我有一个简单的mysql表,我想用它来进行每日统计跟踪。在将一些信息放入其中并测试出我在其上使用的充分查询之后,它没有使用我的索引!我无法弄清楚原因。
我尝试输入更多测试数据--200行。我也尝试更改select *来选择created,value1。没有骰子。
mysql> show create table stat_general\G
*************************** 1. row ***************************
Table: stat_general
Create Table: CREATE TABLE `stat_general` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`created` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
`nameid` smallint(6) NOT NULL,
`value1` int(11) NOT NULL,
`value2` int(11) DEFAULT NULL,
PRIMARY KEY (`id`),
KEY `created` (`created`,`nameid`,`value2`)
) ENGINE=MyISAM AUTO_INCREMENT=49 DEFAULT CHARSET=utf8
1 row in set (0.00 sec)
mysql> select * from stat_general;
+----+---------------------+--------+--------+--------+
| id | created | nameid | value1 | value2 |
+----+---------------------+--------+--------+--------+
| 1 | 2011-06-09 09:43:41 | 1 | 511803 | NULL |
| 2 | 2011-06-09 09:43:41 | 2 | 44 | NULL |
| 3 | 2011-06-09 09:43:41 | 3 | 9128 | NULL |
| 4 | 2011-06-09 09:43:41 | 4 | 219 | NULL |
| 5 | 2011-06-09 09:43:41 | 5 | 0 | NULL |
| 6 | 2011-06-09 09:43:41 | 6 | 0 | NULL |
| 7 | 2011-06-09 09:43:41 | 7 | 0 | NULL |
| 8 | 2011-06-09 09:43:41 | 8 | 0 | NULL |
| 9 | 2011-06-09 09:43:41 | 9 | 0 | NULL |
| 10 | 2011-06-09 09:43:41 | 10 | 140 | NULL |
| 11 | 2011-06-09 09:43:41 | 11 | 0 | NULL |
| 12 | 2011-06-09 09:43:41 | 12 | 146 | NULL |
| 13 | 2011-06-09 09:43:41 | 13 | 0 | NULL |
| 14 | 2011-06-09 09:43:41 | 14 | 1 | NULL |
| 15 | 2011-06-09 09:43:41 | 15 | 8981 | NULL |
| 16 | 2011-06-09 09:43:41 | 16 | 1 | -127 |
| 17 | 2011-06-09 09:43:41 | 16 | 2 | -2 |
| 18 | 2011-06-09 09:43:41 | 16 | 939 | -1 |
| 19 | 2011-06-09 09:43:41 | 16 | 146 | 1 |
| 20 | 2011-06-09 09:43:41 | 16 | 8011 | 3 |
| 21 | 2011-06-09 09:43:41 | 16 | 28 | 127 |
| 22 | 2011-06-09 09:43:41 | 16 | 1 | 128 |
| 23 | 2011-06-09 09:43:41 | 17 | 146 | 1 |
| 24 | 2011-06-09 09:43:41 | 18 | 146 | 1 |
| 25 | 2011-06-09 09:44:08 | 1 | 511803 | NULL |
| 26 | 2011-06-09 09:44:08 | 2 | 44 | NULL |
| 27 | 2011-06-09 09:44:08 | 3 | 9128 | NULL |
| 28 | 2011-06-09 09:44:08 | 4 | 219 | NULL |
| 29 | 2011-06-09 09:44:08 | 5 | 0 | NULL |
| 30 | 2011-06-09 09:44:08 | 6 | 0 | NULL |
| 31 | 2011-06-09 09:44:08 | 7 | 0 | NULL |
| 32 | 2011-06-09 09:44:08 | 8 | 0 | NULL |
| 33 | 2011-06-09 09:44:08 | 9 | 0 | NULL |
| 34 | 2011-06-09 09:44:08 | 10 | 140 | NULL |
| 35 | 2011-06-09 09:44:08 | 11 | 0 | NULL |
| 36 | 2011-06-09 09:44:08 | 12 | 146 | NULL |
| 37 | 2011-06-09 09:44:08 | 13 | 0 | NULL |
| 38 | 2011-06-09 09:44:08 | 14 | 1 | NULL |
| 39 | 2011-06-09 09:44:08 | 15 | 8981 | NULL |
| 40 | 2011-06-09 09:44:08 | 16 | 1 | -127 |
| 41 | 2011-06-09 09:44:08 | 16 | 2 | -2 |
| 42 | 2011-06-09 09:44:08 | 16 | 939 | -1 |
| 43 | 2011-06-09 09:44:08 | 16 | 146 | 1 |
| 44 | 2011-06-09 09:44:08 | 16 | 8011 | 3 |
| 45 | 2011-06-09 09:44:08 | 16 | 28 | 127 |
| 46 | 2011-06-09 09:44:08 | 16 | 1 | 128 |
| 47 | 2011-06-09 09:44:08 | 17 | 146 | 1 |
| 48 | 2011-06-09 09:44:08 | 18 | 146 | 1 |
+----+---------------------+--------+--------+--------+
48 rows in set (0.00 sec)
mysql> select * from stat_general where created>'2011-06-09 9:44' AND nameid=1;
+----+---------------------+--------+--------+--------+
| id | created | nameid | value1 | value2 |
+----+---------------------+--------+--------+--------+
| 25 | 2011-06-09 09:44:08 | 1 | 511803 | NULL |
+----+---------------------+--------+--------+--------+
1 row in set (0.00 sec)
mysql> explain select * from stat_general where created>'2011-06-09 9:44' AND nameid=1;
+----+-------------+--------------+------+---------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+--------------+------+---------------+------+---------+------+------+-------------+
| 1 | SIMPLE | stat_general | ALL | created | NULL | NULL | NULL | 48 | Using where |
+----+-------------+--------------+------+---------------+------+---------+------+------+-------------+
1 row in set (0.00 sec)
答案 0 :(得分:3)
有两个原因:
首先,你的桌子太小了。使用这几行,无论查询如何,打开整个表并逐个过滤行都会更快。
接下来,即使您有更多具有非常相似数据的行,您当前的索引也不够具有选择性。在这种情况下,不值得随机打开磁盘页面;按顺序打开它们并过滤掉无效的行会更快。
答案 1 :(得分:0)
只是为了咧嘴笑......你的查询主要是基于给定的“nameID”还是“Created”。如果NameID(即:特定人员发生了什么),我会将索引和标准改为
KEY `byName` ( `nameid`,`created`,`value2`)
然后你的查询......
where nameid = 1 AND created > '2011-06-09 9:44'
我不是肯定的,但是如果标准是匹配索引的自然顺序,它可以直接利用它来猜测索引...(特别是对于更大的数据集)...否则@Denis是正确的只是查询数据页面,因为它非常小,并且无论是否考虑应该使用哪个索引,都要吹过它们。