David Garamond wrote:
> insert into t1 values (3,char(32));
> Record 3, 5, and 7 is rejected due to duplicate value.
>
> +----+--------------+------+
> | id | length(code) | code |
> +----+--------------+------+
> | 1 | 0 | |
> | 2 | 1 | 0 |
> | 4 | 2 | 0 0 |
> | 6 | 2 | 32 0 |
> +----+--------------+------+
> 4 rows in set (0.00 sec)
>
> It seems the string in CHAR/VARCHAR is stored as null-terminated ala C.
> But why is record 3 rejected too?
Ugh, never mind. I forgot about the automatic trailing blanks removal
feature. Well, since trailing blanks are always removed, it means
CHAR/VARCHAR could never store binary data then...
Is there a 128-bit datatype planned? It would be handy to store
GUID/UUID or IPv6 addresses.
--
dave
