哪个查询是最佳 id<>3 或 {1,2,4} 中的 id

Question

如果一个表有很多行，每行有很多列，哪个查询更快？

如果列的值在 {1,2,3,4} 内

    SELECT * FROM table t WHERE t.column1<>3

或者

    SELECT * FROM table t WHERE t.column1 in {1,2,4}

这两种情况下是 3 次比较还是 1 次比较？

Answer 1

在 sqlserver 2008 和 2012 上测试：

问题N 3中的第一种方法：

CREATE TABLE #t (n INT)

;WITH N1 (N) AS 
(SELECT 1 FROM (VALUES (1), (1), (1), (1), (1), (1), (1), (1), (1),
(1)) n (N)),
N2 (N) AS (SELECT 1 FROM N1 AS N1 CROSS JOIN N1 AS N2),
N3 (N) AS (SELECT 1 FROM N2 AS N1 CROSS JOIN N2 AS N2),
N4 (N) AS (SELECT ROW_NUMBER() OVER(ORDER BY N1.N) FROM N3 AS N1 CROSS JOIN N3 AS N2),
N5 (N) AS (SELECT top 50000000 N%4 + 1 N FROM n4)
INSERT #t
SELECT n from n5

DECLARE @dummy int

DECLARE @a datetime = getdate()
SELECT @dummy = count(*)
FROM #t
WHERE N <> 3

SELECT datediff(ms, 0, getdate() - @a)

结果：

3886 milliseconds    

问题的第二种方法

CREATE TABLE #t (n INT)

;WITH N1 (N) AS 
(SELECT 1 FROM (VALUES (1), (1), (1), (1), (1), (1), (1), (1), (1),
(1)) n (N)),
N2 (N) AS (SELECT 1 FROM N1 AS N1 CROSS JOIN N1 AS N2),
N3 (N) AS (SELECT 1 FROM N2 AS N1 CROSS JOIN N2 AS N2),
N4 (N) AS (SELECT ROW_NUMBER() OVER(ORDER BY N1.N) FROM N3 AS N1 CROSS JOIN N3 AS N2),
N5 (N) AS (SELECT top 50000000 N%4 + 1 N FROM n4)
INSERT #t
SELECT n from n5

DECLARE @dummy int

DECLARE @a datetime = getdate()
SELECT @dummy = count(*) 
FROM #t
WHERE N in (1,2,4)

SELECT datediff(ms, 0, getdate() - @a)

结果：

4223 milliseconds

结论

N 3 更快（3886 vs 4223 毫秒）

是的，这似乎是一个很小的差异，但如果该列被声明为 varchar(10) - CREATE TABLE #t (n INT) - 差异会更显着（经过测试）。

Answer 2

它们是不同的。第一个选择 ID 不为 3 的所有记录，而第二个选择所有 ID 为 1、2 或 4 的记录。例如，不包括 ID = 5 的记录。

因此性能不应该是选择一个查询而不是另一个的原因。 要么您在此表中有数百万条记录，然后IN 方法不起作用，或者您实际上只有这 4 条记录，那么您在问什么？在可读性和故障安全方面选择最合适的方式，不要关心纳秒。

查询优化主要依赖于 RDBMS，那么您使用的是哪个？在 SQL-Server 中应该没有区别。但是你自己检查一下。 If you have two horses and you want to know which of the two is the faster then race your horses.

Answer 3

在 Oracle 中，如果您有 INDEX，将会产生巨大的影响。你可以做一个小测试用例，检查每条语句的解释计划。

• 案例#1：没有索引

设置

出于演示目的，我从 scott.emp 表中创建了一个小表，有 4 行。

SQL> CREATE TABLE t AS SELECT * FROM emp WHERE empno IN (7369, 7499, 7521, 7566);

Table created.

SQL>

由于我在 12c，CTAS 不需要收集统计数据。

让我们看看解释计划：

使用不等于条件

SQL> EXPLAIN PLAN FOR SELECT * FROM t WHERE empno <> 7369;

Explained.

SQL>
SQL> SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY);

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |     3 |   117 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T    |     3 |   117 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

   1 - filter("EMPNO"<>7369)

13 rows selected.

SQL>

使用 IN 条件

SQL> EXPLAIN PLAN FOR SELECT * FROM t WHERE empno IN (7499, 7521, 7566);

Explained.

SQL>
SQL> SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY);

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |     3 |   117 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T    |     3 |   117 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

   1 - filter("EMPNO"=7499 OR "EMPNO"=7521 OR "EMPNO"=7566)

13 rows selected.

SQL>

• 案例#2：有索引

设置

SQL> CREATE INDEX t_indx ON t(empno);

Index created.

SQL>

使用不等于条件

SQL> EXPLAIN PLAN FOR SELECT * FROM t WHERE empno <> 7369;

Explained.

SQL>
SQL> SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY);

PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |     3 |   117 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T    |     3 |   117 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

   1 - filter("EMPNO"<>7369)

13 rows selected.

SQL>

使用 IN 条件

SQL> EXPLAIN PLAN FOR SELECT * FROM t WHERE empno IN (7499, 7521, 7566);

Explained.

SQL>
SQL> SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY);

PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------------
Plan hash value: 4217410154

-----------------------------------------------------------------------------------------------
| Id  | Operation                            | Name   | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                     |        |     3 |   117 |     2   (0)| 00:00:01 |
|   1 |  INLIST ITERATOR                     |        |       |       |            |          |
|   2 |   TABLE ACCESS BY INDEX ROWID BATCHED| T      |     3 |   117 |     2   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN                  | T_INDX |     3 |       |     1   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------------

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

   3 - access("EMPNO"=7499 OR "EMPNO"=7521 OR "EMPNO"=7566)

15 rows selected.

SQL>