>>>>>>In the brief period when I used DBC/DBF I never used the RI or triggers. Didn't trust them. (Steve Sawyer once gave a great presentation on why that distrust was well-founded)
>>>>>>
>>>>>>Now use them even less :-) but a couple quesitons
>>>>>>
>>>>>>Are there any circumstances when they do no fire.
>>>>>>
>>>>>>For example ; Sql insert / delete from a table. Should it fire the RI code to delete children?
>>>>>>
>>>>>>Browse, mark record as deleted. RI fire?
>>>>>>
>>>>>>Manual insert / update ?
>>>>>>
>>>>>>TIA
>>>>>
>>>>>I believe in all these cases the RI triggers should fire. Did you find a case when it's not so?
>>>>
>>>>I'll have to retest, but I think if one browses a parent, marks a row for delete, the casade delete does not happen in children but I will have to test again.
>>>
>>>Try re-testing after moving a record pointer.
>>
>>Maybe that will clear up the problem but it shouldn't be necessary to do that. The cascade to the child table should occur immediately following a DELETE in the parent table.
>
>The trigger only fires when the record is updated on the table. This means when a row if row- or table buffered, you first need to issue a TABLEUPDATE(). Otherwise, I've never encountered the problem as described here.

Yes, I know about TABLEUPDATE(). My point was that the deletes in the parent and child tables should occur "simultaneously." There should be no need to move the record pointer.