[MySQL优化案例]系列 -- 用TIMESTAMP类型取代INT和DATETIME

时间 2021-08-13

标签 mysql sql ide 测试 this spa unix 索引 get 博客栏目 MySQL 繁體版

原文原文链接

做/译者：叶金荣（Email:

），来源：http://imysql.cn，转载请注明做/译者和出处，而且不能用于商业用途，违者必究。

引言：在之前，我老是习惯用 INT UNSIGNED 来存储一个转换成Unix时间戳的时间值，认为这样作从索引，比较等角度来说，都会比较高效。如今咱们来对比下 TIMESTAMP 和 INT UNSIGNED 以及 DATETIME 这3种类型到底谁更好。

1. 准备

建立一个测试表：

mysql> CREATE TABLE `t` (
`d1` int(10) unsigned NOT NULL default '0',
`d2` timestamp NOT NULL default CURRENT_TIMESTAMP,
`d3` datetime NOT NULL,
KEY `d2` (`d2`),
KEY `d1` (`d1`),
KEY `d3` (`d3`)
);

而后建立一个存储过程填充数据：

mysql> DELIMITER //
CREATE PROCEDURE INS_T()
BEGIN
SET @i=1;
WHILE 0<1
DO
SET @i=@i+1;
INSERT INTO i VALUES (1199116800+@i, FROM_UNIXTIME(1199116800+@i), FROM_UNIXTIME(1199116800+@i));
END WHILE;
END;//
DELIMITER ;

时间戳 1199116800 表示 2008-01-01 这个时间点。而后运行存储过程，大概填充几十万条记录后，停止执行，由于上面的存储过程是个死循环，因此须要人工停止。
来看看到底有多少条记录了，以及索引状况：

mysql> select count(*) from t;
+----------+
| count(*) |
+----------+
|   924707 |
+----------+
mysql> analyze table t;
+--------+---------+----------+-----------------------------+
| Table  | Op      | Msg_type | Msg_text                    |
+--------+---------+----------+-----------------------------+
| test.t | analyze | status   | Table is already up to date |
+--------+---------+----------+-----------------------------+
mysql> show index from t;
+-------+------------+----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+
| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment |
+-------+------------+----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+
| t     |          1 | d2       |            1 | d2          | A         |      924707 |     NULL | NULL   |      | BTREE      |         |
| t     |          1 | d1       |            1 | d1          | A         |      924707 |     NULL | NULL   |      | BTREE      |         |
| t     |          1 | d3       |            1 | d3          | A         |      924707 |     NULL | NULL   |      | BTREE      |         |
+-------+------------+----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+

2. 对比

2.1 只检索一条记录

mysql> explain select * from t where d1 = 1199579155;
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
| id | select_type | table | type | possible_keys | key  | key_len | ref   | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
|  1 | SIMPLE      | t     | ref  | d1            | d1   | 4       | const |    1 |       |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
mysql> explain select * from t where d2 = '2008-01-06 08:25:55';
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
| id | select_type | table | type | possible_keys | key  | key_len | ref   | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
|  1 | SIMPLE      | t     | ref  | d2            | d2   | 4       | const |    1 |       |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
mysql> explain select * from t where d3 = '2008-01-06 08:25:55';
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
| id | select_type | table | type | possible_keys | key  | key_len | ref   | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+
|  1 | SIMPLE      | t     | ref  | d3            | d3   | 8       | const |    1 |       |
+----+-------------+-------+------+---------------+------+---------+-------+------+-------+

2.2 范围检索

mysql> explain select * from t where d1 >= 1199894400;
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
| id | select_type | table | type  | possible_keys | key  | key_len | ref  | rows   | Extra       |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
|  1 | SIMPLE      | t     | range | d1            | d1   | 4       | NULL | 121961 | Using where |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
mysql> explain select * from t where d2 >= from_unixtime(1199894400);
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
| id | select_type | table | type  | possible_keys | key  | key_len | ref  | rows   | Extra       |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
|  1 | SIMPLE      | t     | range | d2            | d2   | 4       | NULL | 121961 | Using where |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
mysql> explain select * from t where d3 >= from_unixtime(1199894400);
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
| id | select_type | table | type  | possible_keys | key  | key_len | ref  | rows   | Extra       |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+
|  1 | SIMPLE      | t     | range | d3            | d3   | 8       | NULL | 120625 | Using where |
+----+-------------+-------+-------+---------------+------+---------+------+--------+-------------+

小结：从上面的2次对比中能够看到，对 d1 或 d2 字段检索时的索引长度都是 4，由于 TIMESTAMP 其实是 4字节 的 INT 值。所以，实际应用中，基本上彻底能够采用 TIMESTAMP 来代替另外2种类型了，而且 TIMESTAMP 还能支持自动更新成当前最新时间，何乐而不为呢？