Division in Haskell
While learning about division in Haskell, I realized that the concept is not as trivial as I initially wanted it to be. There are subtle differences between the functions that can easily trip someone who’s not aware of them.
quot performs integer division that rounds towards zero.
div is like
quot, but it rounds below zero—negative infinity.
rem returns the remainder of a division.
mod, on the other hand, performs modular arithmetic.
In the GHCi interaction below,
quotRem returns a tuple of the application of
rem to its arguments, while
divMod returns a tuple of the application
mod to its arguments.
Prelude> quotRem 13 5 (2,3) Prelude> quotRem (-13) 5 (-2,-3) Prelude> quotRem 13 (-5) (-2,3) Prelude> quotRem 5 13 (0,5) Prelude> quotRem (-5) 13 (0,-5) Prelude> quotRem 5 (-13) (0,5) Prelude> divMod 13 5 (2,3) Prelude> divMod (-13) 5 (-3,2) Prelude> divMod 13 (-5) (-3,-2) Prelude> divMod 5 13 (0,5) Prelude> divMod (-5) 13 (-1,8) Prelude> divMod 5 (-13) (-1,-8)
Giving special attention to negative numbers, here are some observations about it:
divModbehave the same, if all the arguments are positive.
0, if the dividend is less than the divisor.
remfollows the sign of the dividend.
remreturns the dividend, if the dividend is less than the divisor.
divrounds off the divisor to the negative infinity, if either the dividend or divisor is negative.
modfollows the sign of the divisor.
0, if the dividend is less than the divisor, and both arguments are positive.
modreturns the dividend, if the dividend is less than the divisor, and both arguments are positive.
-1if the dividend is negative, and its absolute value is less than the divisor.
0if the dividend is
0, and the divisor is not zero.
modreturns the difference of the absolute values of the divisor and the dividend, following the sign of the divisor, if either the divisor or the dividend is negative.