码迷,mamicode.com
首页 > 数据库 > 详细

oracle 表连接 - sort merge joins 排序合并连接

时间:2014-12-13 10:53:51      阅读:291      评论:0      收藏:0      [点我收藏+]

标签:merge join   sort join   排序合并连接   

一. sort merge joins连接(排序合并连接) 原理

指的是两个表连接时, 通过连接列先分别排序后, 再通过合并操作来得到最后返回的结果集的方法. 

假如表 T1 和 T2 的连接方式是排序合并连接, oracle 执行步骤如下:
(1) 根据 sql 语句中的谓词条件(如果有) 访问 T1 表, 得到一个过滤的结果集, 然后按照 T1 中的连接列对结果集进行排序
(2) 根据 sql 语句中的谓词条件(如果有) 访问 T2 表, 得到一个过滤的结果集, 然后按照 T2 中的连接列对结果集进行排序
(3) 将 1 和 2 的结果集合并起来, 对记录进行匹配得到最后的结果集.

通常来说, sort merge joins连接(排序合并连接) 使用并不广泛, 因为在大部分情况下使用 nested loops 或者 hash joins 都能获得比它更好的执行效率, 但是由于 hash joins 只能用于等值连接条件, 所以在非等值条件连接以及非 like 非 "<>" 情况下, 如果连接列上已经有排序, 使用 sort merge joins连接方式能获得比较好的执行效率


二. sort merge joins连接(排序合并连接) 特性

(1) 驱动表最多访问一次, 如果独立的谓词条件(不涉及驱动表字段的函数或者表达式等)不成立, 则不用再去访问驱动表
(2) 被驱动表最多访问一次. 如果驱动表没有记录, 被驱动表不用访问
(3) 驱动表的选择对于执行成本以及性能没有太大的影响
(4) 支持大部分的连接条件, 比如 ">" "<" ">=" "<=", 不支持 like, "<>"


构造试验数据

SQL> CREATE TABLE t1 (
  2    id NUMBER NOT NULL,
  3    n NUMBER,
  4    pad VARCHAR2(4000),
  5    CONSTRAINT t1_pk PRIMARY KEY(id)
  6  );

Table created.

SQL> CREATE TABLE t2 (
  2    id NUMBER NOT NULL,
  3    t1_id NUMBER NOT NULL,
  4    n NUMBER,
  5    pad VARCHAR2(4000),
  6    CONSTRAINT t2_pk PRIMARY KEY(id),
  7    CONSTRAINT t2_t1_fk FOREIGN KEY (t1_id) REFERENCES t1
  8  );

Table created.

SQL> CREATE TABLE t3 (
  2    id NUMBER NOT NULL,
  3    t2_id NUMBER NOT NULL,
  4    n NUMBER,
  5    pad VARCHAR2(4000),
  6    CONSTRAINT t3_pk PRIMARY KEY(id),
  7    CONSTRAINT t3_t2_fk FOREIGN KEY (t2_id) REFERENCES t2
  8  );

Table created.
SQL> CREATE TABLE t4 (
  2    id NUMBER NOT NULL,
  3    t3_id NUMBER NOT NULL,
  4    n NUMBER,
  5    pad VARCHAR2(4000),
  6    CONSTRAINT t4_pk PRIMARY KEY(id),
  7    CONSTRAINT t4_t3_fk FOREIGN KEY (t3_id) REFERENCES t3
  8  );


Table created.

SQL> execute dbms_random.seed(0)

PL/SQL procedure successfully completed.


SQL> INSERT INTO t1 SELECT rownum, rownum, dbms_random.string('a',50) FROM dual CONNECT BY level <= 10 ORDER BY dbms_random.random;

10 rows created.

SQL> INSERT INTO t2 SELECT 100+rownum, t1.id, 100+rownum, t1.pad FROM t1, t1 dummy ORDER BY dbms_random.random;

100 rows created.

SQL> INSERT INTO t3 SELECT 1000+rownum, t2.id, 1000+rownum, t2.pad FROM t2, t1 dummy ORDER BY dbms_random.random;

1000 rows created.

