>This message is related to the performance issue threads that were created a few days ago.
>
>I made some robust test this morning on my 100000 records tables. Here's my result related to high volume tables.
>
>2.1.2 SCAN was slowing down the test drastically. I was using a SCAN to replace a value in a few records. I now have to use a SQL and apply the SCAN on the SQL and use the SEEK to be able to apply my REPLACE
>
What about REPLACE ALL FOR?
>3.1 If I delete a record in a child table, I just put a negative value of my record so the relation will be lost resulting in the same effect as SET DELETE ON. Yes, I still apply DELETE on that record so it will remove on the pack or will be recycle on the next ADD
>
>3.2 On my pick lists, I add to add the FOR UniqueId>0 so I won't get the deleted records
This is not consistent with INSERT TRIGGER for Child table which is usually RESTRICTED.
Edward Pikman
Independent Consultant