> '|| COALESCE(string_val, int_val) );
    EXCEPTION WHEN OTHERS THEN DBMS_OUTPUT.PUT_LINE('4. COALESCE(string_val, int_val) -> '||SQLERRM ); 
    END;

    DBMS_OUTPUT.PUT_LINE( '5. COALESCE(SYSDATE,SYSTIMESTAMP) -> '|| COALESCE(SYSDATE,SYSTIMESTAMP) );

END;

Output:

1. NVL(int_val,string_val) -> ORA-06502: PL/SQL: numeric or value error: character to number conversion error
2. NVL(string_val, int_val) -> foo
3. COALESCE(int_val,string_val) -> 1
4. COALESCE(string_val, int_val) -> ORA-06502: PL/SQL: numeric or value error: character to number conversion error
5. COALESCE(SYSDATE,SYSTIMESTAMP) -> 2016-11-30 09:55:55.000000 +1:0 --> This is a TIMESTAMP value, not a DATE value!

NVL:将null替换为value。

COALESCE:从表达式列表中返回第一个非空表达式。

表:PRICE_LIST

+----------------+-----------+
| Purchase_Price | Min_Price |
+----------------+-----------+
| 10             | null      |
| 20             |           |
| 50             | 30        |
| 100            | 80        |
| null           | null      |
+----------------+-----------+   

的例子如下 [1]设定销售价格，所有产品增加10%利润。 [2]如果没有进货目录价格，则销售价格为最低价。清仓甩卖。 [3]如果也没有最低价格，则将销售价格设置为默认价格“50”。

SELECT
     Purchase_Price,
     Min_Price,
     NVL(Purchase_Price + (Purchase_Price * 0.10), Min_Price)    AS NVL_Sales_Price,
COALESCE(Purchase_Price + (Purchase_Price * 0.10), Min_Price,50) AS Coalesce_Sales_Price
FROM 
Price_List

用现实生活中的实例解释。

+----------------+-----------+-----------------+----------------------+
| Purchase_Price | Min_Price | NVL_Sales_Price | Coalesce_Sales_Price |
+----------------+-----------+-----------------+----------------------+
| 10             | null      | 11              |                   11 |
| null           | 20        | 20              |                   20 |
| 50             | 30        | 55              |                   55 |
| 100            | 80        | 110             |                  110 |
| null           | null      | null            |                   50 |
+----------------+-----------+-----------------+----------------------+

你可以看到，使用NVL，我们可以实现规则[1]，[2]，但使用COALSECE，我们可以实现所有三个规则。