[BioSQL-l] version NOT NULL

[Hilmar Lapp]

On Thursday, March 13, 2003, at 04:53  AM, Yves Bastide wrote:

> Aaron J Mackey wrote:
>> INT NOT NULL implies INT NOT NULL DEFAULT 0
>
> Nope, that's a MySQLism.

Exactly. No other decent RDBMS does this.

>
>> But I wouldn't mind adding that to be explicit about it.
>
> Is it worth it, if every adaptor pretty has to be explicit?
>

I think it's simply wrong. It pretends the schema would handle 
something that really the adaptor code needs to handle no matter what. 
If an adaptor doesn't, I'd rather see an exception thrown for an 
attempt to insert a NULL value there.

	-hilmar
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------