R/round_half_up.R

In janitor: Simple Tools for Examining and Cleaning Dirty Data

#' @title Round a numeric vector; halves will be rounded up, ala Microsoft Excel.
#'
#' @description
#' In base R \code{round()}, halves are rounded to even, e.g., 12.5 and 11.5 are both rounded to 12.  This function rounds 12.5 to 13 (assuming \code{digits = 0}).  Negative halves are rounded away from zero, e.g., -0.5 is rounded to -1.
#'
#' This may skew subsequent statistical analysis of the data, but may be desirable in certain contexts.  This function is implemented exactly from \url{https://stackoverflow.com/a/12688836}; see that question and comments for discussion of this issue.
#'
#' @param x a numeric vector to round.
#' @param digits how many digits should be displayed after the decimal point?
#' @export
#' @examples
#' round_half_up(12.5)
#' round_half_up(1.125, 2)
#' round_half_up(1.125, 1)
#' round_half_up(-0.5, 0) # negatives get rounded away from zero
#'
round_half_up <- function(x, digits = 0) {
  posneg <- sign(x)
  z <- abs(x) * 10 ^ digits
  z <- z + 0.5 + sqrt(.Machine$double.eps)
  z <- trunc(z)
  z <- z / 10 ^ digits
  z * posneg
}

#' @title Round a numeric vector to the specified number of significant digits; halves will be rounded up.
#'
#' @description
#' In base R \code{signif()}, halves are rounded to even, e.g.,
#' \code{signif(11.5, 2)} and \code{signif(12.5, 2)} are both rounded to 12.
#' This function rounds 12.5 to 13 (assuming \code{digits = 2}). Negative halves
#' are rounded away from zero, e.g., \code{signif(-2.5, 1)} is rounded to -3.
#'
#' This may skew subsequent statistical analysis of the data, but may be
#' desirable in certain contexts. This function is implemented from 
#' \url{https://stackoverflow.com/a/1581007/}; see that question and
#' comments for discussion of this issue.
#'
#' @param x a numeric vector to round.
#' @param digits integer indicating the number of significant digits to be used.
#' @export
#' @examples
#' signif_half_up(12.5, 2)
#' signif_half_up(1.125, 3)
#' signif_half_up(-2.5, 1) # negatives get rounded away from zero
#'
signif_half_up <- function(x, digits = 6) {
  
  xs <- which(x != 0 & !is.na(x) & !is.infinite(x))
  
  y <- rep(0, length(x))
  z <- x
  
  y[xs] <- 10 ^ (digits - ceiling(log10(abs(x[xs]))))
  
  z[xs] <- round_half_up(x[xs] * y[xs]) / y[xs]
  
  return(z)
  
}