Mysql索引會失效的幾種情況分析

Mysql索引會失效的幾種情況分析

CREATE TABLE `members` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `name` varchar(100) COLLATE utf8mb4_unicode_ci NOT NULL COMMENT '用戶名',
  `password` varchar(255) COLLATE utf8mb4_unicode_ci DEFAULT NULL COMMENT '密碼',
  `qq` int(10) DEFAULT NULL COMMENT '密碼',
  `register_ip` varchar(255) COLLATE utf8mb4_unicode_ci DEFAULT NULL COMMENT '註冊IP',
  `wechat` varchar(255) COLLATE utf8mb4_unicode_ci DEFAULT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `name` (`name`),
  UNIQUE KEY `password` (`password`(191)) USING BTREE,
  KEY `qq` (`qq`)
) ENGINE=MyISAM AUTO_INCREMENT=286 DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;

1.explain 查看索引使用

• type列，連接類型。一個好的SQL語句至少要達到range級別。杜絕出現all級別。
• possible_keys：顯示可能應用在這張表中的索引，一個或多個。查詢涉及到的字段上若存在索引，則該索引獎盃列出，但不一定被查詢實際使用。
• key列，使用到的索引名。如果沒有選擇索引，值是NULL。可以採取強制索引方式。
• key_len列，索引長度。
• rows列，掃描行數。該值是個預估值。
• extra列，詳細說明。注意，常見的不太友好的值，如下：Using filesort，Using temporary。

各列表的值詳情：https://www.cnblogs.com/miskis/p/9081187.html

1.1type類型

type	解釋
null	–
system	const的特例，僅返回一條數據的時候。
const	查找主鍵索引，返回的數據至多一條（0或者1條）。 屬於精確查找
eq_ref	查找唯一性索引，返回的數據至多一條。屬於精確查找
ref	查找非唯一性索引，返回匹配某一條件的多條數據。屬於精確查找、數據返回可能是多條
range	查找某個索引的部分索引，一般在where子句中使用 < 、>、in、between等關鍵詞。只檢索給定範圍的行，屬於範圍查找
index	查找所有的索引樹，比ALL要快的多，因爲索引文件要比數據文件小的多。
ALL	不使用任何索引，進行全表掃描，性能最差。

從下到上逐漸變好，使用的索引至少要達到range 級別。

2.使用or / !=(<>) / not null / is null

or影響

mysql> explain select * from members where name = '1' or wechat = '1';
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
| id | select_type | table   | type | possible_keys      | key  | key_len | ref  | rows | Extra       |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
|  1 | SIMPLE      | members | ALL  | name,name_password | NULL | NULL    | NULL |    2 | Using where |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+

!=

mysql> explain select * from members where name != '1';
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
| id | select_type | table   | type | possible_keys      | key  | key_len | ref  | rows | Extra       |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
|  1 | SIMPLE      | members | ALL  | name,name_password | NULL | NULL    | NULL |    2 | Using where |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+

is null

mysql> explain select * from members where name is not null;
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
| id | select_type | table   | type | possible_keys      | key  | key_len | ref  | rows | Extra       |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
|  1 | SIMPLE      | members | ALL  | name,name_password | NULL | NULL    | NULL |    2 | Using where |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+

is not null

mysql> explain select * from members where name is not null;
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
| id | select_type | table   | type | possible_keys      | key  | key_len | ref  | rows | Extra       |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
|  1 | SIMPLE      | members | ALL  | name,name_password | NULL | NULL    | NULL |    2 | Using where |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+

3. 字符串類型，where的時候帶引號

mysql> explain select * from members where name = '1';
+----+-------------+---------+-------+--------------------+------+---------+-------+------+-------+
| id | select_type | table   | type  | possible_keys      | key  | key_len | ref   | rows | Extra |
+----+-------------+---------+-------+--------------------+------+---------+-------+------+-------+
|  1 | SIMPLE      | members | const | name,name_password | name | 402     | const |    1 |       |
+----+-------------+---------+-------+--------------------+------+---------+-------+------+-------+
1 row in set

mysql> explain select * from members where name = 1;
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
| id | select_type | table   | type | possible_keys      | key  | key_len | ref  | rows | Extra       |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+
|  1 | SIMPLE      | members | ALL  | name,name_password | NULL | NULL    | NULL |    2 | Using where |
+----+-------------+---------+------+--------------------+------+---------+------+------+-------------+

4. 不建議使用%前綴模糊查詢

例如LIKE“%name”或者LIKE“%name%”，這種查詢會導致索引失效而進行全表掃描。
但是可以使用LIKE “name%”。

mysql> explain select name from members where name like '%1';
+----+-------------+---------+-------+---------------+------+---------+------+------+--------------------------+
| id | select_type | table   | type  | possible_keys | key  | key_len | ref  | rows | Extra                    |
+----+-------------+---------+-------+---------------+------+---------+------+------+--------------------------+
|  1 | SIMPLE      | members | index | NULL          | name | 402     | NULL |    2 | Using where; Using index |
+----+-------------+---------+-------+---------------+------+---------+------+------+--------------------------+
1 row in set

mysql> explain select name from members where name like '1%';
+----+-------------+---------+-------+--------------------+------+---------+------+------+--------------------------+
| id | select_type | table   | type  | possible_keys      | key  | key_len | ref  | rows | Extra                    |
+----+-------------+---------+-------+--------------------+------+---------+------+------+--------------------------+
|  1 | SIMPLE      | members | index | name,name_password | name | 402     | NULL |    2 | Using where; Using index |
+----+-------------+---------+-------+--------------------+------+---------+------+------+--------------------------+

那如何查詢%name%？答案：使用全文索引。

# 更新全文索引
ALTER TABLE `members` ADD FULLTEXT INDEX `name` (`name`);
mysql> explain select name from members where match(name) against ('1' in boolean mode);
+----+-------------+---------+----------+---------------+------+---------+-----+------+-------------+
| id | select_type | table   | type     | possible_keys | key  | key_len | ref | rows | Extra       |
+----+-------------+---------+----------+---------------+------+---------+-----+------+-------------+
|  1 | SIMPLE      | members | fulltext | name          | name | 0       |     |    1 | Using where |
+----+-------------+---------+----------+---------------+------+---------+-----+------+-------------+

5.索引列上少計算

不在索引列上做任何操作（計算、函數、(自動or手動)類型轉換），會導致索引失效而轉向全表掃描
其中後面那句varchar引號不能丟也是如此，如果是丟了，則會進行自動類型轉換，就相當於計算。

待續。。。。。