Site icon R-bloggers

Convert decimal to IEEE-754 in R

[This article was first published on Statistic on aiR, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
For some theory on the standard IEEE-754, you can read the Wikipedia page. Here I will post only the code of the function to make the conversion in R.


First we write some functions to convert decimal numbers to binary numbers:

decInt_to_8bit <- function(x, precs) { q <- c() r <- c() xx <- c() for(i in 1:precs){ xx[1] <- x q[i] <- xx[i] %/% 2 r[i] <- xx[i] %% 2 xx[i+1] <- q[i] } rr <- rev(r) return(rr) } devDec_to_8bit <- function(x, precs) { nas <- c() nbs <- c() xxs <- c() for(i in 1:precs) { xxs[1] <- x*2 nas[i] <- (xxs[i]) - floor(xxs[i]) nbs[i] <- trunc(xxs[i], 1) xxs[i+1] <- nas[i]*2 } return(nbs) }

For example, in 8-bit:

decInt_to_8bit(11, 8) [1] 0 0 0 0 1 0 1 1


devDec_to_8bit(0.625, 8) [1] 1 0 1 0 0 0 0 0


devDec_to_8bit(0.3, 8) [1] 0 1 0 0 1 1 0 0 devDec_to_8bit(0.3, 16) [1] 0 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0

We can delete the extra-zeros from the vectors, using these functions:

remove.zero.aft <- function(a) { n <- length(a) for(i in n:1){ if (a[n]==0) a <- a[-n] else return(a) n <- n-1 } } remove.zero.bef <- function(a) { n <- length(a) for(i in 1:n){ if (a[1]==0) a <- a[-1] else return(a) } }

So we have:

remove.zero.bef(decInt_to_8bit(11, 8)) [1] 1 0 1 1 remove.zero.aft(devDec_to_8bit(0.625, 8)) [1] 1 0 1

Binding these functions, we have:

dec.to.nbit <- function(x,n) { aa <- abs(trunc(x, 1)) bb <- abs(x) - abs(trunc(x)) q <- c() r <- c() xx <- c() for(i in 1:n){ xx[1] <- aa q[i] <- xx[i] %/% 2 r[i] <- xx[i] %% 2 xx[i+1] <- q[i] } rr <- rev(r) nas <- c() nbs <- c() xxs <- c() for(i in 1:n) { xxs[1] <- bb*2 nas[i] <- (xxs[i]) - floor(xxs[i]) nbs[i] <- trunc(xxs[i], 1) xxs[i+1] <- nas[i]*2 } bef <- paste(remove.zero.bef(rr), collapse="") aft <- paste(remove.zero.aft(nbs), collapse="") bef.aft <- c(bef, aft) strings <- paste(bef.aft, collapse=".") return(strings) }

Example:

dec.to.nbit(11.625,8) [1] "1011.101"




Now we can write the code for the decimal to IEEE-754 single float conversion in R:

dec.to.ieee754 <- function(x) { aa <- abs(trunc(x, 1)) bb <- abs(x) - abs(trunc(x)) rr <- decInt_to_8bit(aa, 32) ppc <- 24 - length(remove.zero.bef(rr)) nbs <- devDec_to_8bit(bb, ppc) bef <- remove.zero.bef(rr) aft <- remove.zero.aft(nbs) exp <- length(bef) - 1 mantissa <- c(bef[-1], aft) exp.bin <- decInt_to_8bit(exp + 127, 16) exp.bin <- remove.zero.bef(exp.bin) first <- c() if (sign(x)==1) first=c(0) if (sign(x)==-1) first=c(1) ieee754 <- c(first, exp.bin, mantissa, rep(0, 23-length(mantissa))) ieee754 <- paste(ieee754, collapse="") return(ieee754) }

The numbers 11.625 and 11.33 in IEEE-754 are:

dec.to.ieee754(11.625) [1] "01000001001110100000000000000000" dec.to.ieee754(11.33) [1] "01000001001101010100011110101110"

You can verify the output with this Online Binary-Decimal Converter

To leave a comment for the author, please follow the link and comment on their blog: Statistic on aiR.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.