Is there a better way of getting the quotient and remainder without using the '/', '%', and '*' operators?
The majority of the time, I think, by far the best way of computing quotient and remainder is with the /
and %
operators. (It's their job, after all!)
If you need both quotient and remainder at the same time, there's also the div
function, which does exactly that. (Its return value is a struct containing quot
and rem
members.) It's a bit cumbersome to use, but it might be more efficient if it means executing a single divide instruction (which tends to give you both quotient and remainder at the machine level anyway) instead of two. (Or, these days, with a modern optimizing compiler, I suspect it wouldn't make any difference anyway.)
But no matter how you do it, there's always a question when it comes to negative numbers. If the dividend or the divisor is negative, what sign(s) should the quotient and remainder be? There are basically two answers. (Actually, there are more than two, but these are the two I think about.)
In Euclidean division, the remainder is always positive, and is therefore always in the range [0,divisor) (that is, between 0 and divisor-1).
In many programming languages (including C), the remainder always has the sign of the dividend.
See the Wikipedia article on modulo operation for much, much more information on these and other alternatives.
If your programming language gives you the second definition (as C does), but you want a remainder that's never negative, one way to fix it is just to do a regular division, test whether the remainder is negative, and if it is, increment the remainder by the divisor and decrement the quotient by 1:
quotient = x / y;
remainder = x % y;
if(remainder < 0) {
reminder += y;
quotient--;
}
This works because of the formal definition of integer division with remainder:
div(x, y) → q, r such that y × q + r = x
If you subtract 1 from q and add y to r, you get the same result, and it's still x.
Here's a sample program demonstrating three different alternatives. I've encapsulated the little "adjust quotient for nonnegative remainder" algorithm as the function euclid()
, which returns the same div_t
type as div()
does:
#include <stdio.h>
#include <stdlib.h>
div_t euclid(int, int);
int main()
{
int x = -7, y = 3;
printf("operators: q = %d, r = %d\n", x/y, x%y);
div_t qr = div(x, y);
printf("div: q = %d, r = %d\n", qr.quot, qr.rem);
qr = euclid(x, y);
printf("euclid: q = %d, r = %d\n", qr.quot, qr.rem);
}
div_t euclid(int x, int y)
{
div_t qr = div(x, y);
if(qr.rem < 0) {
qr.quot--;
qr.rem += y;
}
return qr;
}
To really explore this, you'll want to try things out for all four cases:
- positive dividend, positive divisor (the normal case, no ambiguity)
- negative dividend, positive divisor (the case I showed)
- positive dividend, negative divisor
- negative dividend, negative divisor