SQL> INSERT INTO t4 SELECT 10000+rownum, t3.id, 10000+rownum, t3.pad FROM t3, t1 dummy ORDER BY dbms_random.random;

10000 rows created.

SQL> COMMIT;

Commit complete.
使用 hint 让执行计划以 T3 作为驱动表
SQL> select /*+ leading(t3) use_merge(t4) */ *
  2  from t3, t4
  3  where t3.id = t4.t3_id and t3.n = 1100;


10 rows selected.


SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last'));


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


SQL_ID  g0rdyg9hdh9m0, child number 0
-------------------------------------
select /*+ leading(t3) use_merge(t4) */ * from t3, t4 where t3.id =
t4.t3_id and t3.n = 1100


Plan hash value: 3831111046


-----------------------------------------------------------------------------------------------------------------
| Id  | Operation           | Name | Starts | E-Rows | A-Rows |   A-Time   | Buffers |  OMem |  1Mem | Used-Mem |
-----------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT    |      |      1 |        |     10 |00:00:00.02 |     119 |       |       |          |
|   1 |  MERGE JOIN         |      |      1 |     10 |     10 |00:00:00.02 |     119 |       |       |          |
|   2 |   SORT JOIN         |      |      1 |      1 |      1 |00:00:00.01 |      15 |  2048 |  2048 | 2048  (0)|
|*  3 |    TABLE ACCESS FULL| T3   |      1 |      1 |      1 |00:00:00.01 |      15 |       |       |          |
|*  4 |   SORT JOIN         |      |      1 |  10000 |     10 |00:00:00.02 |     104 |   974K|   535K|  865K (0)|
|   5 |    TABLE ACCESS FULL| T4   |      1 |  10000 |  10000 |00:00:00.01 |     104 |       |       |          |
-----------------------------------------------------------------------------------------------------------------


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


   3 - filter("T3"."N"=1100)
   4 - access("T3"."ID"="T4"."T3_ID")
       filter("T3"."ID"="T4"."T3_ID")
使用 hint 让执行计划以 T4 作为驱动表
SQL> select /*+ leading(t4) use_merge(t3) */ *
  2  from t3, t4
  3  where t3.id = t4.t3_id and t3.n = 1100;


10 rows selected.


SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last'));


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


SQL_ID  gxuwn06y1c1az, child number 0
-------------------------------------
select /*+ leading(t4) use_merge(t3) */ * from t3, t4 where t3.id =
t4.t3_id and t3.n = 1100


Plan hash value: 875334572


-----------------------------------------------------------------------------------------------------------------
| Id  | Operation           | Name | Starts | E-Rows | A-Rows |   A-Time   | Buffers |  OMem |  1Mem | Used-Mem |
-----------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT    |      |      1 |        |     10 |00:00:00.04 |     119 |       |       |          |
|   1 |  MERGE JOIN         |      |      1 |     10 |     10 |00:00:00.04 |     119 |       |       |          |
|   2 |   SORT JOIN         |      |      1 |  10000 |   1001 |00:00:00.04 |     104 |   974K|   535K|  865K (0)|
|   3 |    TABLE ACCESS FULL| T4   |      1 |  10000 |  10000 |00:00:00.01 |     104 |       |       |          |
|*  4 |   SORT JOIN         |      |   1001 |      1 |     10 |00:00:00.01 |      15 |  2048 |  2048 | 2048  (0)|
|*  5 |    TABLE ACCESS FULL| T3   |      1 |      1 |      1 |00:00:00.01 |      15 |       |       |          |
-----------------------------------------------------------------------------------------------------------------


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


   4 - access("T3"."ID"="T4"."T3_ID")
       filter("T3"."ID"="T4"."T3_ID")
   5 - filter("T3"."N"=1100)
从返回的执行计划结果中我们可以看到:
1. 以 T3 为驱动表和以 T4 为驱动表, 两者的 cost (A-Time) 和 buffers 都差不多

