That is correct.
Use NOT NULL in the definition to force unique keys.
jon.barker@stripped wrote:
>>Description:
>>
> If a null column is used in a unique index then mysql fails to throw a
> key violation when multiple identical inserts (with nulls) are
> performed.
>
>>How-To-Repeat:
>>
> create table test (a int null, b int null);
> alter table test add unique index index_ab (a,b);
> insert into test (a,b) values(1,null);
> insert into test (a,b) values(1,null);
> insert into test (a,b) values(1,null);
> select * from test;
> +------+------+
> | a | b |
> +------+------+
> | 1 | NULL |
> | 1 | NULL |
> | 1 | NULL |
> +------+------+
> 3 rows in set (0.00 sec)
>
> insert into test (a,b) values(1,1);
> insert into test (a,b) values(1,1);
> ERROR 1062: Duplicate entry '1-1' for key 1
>
>>Fix:
>>
> No fix.
>
>>Submitter-Id: <submitter ID>
>>Originator: Jon Barker
>>Organization:
>>
>
>
>>MySQL support: none
>>Synopsis: NULLs don't work correctly in unique indexes.
>>Severity: serious
>>Priority:
>>Category: mysql
>>Class:
>>Release: mysql-3.23.41 (Official MySQL RPM)
>>
>
>>Environment:
>>
>
>
