用于替换 R 中值的自定义函数

Question

I'm trying to create a function to replace outliers in a variable:

outlier_upper_replace <- function(x, level=1.5) {
  lowerq = quantile(x, na.rm=TRUE)[2]
  upperq = quantile(x, na.rm=TRUE)[4]
  iqr = upperq - lowerq
  outlier_upper = (iqr * level) + upperq
  x[x > outlier_upper] <- outlier_upper
}

I can't understand why outlier_upper_replace(prod$varname) is not replacing the value.

If I run the following code, which is a particular case of the above function, the value is replaced:

level = 1.5
lowerq = quantile(prod$varname, na.rm=TRUE)[2]
upperq = quantile(prod$varname, na.rm=TRUE)[4]
iqr = upperq - lowerq
outlier_upper = (iqr * level) + upperq
prod$varname[prod$varname > outlier_upper] <- outlier_upper

Any insights about why this is happening?

EDIT: Example

set.seed(1)
x <- rnorm(100)
x <- c(-10, x, 10)

boxplot(x, main="start boxplot")

#function to bin-in upper outliers
outlier_upper_replace <- function(x, level=1.5) {
  lowerq = quantile(x, na.rm=TRUE)[2]
  upperq = quantile(x, na.rm=TRUE)[4]
  iqr = upperq - lowerq
  outlier_upper = (iqr * level) + upperq
  x[x > outlier_upper] <- outlier_upper
  return(x)
}

outlier_upper_replace(x)

boxplot(x, main="end boxplot")

Answer 1

Add

return(x)

to the end of your function.

The return value of an R function is the result of the last expression in the function body. Since you modify the vector in the last expression there is no result and you have to explicitly return the result.