Number vs. Number(p,s)

I am trying to decide whether I wan't to use floating point (NUMBER) or fixed point
(NUMBER(p, s)) datatypes for my database design. Apart from having more control with the fixed point option, is there any advantage memory wise? For instance, if I know that I have a number that will never be greater than, say
99999.99, should I bother and assign it to NUMBER(7,2), or just go ahead and make it a NUMBER with no scale or precision declared?

I just don't see the advantage of specifically declaring a precision and scale, when you can just make it a floating point. Especially, when you consider that you are making a guess (albeit an edjucated guess) as to the scale and precision of you column. If this guess turns out to be incorrect, and say you really should have chosen NUMBER(8,2), for example, your application will give you an error.

Any advice here would be appreciated.

Thomas Received on Sun May 13 2001 - 19:52:28 CDT