MySQL在执行SQL查询时可能会用到临时表，一般而言，用到临时表就意味着性能较低。MySQL临时表可分：内存临时表和磁盘临时表，磁盘临时表简介MySQL在执行SQL查询时可能会用到临时表，一般而言，用到临时表就意味着性能较低。MySQL临时表可分：内存临时表和磁盘临时表，，磁盘临时表使用MyISAM存储，而内存临时表则使用Memory引擎。MySQL使用内存临时表来存放查询的中间结果集，如果中间结果集大于临时表的容量设定，又或者中间结果集含TEXT或BLOB列类型字段，则MySQL会把其转化为磁盘临时表。Linux平台缺省是/tmp,/tmp目录小的系统要注意啦。
临时表特性下面列举几个内存临时表的特性
1.表结构(.frm)在磁盘，数据在内存
2.缺省使用哈希索引
3.定长存储(BTW：即使是varchar也是定长)
4.只支持表锁
5.不支持TEXT和BLOB列类型
相关参数1 大小参数MIN{tmp_table_size,max_heap_table_size}mysql> show global variables like '%table_size';+---------------------+----------+| Variable_name | Value |+---------------------+----------+| max_heap_table_size | 16777216 || tmp_table_size | 16777216 |+---------------------+----------+2 rows in set (0.00 sec) 2 数量参数mysql> show global status like 'created_tmp%';+-------------------------+-------+| Variable_name | Value |+-------------------------+-------+| Created_tmp_disk_tables | 7 || Created_tmp_files | 6 || Created_tmp_tables | 90 |+-------------------------+-------+3 rows in set (0.00 sec)通过ps_helper我们可以细化这些临时表到底是谁的临时表mysql> SELECT query, exec_count, memory_tmp_tables, disk_tmp_tables, avg_tmp_tables_per_query, tmp_tables_to_disk_pct FROM statements_with_temp_tables LIMIT 5;+-------------------------------------------------------------------+------------+-------------------+-----------------+--------------------------+------------------------+| query | exec_count | memory_tmp_tables | disk_tmp_tables | avg_tmp_tables_per_query | tmp_tables_to_disk_pct |+-------------------------------------------------------------------+------------+-------------------+-----------------+--------------------------+------------------------+| SELECT IF ( ( `locate` ( ? , ` ... . `COMPRESSED_SIZE` ) ) DESC | 2 | 4 | 2 | 2 | 50 || SELECT IF ( ( `locate` ( ? , ` ... MPRESSED_SIZE` = ? ) , ? , ... | 2 | 4 | 2 | 2 | 50 || SELECT IF ( `isnull` ( `inform ... ` = `performance_schema` . ... | 2 | 4 | 2 | 2 | 50 || SELECT IF ( `isnull` ( `inform ... by_thread_by_event_name` . ... | 2 | 4 | 2 | 2 | 50 || SHOW FULL FIELDS FROM `stateme ... ` , `performance_schema` . ... | 2 | 4 | 2 | 2 | 50 |+-------------------------------------------------------------------+------------+-------------------+-----------------+--------------------------+------------------------+5 rows in set (0.00 sec)