2. 以 T3 为驱动表时, T3 访问一次, T4 也是访问一次; 以 T4 为驱动表时, T4 访问一次, T3 也是访问一次

3. 需要排序, 如果 PGA 空间重足时在 PGA 中排序, 不如果不足则交换到磁盘上排序


另外, 在执行计划中有几个统计信息列 0Mem, 1Mem, Use_Mem 需要介绍一下
  • 0Mem 指的是预计在 PGA 中排序需要的内存大小
  • 1Mem 指的是当内存大小(PGA)不足以进行排序, 预计将数据一次交换到磁盘空间的内存大小
  • Used-Mem 指的是执行时实际使用的内存大小, 其中括号中的数字代表进行磁盘交换的次数, 0 代表没有进行磁盘交换

三. sort merge joins连接(排序合并连接) 优化
SQL> select /*+ leading(t3) use_merge(t4) */ *
  2  from t3, t4
  3  where t3.id = t4.t3_id and t3.n = 1100 and t4.n = 10034;


SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last'));


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


SQL_ID  bg9h60c7ak3ud, child number 0
-------------------------------------
select /*+ leading(t3) use_merge(t4) */ * from t3, t4 where t3.id =
t4.t3_id and t3.n = 1100 and t4.n = 10034


Plan hash value: 3831111046


-----------------------------------------------------------------------------------------------------------------
| Id  | Operation           | Name | Starts | E-Rows | A-Rows |   A-Time   | Buffers |  OMem |  1Mem | Used-Mem |
-----------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT    |      |      1 |        |      1 |00:00:00.01 |     119 |       |       |          |
|   1 |  MERGE JOIN         |      |      1 |      1 |      1 |00:00:00.01 |     119 |       |       |          |
|   2 |   SORT JOIN         |      |      1 |      1 |      1 |00:00:00.01 |      15 |  2048 |  2048 | 2048  (0)|
|*  3 |    TABLE ACCESS FULL| T3   |      1 |      1 |      1 |00:00:00.01 |      15 |       |       |          |
|*  4 |   SORT JOIN         |      |      1 |      1 |      1 |00:00:00.01 |     104 |  2048 |  2048 | 2048  (0)|
|*  5 |    TABLE ACCESS FULL| T4   |      1 |      1 |      1 |00:00:00.01 |     104 |       |       |          |
-----------------------------------------------------------------------------------------------------------------


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


   3 - filter("T3"."N"=1100)
   4 - access("T3"."ID"="T4"."T3_ID")
       filter("T3"."ID"="T4"."T3_ID")
   5 - filter("T4"."N"=10034)


SQL> create index t4_n on t4(n);


Index created.


SQL> select /*+ leading(t3) use_merge(t4) */ *
  2  from t3, t4
  3  where t3.id = t4.t3_id and t3.n = 1100 and t4.n = 10034;


SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last'));


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


SQL_ID  bg9h60c7ak3ud, child number 0
-------------------------------------
select /*+ leading(t3) use_merge(t4) */ * from t3, t4 where t3.id =
t4.t3_id and t3.n = 1100 and t4.n = 10034


Plan hash value: 1501658231


------------------------------------------------------------------------------------------------------------------------------------
| Id  | Operation                     | Name | Starts | E-Rows | A-Rows |   A-Time   | Buffers | Reads  |  OMem |  1Mem | Used-Mem |
------------------------------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |      |      1 |        |      1 |00:00:00.01 |      18 |      1 |       |       |          |
|   1 |  MERGE JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |      18 |      1 |       |       |          |
|   2 |   SORT JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |      15 |      0 |  2048 |  2048 | 2048  (0)|
|*  3 |    TABLE ACCESS FULL          | T3   |      1 |      1 |      1 |00:00:00.01 |      15 |      0 |       |       |          |
|*  4 |   SORT JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |       3 |      1 |  2048 |  2048 | 2048  (0)|
|   5 |    TABLE ACCESS BY INDEX ROWID| T4   |      1 |      1 |      1 |00:00:00.01 |       3 |      1 |       |       |          |
|*  6 |     INDEX RANGE SCAN          | T4_N |      1 |      1 |      1 |00:00:00.01 |       2 |      1 |       |       |          |
------------------------------------------------------------------------------------------------------------------------------------


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


   3 - filter("T3"."N"=1100)
   4 - access("T3"."ID"="T4"."T3_ID")
       filter("T3"."ID"="T4"."T3_ID")
   6 - access("T4"."N"=10034)


SQL> create index t3_n on t3(n);


Index created.


