oracle执行计划中filter,【DB笔试面试258】在Oracle中，执行计划里的access和filter有什么区别？...

LHR@orclasm > set autot on

LHR@orclasm > SELECT T.JOB, COUNT(1) FROM SCOTT.EMP T GROUP BY T.JOB  HAVING COUNT(1)>3;

JOB         COUNT(1)

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

CLERK              4

SALESMAN           4

Execution Plan

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

Plan hash value: 2138686577

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

| Id  | Operation           | Name | Rows  | Bytes | Cost (%CPU)| Time     |

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

|   0 | SELECT STATEMENT    |      |     1 |     8 |     4  (25)| 00:00:01 |

|*  1 |  FILTER             |      |       |       |            |          |

|   2 |   HASH GROUP BY     |      |     1 |     8 |     4  (25)| 00:00:01 |

|   3 |    TABLE ACCESS FULL| EMP  |    14 |   112 |     3   (0)| 00:00:01 |

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

Predicate Information (identified by operation id):

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

1 - filter(COUNT(*)>3)

Statistics

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

25  recursive calls

4  db block gets

6  consistent gets

0  physical reads

1544  redo size

660  bytes sent via SQL*Net to client

519  bytes received via SQL*Net from client

2  SQL*Net roundtrips to/from client

0  sorts (memory)

0  sorts (disk)

2  rows processed

很显然ID1的filter操作只有一个子节点ID2，在这种情况下的filter操作也就是单纯的过滤操作。