[FIXED] How to convert a binary fraction number into decimal fraction number in R?

Issue

I need to write a function that converts a binary fraction number into decimal fraction number in R. e.g. f(0.001) # 0.125

What I did: I searched for the related functions in R packages:

DescTools::BinToDec(0.001) # NA
DescTools::BinToDec("0.001") # NA
base::strtoi(0.001, base=2) # NA
base::strtoi("0.001", base=2) # NA

base::packBits(intToBits(0.001), "integer") # 0
base::packBits(intToBits("0.001"), "integer") # 0

compositions::unbinary(0.001) # 0.001
compositions::unbinary("0.001") # NA

I searched in SOF, found the following:

base2decimal <- function(base_number, base = 2) {
  split_base <- strsplit(as.character(base_number), split = "")
  return(sapply(split_base, function(x) sum(as.numeric(x) * base^(rev(seq_along(x) - 1)))))}
base2decimal(0.001) # NA
base2decimal("0.001") # NA

0.001 is:

(0 * 2^(-1)) + (0 * 2^(-2)) + (1 * 2^(-3)) # 0.125
(0 * 1/2) + (0 * (1/2)^2) + (1 * (1/2)^3) # 0.125
(0 * 0.5) + (0 * (0.5)^2) + (1 * 0.5^3) # 0.125

So, something like sum of the inner product (0,0,1) * (0.5^1, 0.5^2, 0.5^3) seems to finish the problem, I could not figure out how to do this in general case.

javascript case:
How to convert a binary fraction number into decimal fraction number? How to convert binary fraction to decimal lisp case:
Convert fractions from decimal to binary

Solution

You can extend the solution you posted in your question to also include the negative powers of two starting at the position of the decimal separator as follows:

base2decimal <- function(base_number, base = 2) {
    base_number = paste(as.character(base_number), ".", sep = "")
    return (mapply(function (val, sep) {
                      val = val[-sep];
                      if (val[[1]] == "-") {
                          sign = -1
                          powmax = sep[[1]] - 3
                          val = val[-1]
                      } else {
                          sign = 1
                          powmax = sep[[1]] - 2
                      };
                      sign * sum(as.numeric(val) * (base ^ seq(powmax, by = -1, length = length(val))))},
        strsplit(base_number, NULL), gregexpr("\\.", base_number)))
}

This code also works for other bases less than (or equal) 10:

base2decimal(c('0.101', '.101', 0.101, 1101.001, 1101, '-0.101', '-.101', -0.101, -1101.001, -1101))
#[1]   0.625   0.625   0.625  13.125  13.000  -0.625  -0.625  -0.625 -13.125
#[10] -13.000
base2decimal(1110.111)
# 14.875
base2decimal(256.3, 8)
# [1] 174.375

Answered By - martin_joerg

Answer Checked By - Cary Denson (PHPFixing Admin)