SQL> select /*+ leading(t3) use_merge(t4) */ *
  2  from t3, t4
  3  where t3.id = t4.t3_id and t3.n = 1100 and t4.n = 10034;


SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last'));


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


SQL_ID  bg9h60c7ak3ud, child number 0
-------------------------------------
select /*+ leading(t3) use_merge(t4) */ * from t3, t4 where t3.id =
t4.t3_id and t3.n = 1100 and t4.n = 10034


Plan hash value: 1827980052


------------------------------------------------------------------------------------------------------------------------------------
| Id  | Operation                     | Name | Starts | E-Rows | A-Rows |   A-Time   | Buffers | Reads  |  OMem |  1Mem | Used-Mem |
------------------------------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |      |      1 |        |      1 |00:00:00.01 |       6 |      1 |       |       |          |
|   1 |  MERGE JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |       6 |      1 |       |       |          |
|   2 |   SORT JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |       3 |      1 |  2048 |  2048 | 2048  (0)|
|   3 |    TABLE ACCESS BY INDEX ROWID| T3   |      1 |      1 |      1 |00:00:00.01 |       3 |      1 |       |       |          |
|*  4 |     INDEX RANGE SCAN          | T3_N |      1 |      1 |      1 |00:00:00.01 |       2 |      1 |       |       |          |
|*  5 |   SORT JOIN                   |      |      1 |      1 |      1 |00:00:00.01 |       3 |      0 |  2048 |  2048 | 2048  (0)|
|   6 |    TABLE ACCESS BY INDEX ROWID| T4   |      1 |      1 |      1 |00:00:00.01 |       3 |      0 |       |       |          |
|*  7 |     INDEX RANGE SCAN          | T4_N |      1 |      1 |      1 |00:00:00.01 |       2 |      0 |       |       |          |
------------------------------------------------------------------------------------------------------------------------------------


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


   4 - access("T3"."N"=1100)
   5 - access("T3"."ID"="T4"."T3_ID")
       filter("T3"."ID"="T4"."T3_ID")
   7 - access("T4"."N"=10034)
从上面的执行计划中可以看出, 全表扫描后最后使用的 buffer 为 119, 在一个表上建立索引使用索引范围扫描后 buffer 为 18, 在两个表上建立的索引使用索引范围扫描后 buffer 为 6. 由此可以见, 在表的谓词条件上如果有索引的话, 将会提高执行效率.

此外, 由于 sort merge joins 需要先在 PGA 中进行排序,, 如果 PGA 空间不足, 就会将数据交换到磁盘上进行排序。由于, 磁盘相对于内存来说是慢速设备,因此在磁盘上排序会比在内存上排序慢, 另外排序排序消耗的时间还需要加上数据在内存和磁盘上传输的时间,因此尽可能减少磁盘排序的次数也就会提高执行效率, 有两种方法会减少磁盘排序:

1. 增大 PGA 的大小, 如果是 oracle 10g,需要增加参数 pga_aggregate_target 的大小,如果是 oracle 11g,则增加 memory_target 的大小

2. 减少排序的数据量, 一些不需要的字段就不要写在 select 后面


四. 小结

遇到 sql 调优时,如果执行计划显示表的连接方式是 sort merge join:

首先,看看 sql 语句是不是表的连接方式有没有可能转换为 hash join(等值连接条件)

其次,只能使用 sort merge join 时看看表的谓词条件上是不是有索引

最后,看看执行计划排序占用的内存大小是不是在磁盘上有排序, 是不是能够避免在磁盘上排序




参考: <<基于 oracle 的 sql 优化>>

          <<收获, 不止 oracle>>

         <<Troubleshooting oracle performance>>

oracle 表连接 - sort merge joins 排序合并连接

标签:merge join   sort join   排序合并连接   

原文地址:http://blog.csdn.net/dataminer_2007/article/details/41907581

(0)
(0)
   
举报
评论 一句话评论(0
登录后才能评论!
© 2014 mamicode.com 版权所有  联系我们:gaon5@hotmail.com
迷上了代码!