Re: Foreign Key Problems

jfr_at_trwlasd.com wrote:
> bs writes
> > Just a little technical note about FK constraints in V6. It's true
> > that they're not enforced in DML, but the actually ARE enforced in DDL.
> > Just try DROPping the Dept table if there's a FK pointing to it from the
> > Emp table and you'll see for yourself.
>
> Duh :-) This was one of the stupidest things Oracle EVER did, IMHO.
>
> They implemented the SYNTAX of this feature in V6 so they could claim
> SYNTACTICAL compatibility with the ANSI SQL Referential Integrity
> Enhancement (1989) and then didn't implement the functional capability so
> that it would useful in any way, shape or form. I would like to know
> exactly WHAT good comes of enforcing the constraints on the table
> definitions but not on the data????
>

• omitted
  >
  > Now, so that I won't get flamed for not saying SOMETHING nice, I will say
  > that V7 has an EXCELLENT referential constraints implementation that works
  > and works well. So, when they finally got around to it, they indeed put
  > in a mighty fine "engine".
  >
  Agreed. The reason Oracle _probably_ (imnsho) did it that way was: 1) the marketing types wanted that compatibility statement; and 2) it gave a growth path so that at least the database could be built properly 3) Sql*Forms could create it's own triggers based upon them. This is why ddl probably enforces them... so that the form triggers remain consistant.

--
===========================================================================
Thierry Lach                           | Sufficiently superior technology |
Work: thierry_at_eccdb1.pms.ford.com      | is indistinguishable from magic. |
Home: curlie!thierry_at_sycom.mi.org      |==================================|
#include <std.disclaimer>              |         We Do Magic Here!        |
===========================================================================