key on more table

Hi,
do you know how can I declare a key on more tables?

For example, if I have the table TABLE1 and the table TABLE2, every one
with a field called key, how can OI declare in mysql that the key is
composed of TABLE1.key AND TABLE2.key?

Thanks a lot
Marco Righi

--
-------
email: thinkmarco@yahoo.it
-------


---
avast! Antivirus: In partenza messaggio pulito.
Virus Database (VPS): 0546-1, 15/11/2005
Controllato il: 15/11/2005 14.06.45
avast! - copyright (c) 1988-2005 ALWIL Software.
http://www.avast.com

myname@yahoo.it wrote:
> Hi,
> do you know how can I declare a key on more tables?
>
> For example, if I have the table TABLE1 and the table TABLE2, every one
> with a field called key, how can OI declare in mysql that the key is
> composed of TABLE1.key AND TABLE2.key?
>
> Thanks a lot
> Marco Righi
>

Macro,

You can't. The purpose of a key is to uniquely identify a row in a
single table.

--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attglobal.net
==================

Jerry Stuckle wrote:
> myname@yahoo.it wrote:
>
>>Hi,
>>do you know how can I declare a key on more tables?
>>
>>For example, if I have the table TABLE1 and the table TABLE2, every one
>>with a field called key, how can OI declare in mysql that the key is
>>composed of TABLE1.key AND TABLE2.key?
>>
>>Thanks a lot
>>Marco Righi
>>
>
>
> Macro,
>
> You can't. The purpose of a key is to uniquely identify a row in a
> single table.
>

s/uniquely//

myname@yahoo.it wrote:
> Hi,
> do you know how can I declare a key on more tables?
>
> For example, if I have the table TABLE1 and the table TABLE2, every one
> with a field called key, how can OI declare in mysql that the key is
> composed of TABLE1.key AND TABLE2.key?
>
> Thanks a lot
> Marco Righi
>
I don't know what you are trying to accomplish, but if you want a key
that appears as a field in more than one table and is still unique in
all those tables, you can introduce an identity table.

This identity table "generates" the keys with an auto-increment field
(primary key), and it is referrred to by foreign keys in the tables that
"share" the relation. Any time a foreign key field needs to be written,
a row is inserted in the identity table and its last genereated ID is used.
This is no hard check for uniqueness: the code that adds records must
generate an ID first. The uniqueness comes from the fact that all
generated IDs are different.

When I first encountered an ID-only table in my own database scheme, I
had given it a really hard thought. Somehow it didn't feel right. But
every other option would have given me a worse scheme and now I think
there can be situations where it is useful.

Best regards.