>My idea was that you could program your own cascade deletes, if you feel like you can do it more efficiently than the RI. You think not (I have the same opinion for my own stuff).
>
>If the slowness of cascading 4000 deletes to two child tables is expected, then I should probably cut the persistent relations with ALTER TABLE ...
>
>Your comment about the indices is a good one. How many _do_ you have?
>I don't know that there is any good "should take" rule. Max Rubels might, but I don't < s >.
>
>If your system depends on frequently deleting 4000 records...well maybe there is a way to design around it. What is the situation?

It won't be frequent; more like every few months, not counting testing. I opened a BROWSE and deleted individual parent records. Each delete and cascade took a second or so, suggesting that the whole delete would take a couple of hours if I hadn't closed them with the Task Manager. I expect that my program will have to cut the relations and restore them.