It would help spread the index out and not have hot blocks. eg instead of
indexing
10001, 10002, 10003,...
it would be
10001, 20001, 30001,...
Hence there would not be hot blocks if a lot of inserts were happening in a
short time.
Jim
"Ginger Mall" <bocgco_at_hongkong.com> wrote in message
news:3e6433d0$1_at_shknews01...
> I have read from books talking about "reverse index". I know it reverses
> the order of the key value. However I am not quite understand why this
kind
> of arrangement could help the B-Tree index to be more efficient. Could
> anyone explain this to me ?
>
> Thank you very much.
>
>
Received on Mon Mar 03 2003 - 23:20:47 CST
