排序合并连接 (Sort Merge Join)是一种两个表在做连接时用排序操作(Sort)和合并操作(Merge)来得到连接结果集的连接方法。
对于排序合并连接的优缺点及适用场景例如以下：
a,通常情况下。排序合并连接的运行效率远不如哈希连接，但前者的使用范围更广。由于哈希连接仅仅能用于等值连接条件，而排序合并连接还能用于其它连接条件(如<,<=,>.>=)
b,通常情况下。排序合并连接并不适合OLTP类型的系统。其本质原因是对于由于OLTP类型系统而言，排序是很昂贵的操作，当然，假设能避免排序操作就例外了。
oracle表之间的连接之排序合并连接(Merge Sort Join)，其特点例如以下：
1，驱动表和被驱动表都是最多仅仅被訪问一次。
2，排序合并连接的表无驱动顺序。
3，排序合并连接的表须要排序，用到SORT_AREA_SIZE。
4，排序合并连接不适用于的连接条件是：不等于<>。like,当中大于>，小于<，大于等于>=，小于等于<=，是能够适用于排序合并连接
5。排序合并连接，假设有索引就能够排除排序。
以下我来做个实验来证实如上的结论：
详细的測试基础表请查看本人Blog 例如以下链接：
oracle表连接之----〉嵌套循环(Nested Loops Join)
1。驱动表和被驱动表的訪问次数：
SQL> select /*+ ordered use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id;
SQL> select sql_id, child_number, sql_text from v$sql where sql_text like '%use_merge%';
SQL_ID CHILD_NUMBER SQL_TEXT
------------- ------------ --------------------------------------------------------------------------------
85u4h9hfqa5ar 0 select sql_id, child_number, sql_text from v$sql where sql_text like '%use_merg
6xph9fhapys39 0 select /*+ ordered use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id
SQL> select * from table(dbms_xplan.display_cursor('6xph9fhapys39',0,'allstats last'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
SQL_ID 6xph9fhapys39, child number 0
-------------------------------------
select /*+ ordered use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id
Plan hash value: 412793182
--------------------------------------------------------------------------------
| Id | Operation | Name |
Starts | E-Rows | A-Rows | A-Time | Buf
--------------------------------------------------------------------------------
| 1 | MERGE JOIN | | 1 | 100 | 100 |00:00:00.07 |
| 2 | SORT JOIN | | 1 | 100 | 100 |00:00:00.01 |
| 3 | TABLE ACCESS FULL| T1 | 1 | 100 | 100 |00:00:00.01 |
|* 4 | SORT JOIN | | 100 | 100K| 100 |00:00:00.07 |
| 5 | TABLE ACCESS FULL| T2 | 1 | 100K| 100K|00:00:00.01 |
--------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("T1"."ID"="T2"."T1_ID")
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
filter("T1"."ID"="T2"."T1_ID")
Note
-----
- dynamic sampling used for this statement
26 rows selected
从上面的实验能够看出排序合并连接和HASH连接时一样的。T1和T2 表都仅仅会被訪问0次或者1次。
select /*+ ordered use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id and 1=2;此语句T1和T2表就会是被訪问0次。
自己能够做试验測试下。
总结：排序合并连接根本就没有驱动和被驱动表的概念，而嵌套循环连接和哈希连接就要考虑驱动和被驱动表的情况。！
2，排序合并的表的驱动顺序
以下是T1为驱动表的运行计划
select /*+ leading(t1) use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id and t1.num=20;
select sql_id,child_number,sql_text from v$sql where sql_text like '%from t1,t2 where t1.id=t2.t1_id and t1.num=20%';
SQL> select * from table(dbms_xplan.display_cursor('8z4jvhnnfhxyf',0,'allstats last'));
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
SQL_ID 8z4jvhnnfhxyf, child number 0
-------------------------------------
select /*+ leading(t1) use_merge(t2)*/ * from t1,t2 where t1.id=t2.t1_id and t1.num=20
Plan hash value: 412793182
-----------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
-----------------------------------------------------------------------------------------------------------------
| 1 | MERGE JOIN | | 1 | 1 | 1 |00:00:00.58 | 3462 | | | |
| 2 | SORT JOIN | | 1 | 1 | 1 |00:00:00.01 | 6 | 2048 | 2048 |2048
(0)|
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|* 3 | TABLE ACCESS FULL| T1 | 1 | 1 | 1 |00:00:00.01 | 6 | | | |
|* 4 | SORT JOIN | | 1 | 100K| 1 |00:00:00.58 | 3456 | 14M| 1490K| 12M
(0)|
| 5 | TABLE ACCESS FULL| T2 | 1 | 100K| 100K|00:00:00.01 | 3456 | | | |
-----------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("T1"."NUM"=20)
4 - access("T1"."ID"="T2"."T1_ID")
filter("T1"."ID"="T2"."T1_ID")
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
23 rows selected.
