Re: Problems with cascade deletes ?

Have you check the status of the FK it might be disabled and as such the RI is not inforced.

Also bear in mind that FK constriants require an index or you'll hit severe locking issues.

HTH Colin.
Serge Beaulé wrote in message <3565BAA5.E587429C_at_informission.ca>...
>We are using NT Workgroup 7.3.3 and occasionnaly find ourselves with
>orphan records
>in a table that is linked to its parent with a FK using a delete
>cascade.
>All we do is delete some records in a few tables and let Oracle delete
>the corresponding records in the child tables.
>And it's only one or two child table per parent, not more than one
>generation down.
>I can't say that our app is perfect but still I can't figure out how
>Oracle would permit the presence of these orphans after our delete
>operation. We do not update any PKs.
>It does not happen all the time, and of course when I trace it it does
>not. But we still have that situation in all the schemas where
>the operation is done.
>Would anyone know of a way for that situation to happen either because
>of an Oracle bug or a misuse of FK in our part.
>The only bizzare hint that we have is that the child table having orphan
>records is the only one where the parent PK column is generated from a
>sequence instead of character like the others.
>
>Thanks for all the help ! !
>
>Serge Beaulé, DBA
>Informission Group Inc.
>serge.beaule_at_informission.ca
>
Received on Sat May 23 1998 - 20:54:46 CDT