qb_name

实验1：优化器对语句进行了merge

SELECT COUNT(DISTINCT T1_VC || T2_VC)
  FROM (SELECT T1.SMALL_VC T1_VC, T2.SMALL_VC T2_VC
          FROM T1, T2
         WHERE T2.ID = T1.ID
           AND T1.N1 <= 1000);

执行计划
----------------------------------------------------------
Plan hash value: 100109877

------------------------------------------------------------------
| Id  | Operation	      | Name	 | Rows  | Bytes | Cost  |
------------------------------------------------------------------
|   0 | SELECT STATEMENT      | 	 |     1 |    22 |   115 |
|   1 |  SORT AGGREGATE       | 	 |     1 |    22 |	 |
|   2 |   VIEW		      | VW_DAG_0 |  1000 | 22000 |   115 |
|   3 |    HASH GROUP BY      | 	 |  1000 | 34000 |   115 |
|*  4 |     HASH JOIN	      | 	 |  1000 | 34000 |    63 |
|*  5 |      TABLE ACCESS FULL| T1	 |  1000 | 19000 |    31 |
|   6 |      TABLE ACCESS FULL| T2	 | 10000 |   146K|    31 |
------------------------------------------------------------------

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

   4 - access("T2"."ID"="T1"."ID")
   5 - filter("T1"."N1"<=1000)

Note
-----
   - cpu costing is off (consider enabling it)

实验2：使用no_merge hint，见Id=4处的不同

SELECT COUNT(DISTINCT T1_VC || T2_VC)
  FROM (SELECT /*+ no_merge */
         T1.SMALL_VC T1_VC, T2.SMALL_VC T2_VC
          FROM T1, T2
         WHERE T2.ID = T1.ID
           AND T1.N1 <= 1000);

执行计划
----------------------------------------------------------
Plan hash value: 1355605975

-------------------------------------------------------------------
| Id  | Operation	       | Name	  | Rows  | Bytes | Cost  |
-------------------------------------------------------------------
|   0 | SELECT STATEMENT       |	  |	1 |    22 |   114 |
|   1 |  SORT AGGREGATE        |	  |	1 |    22 |	  |
|   2 |   VIEW		       | VW_DAG_0 |  1000 | 22000 |   114 |
|   3 |    HASH GROUP BY       |	  |  1000 | 22000 |   114 |
|   4 |     VIEW	       |	  |  1000 | 22000 |    63 |
|*  5 |      HASH JOIN	       |	  |  1000 | 34000 |    63 |
|*  6 |       TABLE ACCESS FULL| T1	  |  1000 | 19000 |    31 |
|   7 |       TABLE ACCESS FULL| T2	  | 10000 |   146K|    31 |
-------------------------------------------------------------------

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

   5 - access("T2"."ID"="T1"."ID")
   6 - filter("T1"."N1"<=1000)

Note
-----
   - cpu costing is off (consider enabling it)

实验3：改变hint位置,此处不起作用

SELECT /*+ no_merge */
 COUNT(DISTINCT T1_VC || T2_VC)
  FROM (SELECT /*+ qb_name(bb) */
         T1.SMALL_VC T1_VC, T2.SMALL_VC T2_VC
          FROM T1, T2
         WHERE T2.ID = T1.ID
           AND T1.N1 <= 1000);

执行计划
----------------------------------------------------------
Plan hash value: 100109877

------------------------------------------------------------------
| Id  | Operation	      | Name	 | Rows  | Bytes | Cost  |
------------------------------------------------------------------
|   0 | SELECT STATEMENT      | 	 |     1 |    22 |   115 |
|   1 |  SORT AGGREGATE       | 	 |     1 |    22 |	 |
|   2 |   VIEW		      | VW_DAG_0 |  1000 | 22000 |   115 |
|   3 |    HASH GROUP BY      | 	 |  1000 | 34000 |   115 |
|*  4 |     HASH JOIN	      | 	 |  1000 | 34000 |    63 |
|*  5 |      TABLE ACCESS FULL| T1	 |  1000 | 19000 |    31 |
|   6 |      TABLE ACCESS FULL| T2	 | 10000 |   146K|    31 |
------------------------------------------------------------------

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

   4 - access("T2"."ID"="T1"."ID")
   5 - filter("T1"."N1"<=1000)

Note
-----
   - cpu costing is off (consider enabling it)

实验4：使用 no_merge(@bb) ，@bb来自/*+ qb_name(bb)，此处作用与实验2一样，实验完毕

SELECT /*+ no_merge(@bb) */
 COUNT(DISTINCT T1_VC || T2_VC)
  FROM (SELECT /*+ qb_name(bb) */
         T1.SMALL_VC T1_VC, T2.SMALL_VC T2_VC
          FROM T1, T2
         WHERE T2.ID = T1.ID
           AND T1.N1 <= 1000);

执行计划
----------------------------------------------------------
Plan hash value: 1355605975

-------------------------------------------------------------------
| Id  | Operation	       | Name	  | Rows  | Bytes | Cost  |
-------------------------------------------------------------------
|   0 | SELECT STATEMENT       |	  |	1 |    22 |   114 |
|   1 |  SORT AGGREGATE        |	  |	1 |    22 |	  |
|   2 |   VIEW		       | VW_DAG_0 |  1000 | 22000 |   114 |
|   3 |    HASH GROUP BY       |	  |  1000 | 22000 |   114 |
|   4 |     VIEW	       |	  |  1000 | 22000 |    63 |
|*  5 |      HASH JOIN	       |	  |  1000 | 34000 |    63 |
|*  6 |       TABLE ACCESS FULL| T1	  |  1000 | 19000 |    31 |
|   7 |       TABLE ACCESS FULL| T2	  | 10000 |   146K|    31 |
-------------------------------------------------------------------

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

   5 - access("T2"."ID"="T1"."ID")
   6 - filter("T1"."N1"<=1000)

Note
-----
   - cpu costing is off (consider enabling it)