Currently all integers are converted to R's 32 bit integers. I think we should use doubles here with maybe a custom vctrs class to avoid overflow.
Bonus: think about using Arbitrary precision library such as rmpf if the integer is too large for doubles.
Currently all integers are converted to R's 32 bit integers. I think we should use doubles here with maybe a custom vctrs class to avoid overflow.
Bonus: think about using Arbitrary precision library such as rmpf if the integer is too large for doubles.