MySQL 文件排序 &索引排序

排序操作是非常消耗cpu的操作，当系统设置不当或query取出的字段过多时，还可以造成MySQL不得不放弃优化后的排序算法，而使用较为古老的需要两次IO的读取表数据的排序算法，使排序效率非常低下。

利用索引进行排序操作，主要是利用了索引的有序性。在通过索引进行检索的过程中，就已经得到了有序的数据访问顺序，依次读取结果数据后就不需要进行排序操作，进而避免了此操作，提高了排序结果集的query性能。

很简单，尽量使用索引排序，这就对了。

做个实验，表结构如下

mysql> show create table artist \G
*************************** 1. row ***************************
       Table: artist
Create Table: CREATE TABLE `artist` (
  `artist_id` int(10) unsigned NOT NULL,
  `type` enum('Band','Person','Unknown','Combination') NOT NULL,
  `name` varchar(255) NOT NULL,
  `gender` enum('Male','Female') DEFAULT NULL,
  `founded` year(4) DEFAULT NULL,
  `country_id` smallint(5) unsigned DEFAULT NULL,
  PRIMARY KEY (`artist_id`),
  UNIQUE KEY `name` (`name`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1
1 row in set (0.00 sec)

mysql> Explain select name, founded from artist where name like 'AUSTRALIA%' order by founded \G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: artist
         type: range
possible_keys: name
          key: name
      key_len: 257
          ref: NULL
         rows: 22
        Extra: Using index condition; Using filesort
1 row in set (0.00 sec)

mysql> show session status like '%sort%';
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| Sort_merge_passes | 0     |
| Sort_range        | 0     |
| Sort_rows         | 0     |
| Sort_scan         | 0     |
+-------------------+-------+
4 rows in set (0.00 sec)

mysql> select name, founded from artist where name like 'AUSTRALIA%' order by founded \G
......
22 rows in set (0.00 sec)

mysql> show session status like '%sort%';
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| Sort_merge_passes | 0     |
| Sort_range        | 1     |
| Sort_rows         | 22    |
| Sort_scan         | 0     |
+-------------------+-------+
4 rows in set (0.00 sec)

几个参数

sort_merge_passes  由于sort buffer不够大，不得不将需要排序的数据进行分段，然后再通过sort merge的算法完成整个过程的merge总次数，一般整个参数用来参考sort buffer size 是否足够。

sort range session/global级别（单位：次） 通过range scan完成的排序总次数。

sort rows   session/global 级别（单位：row) 排序的总行数。

sort scan   通过扫描表完成的排序总次数。

索引排序，其中extra中的filesort不见了

mysql> Explain select name, founded from artist where name like 'AUSTRALIA%' order by name \G
*************************** 1. row ***************************
           id: 1
  select_type: SIMPLE
        table: artist
         type: range
possible_keys: name
          key: name
      key_len: 257
          ref: NULL
         rows: 22
        Extra: Using index condition
1 row in set (0.00 sec)

mysql> show session status like '%sort%';
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| Sort_merge_passes | 0     |
| Sort_range        | 0     |
| Sort_rows         | 0     |
| Sort_scan         | 0     |
+-------------------+-------+
4 rows in set (0.00 sec)

mysql> select name, founded from artist where name like 'AUSTRALIA%' order by name \G
......
22 rows in set (0.00 sec)

mysql> show session status like '%sort%';
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| Sort_merge_passes | 0     |
| Sort_range        | 0     |
| Sort_rows         | 0     |
| Sort_scan         | 0     |
+-------------------+-------+
4 rows in set (0.00 sec)