Oracle SQL：执行过滤的最有效方式

Question

当我开始在大表上编写 SQL 查询时，我想知道哪个最有效。在我看来，它们看起来都一样。寻找有关 SQL 如何分解和执行这些查询的解释

选项 1：

select a.*
from table_a a
where a.column in (select filter from table_b)

选项 2：

select a.*
from table_a a, table_b b
where a.column=b.filter

选项 3：

select a.*
from table_a a
inner join table_b b
on a.column=b.filter

Answer 1

在最新版本的 Oracle 中，优化器足够聪明，可以完成它的工作。所以没关系，您的两个查询都将在内部进行优化以有效地完成任务。优化器可能会进行查询重写并选择高效的执行计划。

让我们通过EMP 和DEPT 表的小例子来理解这一点。

查询 1：

SQL> explain plan for
  2  select a.*
  3  from emp a
  4  WHERE A.deptno IN (SELECT deptno FROM dept);

Explained.

SQL>
SQL> SELECT * FROM TABLE(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------
Plan hash value: 3956160932

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |    14 |   546 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| EMP  |    14 |   546 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------

   1 - filter("A"."DEPTNO" IS NOT NULL)

13 rows selected.

SQL>

查询 2：

SQL> explain plan for
  2  SELECT A.*
  3  FROM emp A, dept b
  4  where a.deptno=b.deptno;

Explained.

SQL>
SQL> SELECT * FROM TABLE(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------
Plan hash value: 3956160932

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |    14 |   546 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| EMP  |    14 |   546 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------

   1 - filter("A"."DEPTNO" IS NOT NULL)

13 rows selected.

SQL>

查询 3：

SQL> explain plan for
  2  select a.*
  3  from emp a
  4  INNER JOIN dept b
  5  ON A.deptno=b.deptno;

Explained.

SQL>
SQL> SELECT * FROM TABLE(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------
Plan hash value: 3956160932

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |    14 |   546 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| EMP  |    14 |   546 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

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

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------

   1 - filter("A"."DEPTNO" IS NOT NULL)

13 rows selected.

SQL>

因此，无论您以不同方式编写查询，优化器都会选择最有效的执行计划。除非您使用提示等强制优化器选择不同的计划。您总是会看到使用相同的计划执行。