On Nov 2, 2004, at 12:56 PM, Kumar S wrote:
> Hello Oliver,
> Thank you for your mail with suggestions. I decided
> to keep the con_exp_id as primary key because of the
> fact that a contact can do as many experiments as he
> can and one experiment can be done many contacts.
> Thus these two tables share a Many to Many
> relationship.
>
>
> Thus for other tables con_exp_id will be made as
> Foreign key .
>
> For example:
>
> Table: Chip_table
> chip_id (PK)
> con_exp_id (FK)
> ......
> .....
>
> Here it is easy for me to anchor con_exp_id as FK from
> con_exp_link table.
>
> Any other better alternative?
>
> Thanks
>
> Kumar
>
Kuman,
This implies that you will have a row in the chip table for every
con_exp row, so you wouldn't need the con_exp table at all? Probably
what you mean is that for every con_exp row, there are foreign keys
referencing the primary keys in your chip and contacts tables? That
gives you the many-to-many relationship and guarantees that you have an
entry in both the contacts and chip tables for every row in the con_exp
table.
Sean