MySQL索引對NULL值的處理

 

# 索引不會包含有NULL值的列
只要列中包含有NULL值都將不會被包含在索引中，復合索引中只要有一列含有NULL值，那么這一列對於此復合索引就是無效的。所以我們在數據庫設計時不要讓字段的默認值為NULL。

在很多庫表設計規范、某某軍規的文章中，是不是經常會看到類似這樣的內容。小編也經常看到這樣的內容，並且在編寫規范的時候，准備也把這一條加進去。但在按部就班之余，小編抽空驗證了一下，發現事實卻並非如此！

小編使用的MySQL版本是社區版 5.7.21

 

新建測試表 t1，插入不含NULL值得100行數據，然后插入1行帶NULL的數據 insert into t1(id) values(101); 表中有主鍵id，索引a

CREATE TABLE `t1` (
  `id` int(11) NOT NULL,
  `a` int(11) DEFAULT NULL,
  `b` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`),
  KEY `a` (`a`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4

 

測試1，包含NULL單列索引的查詢，可以看到即使是查找 IS NULL的行，也是可以用上索引的

測試1：

desc select * from t1 where a > 82;
+----+-------------+-------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+
| id | select_type | table | partitions | type  | possible_keys | key  | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+-------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | t1    | NULL       | range | a             | a    | 5       | NULL |   18 |   100.00 | Using index condition |
+----+-------------+-------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+

desc select * from t1 where a is NULL;
+----+-------------+-------+------------+------+---------------+------+---------+-------+------+----------+-----------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref   | rows | filtered | Extra                 |
+----+-------------+-------+------------+------+---------------+------+---------+-------+------+----------+-----------------------+
|  1 | SIMPLE      | t1    | NULL       | ref  | a             | a    | 5       | const |    1 |   100.00 | Using index condition |
+----+-------------+-------+------------+------+---------------+------+---------+-------+------+----------+-----------------------+

desc select * from t1 where a = 20 or a is null; 
+----+-------------+-------+------------+-------------+---------------+--------+---------+-------+------+----------+--------------------------+
| id | select_type | table | partitions | type        | possible_keys | key    | key_len | ref   | rows | filtered | Extra                    |
+----+-------------+-------+------------+-------------+---------------+--------+---------+-------+------+----------+--------------------------+
|  1 | SIMPLE      | t1    | NULL       | ref_or_null | idx_ab        | idx_ab | 5       | const |    2 |   100.00 | Using where; Using index |
+----+-------------+-------+------------+-------------+---------------+--------+---------+-------+------+----------+--------------------------+

注意對 NULL 值的檢索只能使用 is null / is not null / <=>，不能使用=,<,>這樣的運算符（mysql中可以用a <=> NULL 表示查找 a is NULL'的行）

 

測試2，包含NULL復合索引的查詢，首先加一個復合索引 alter table t1 drop index a,add index idx_ab(a,b); 可以看到不管是指定 a is null ，或者指定 b is null ，都可以利用上索引 idx_ab（key_len 可以看出）

測試2：

desc select * from t1 where a=50 and b>20;
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+
| id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows | filtered | Extra                    |
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+
|  1 | SIMPLE      | t1    | NULL       | range | idx_ab        | idx_ab | 10      | NULL |    1 |   100.00 | Using where; Using index |
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+

desc select * from t1 where a=50 and b is null;
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+
| id | select_type | table | partitions | type | possible_keys | key    | key_len | ref         | rows | filtered | Extra                    |
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+
|  1 | SIMPLE      | t1    | NULL       | ref  | idx_ab        | idx_ab | 10      | const,const |    1 |   100.00 | Using where; Using index |
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+

desc select * from t1 where a is null and b>20;     
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+
| id | select_type | table | partitions | type  | possible_keys | key    | key_len | ref  | rows | filtered | Extra                    |
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+
|  1 | SIMPLE      | t1    | NULL       | range | idx_ab        | idx_ab | 10      | NULL |    1 |   100.00 | Using where; Using index |
+----+-------------+-------+------------+-------+---------------+--------+---------+------+------+----------+--------------------------+

desc select * from t1 where a is null and b is null;   
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+
| id | select_type | table | partitions | type | possible_keys | key    | key_len | ref         | rows | filtered | Extra                    |
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+
|  1 | SIMPLE      | t1    | NULL       | ref  | idx_ab        | idx_ab | 10      | const,const |    1 |   100.00 | Using where; Using index |
+----+-------------+-------+------------+------+---------------+--------+---------+-------------+------+----------+--------------------------+

 

由此，只要列中包含有NULL值都將不會被包含在索引中，復合索引中只要有一列含有NULL值，那么這一列對於此復合索引就是無效的。所以我們在數據庫設計時不要讓字段的默認值為NULL。這句的前半句是不對的（可參考官網說明： https://dev.mysql.com/doc/refman/5.7/en/is-null-optimization.html），但是后半句的結論確是可以采納的。

雖然MySQL可以在含有null的列上使用索引，但不代表null和其他數據在索引中是一樣的。不建議列上允許為空，最好限制 not null ，並設置一個默認值，比如0和''空字符串等，如果是datetime類型，可以設置成'1970-01-01 00:00:00'這樣的值。對MySQL來說，null 是一個特殊的值，Conceptually, NULL means “a missing unknown value” and it is treated somewhat differently from other values。 對null做算術運算的結果都是null，count時不會包括null行，null 比空字符串需要更多的存儲空間等。

 

附：上面說到可用通過 key_len 看出使用了索引列的個數，a,b 都是 int 類型，4 byte，為什么 key_len 是 5 byte 和 10 byte 呢？是因為如果索引列定義時允許NULL，其key_len還需要再加 1 bytes. 參考好友王的文章，可以移步我們的站點查看詳情： http://www.fordba.com/spend-10-min-to-understand-how-mysql-use-index.html