CODEWITHSUNDEEP
rfunction
LuckyLuke
LuckyLuke

Reputation: 21

R implementation of function

I having some trouble with defining a function in R. Maybe someone with more experience could quickly help me:

sample data frame:

SALES <- c(21341,1241,5234)
EARNINGS <- c(12562,12356,12352)
df <- data.frame (SALES, EARNINGS)

I am interested in the deviation of a variable from its mean. This deviation (d) is calulated as follows.

p <- 0.1
m <- mean(df$SALES)
s <- sd(df$SALES)
d <- qnorm(1-p,mean=m,sd=s)-m

> d
[1] 13637.03

Now, I tried to execute this calulcation in a function with the following inputs: dataframe, variable (column) and p. But instead of 13637.03 I obtain NA as a result:

calculate.d <- function(x,y,p) {
  m <- mean(x$y)
  s <- sd(x$y)
  d <- qnorm(1-p,mean=m,sd=s)-m
  return(d)}
d <- calculate.d(df,SALES,0.1)

> d
[1] NA

why do the two formulations not give the same result? How do i have to adjust the function to get the desired result?

Upvotes: 0

Views: 68

Answers (1)

Anis R.
Anis R.

Reputation: 6902

If your column name is stored in a variable (in your case a parameter), x$y will not work as is (at least not to my knowledge). What you can do is use x[, y] to retrieve the right column. Note that you should use a string (so "SALES", not SALES):

calculate.d <- function(x, y, p) {
  column <- x[, y]
  m <- mean(column)
  s <- sd(column)
  d <- qnorm(1-p,mean=m,sd=s)-m
  d
}

#notice that "SALES" is a string here
calculate.d(df, "SALES", 0.1)

Upvotes: 1

Related Questions