Elapsed: 00:00:00.01
以下是T2为驱动表的运行计划：
SQL> select * from table(dbms_xplan.display_cursor('bxydvw58bhczf',0,'allstats last'));
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
SQL_ID bxydvw58bhczf, child number 0
-------------------------------------
select /*+ leading(t2) use_merge(t1)*/ * from t1,t2 where t1.id=t2.t1_id and t1.num=20
Plan hash value: 1792967693
-----------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem |
-----------------------------------------------------------------------------------------------------------------
| 1 | MERGE JOIN | | 1 | 1 | 1 |00:00:02.20 | 3462 | | | |
| 2 | SORT JOIN | | 1 | 100K| 21 |00:00:02.20 | 3456 | 14M| 1490K| 12M
(0)|
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
| 3 | TABLE ACCESS FULL| T2 | 1 | 100K| 100K|00:00:00.10 | 3456 | | | |
|* 4 | SORT JOIN | | 21 | 1 | 1 |00:00:00.01 | 6 | 2048 | 2048 |2048
(0)|
|* 5 | TABLE ACCESS FULL| T1 | 1 | 1 | 1 |00:00:00.01 | 6 | | | |
-----------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("T1"."ID"="T2"."T1_ID")
filter("T1"."ID"="T2"."T1_ID")
5 - filter("T1"."NUM"=20)
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
23 rows selected.
Elapsed: 00:00:00.85
从上面的两个运行计划能够看出，不管T1表示驱动表还是被驱动表，效果都是一样的，排序的尺寸一个是2048+12M，一个是12M+2048。
结论：排序合并连接没有驱动的概念。不管哪个表再前面都无所谓。
3，排序合并连接的限制
SQL〉explain plan for select /*+ leading(t1) use_merge(t2)*/ * from t1,t2 where t1.id<>t2.t1_id and t1.num=20;
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 4016936828
---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 5000 | 1083K| 82709 (1)| 00:15:10 |
| 1 | NESTED LOOPS | | 5000 | 1083K| 82709 (1)| 00:15:10 |
| 2 | TABLE ACCESS FULL| T2 | 100K| 10M| 710 (1)| 00:00:08 |
|* 3 | TABLE ACCESS FULL| T1 | 1 | 107 | 1 (0)| 00:00:01 |
---------------------------------------------------------------------------
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("T1"."NUM"=20 AND TO_CHAR("T1"."ID") LIKE
TO_CHAR("T2"."T1_ID"))
16 rows selected.
从上面的运行计划能够看出，优化器走的是NESTED LOOPS JOIN。
SQL> explain plan for select /*+ leading(t1) use_merge(t2)*/ * from t1,t2 where t1.id>t2.t1_id and t1.num=20;
Explained.
Elapsed: 00:00:00.01
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 412793182
------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes |TempSpc| Cost (%CPU)| Time |
------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 5000 | 1083K| | 5080 (1)| 00:00:56 |
| 1 | MERGE JOIN | | 5000 | 1083K| | 5080 (1)| 00:00:56 |
| 2 | SORT JOIN | | 1 | 107 | | 4 (25)| 00:00:01 |
|* 3 | TABLE ACCESS FULL| T1 | 1 | 107 | | 3 (0)| 00:00:01 |
|* 4 | SORT JOIN | | 100K| 10M| 25M| 5076 (1)| 00:00:56 |
| 5 | TABLE ACCESS FULL| T2 | 100K| 10M| | 710 (1)| 00:00:08 |
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("T1"."NUM"=20)
4 - access(INTERNAL_FUNCTION("T1"."ID")>INTERNAL_FUNCTION("T2"."T1_ID"))
filter(INTERNAL_FUNCTION("T1"."ID")>INTERNAL_FUNCTION("T2"."T1_ID"))
19 rows selected.
同理能够实验得出：排序合并连接不适用于的连接条件是：不等于<>，like,当中大于>，小于<。大于等于>=，小于等于<=，是能够适用于排序合并连接