L
Laura Stern
Guest
Think of it this way. The DELETE is going to go through the same work regardless of how the record was found in the first place. So clearly if TABLE-SCAN makes finding each record faster, then the whole process will be faster. You seem to be thinking that if we use the index to find the record then deleting the record and the indexes will somehow be faster. That is not true. Well, maybe index blocks will already be in memory that wouldn't have been there. But these blocks needed to be loaded once regardless. So there would be no gain in performance overall.
Continue reading...
Continue reading...