NUMBER DataType Question!

I have a simple but confusing question.

  In most of the Reference Book, the author usually declares number fields   to be NUMBER(d, f) where d = total no. of digits and f = decimal place.   Then, what is the difference between this and only NUMBER declaration?

I have this question because I am developing an system which involves a large amount of transaction (in number) and calculation. The numeric storage on number (amount, units, unit price...) in each transaction makes me restrict all NUMBER declaration to, say NUMBER(17,6). However, under this restriction, the calculated value (e.g. amount / unit price) become more and more unreliable. For example, I run the system for 2 cycle and the total amount loses $2 dollars finally! When I change the declaration to NUMBER, unreliability face-out and ,amazingly, the system seems to be fine (actually on its speed).

So, I want to clarify the above 2 declaration such that the unrestricted number declaration will (MUST) not make my system overflow in the future and continuing running!

thanks! Received on Sat Feb 14 1998 - 00:00:00 CET