mysql> explain select * from testde1 where testde1.id in(select id from testde2); +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | 1 | PRIMARY | testde1 | NULL | ALL | NULL | NULL | NULL | NULL | 15 | 100.00 | Using where | | 2 | DEPENDENT SUBQUERY | testde2 | NULL | ALL | NULL | NULL | NULL | NULL | 2 | 50.00 | Using where | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ 2 rows in set, 1 warning (0.00 sec)
Note (Code 1003): /* select#1 */ select `test`.`testde1`.`id` AS `id` from `test`.`testde1` where (`test`.`testde1`.`id`,(/* select#2 */ select 1 from `test`.`testde2` where ((`test`.`testde1`.`id`) = `test`.`testde2`.`id`)))
使用DEPENDENT SUBQUERY 關聯子查詢優化,這也是最慢的。這和 select * from testde1 where exists (select * from testde2 where testde1.id=testde2.id);的執行計劃完全一致， testde1大表必須作為驅動表 mysql> explain select * from testde1 where exists (select * from testde2 where testde1.id=testde2.id); +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | 1 | PRIMARY | testde1 | NULL | ALL | NULL | NULL | NULL | NULL | 15 | 100.00 | Using where | | 2 | DEPENDENT SUBQUERY | testde2 | NULL | ALL | NULL | NULL | NULL | NULL | 2 | 50.00 | Using where | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ 2 rows in set, 2 warnings (0.00 sec)
實際就是下面的執行計劃：
mysql> explain delete from testde1 where id in (select id from testde2); +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ | 1 | DELETE | testde1 | NULL | ALL | NULL | NULL | NULL | NULL | 15 | 100.00 | Using where | | 2 | DEPENDENT SUBQUERY | testde2 | NULL | ALL | NULL | NULL | NULL | NULL | 2 | 50.00 | Using where | +----+--------------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+ 2 rows in set (0.00 sec)
這里我們看到小表testde2做了驅動表。 最后來說明一下這個報錯： mysql> delete from testde1 where id in(select testde1.id from testde1,testde2 where testde1.id=testde2.id ); ERROR 1093 (HY000): You can't specify target table 'testde1' for update in FROM clause 我們先不管他有沒有意義，這個報錯再手冊上叫做ER_UPDATE_TABLE_USED，我們首先來分析一下這個報錯 這樣的delete會進行exists展開那么testde1既是修改條件的來源也是修改的對象，這樣是不允許的。那么如何修改呢？ 實際上就需要select testde1.id from testde1,testde2 where testde1.id=testde2.id 的結果保存在一個臨時表中， 不要exists展開，手冊中給出的方法是 方法一、建立一個algorithm=temptable 的視圖 方法二、建立一個普通視圖同時修改SET optimizer_switch = 'derived_merge=off';
其目的都在于不展開選取第二種方式測試： mysql> create view myt1 -> as -> select testde1.id from testde1,testde2 where testde1.id=testde2.id; Query OK, 0 rows affected (0.02 sec)
mysql> show status like '%tmp%'; +-------------------------+-------+ | Variable_name | Value | +-------------------------+-------+ | Created_tmp_disk_tables | 0 | | Created_tmp_files | 0 | | Created_tmp_tables | 2 | +-------------------------+-------+ 3 rows in set (0.01 sec)
看看執行計劃： mysql> explain delete from testde1 where id in (select * from myt1); +----+--------------------+------------+------------+----------------+---------------+-------------+---------+------+------+----------+----------------------------------------------------+ | id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra | +----+--------------------+------------+------------+----------------+---------------+-------------+---------+------+------+----------+----------------------------------------------------+ | 1 | DELETE | testde1 | NULL | ALL | NULL | NULL | NULL | NULL | 13 | 100.00 | Using where | | 2 | DEPENDENT SUBQUERY | | NULL | index_subquery | | | 5 | func | 2 | 100.00 | Using index | | 3 | DERIVED | testde2 | NULL | ALL | NULL | NULL | NULL | NULL | 2 | 100.00 | NULL | | 3 | DERIVED | testde1 | NULL | ALL | NULL | NULL | NULL | NULL | 13 | 10.00 | Using where; Using join buffer (Block Nested Loop) | +----+--------------------+------------+------------+----------------+---------------+-------------+---------+------+------+----------+----------------------------------------------------+ 4 rows in set (0.00 sec)
