Return plot from R function

Question

Why won't my R function return or print a plot? The code is below. All of the code seems to work fine, except for the plot. No matter what I do, I can't get R to create the plot when the function is called. Looking around online, I can't find any reason why this wouldn't work.

powerc.fun <- function(n,sigma,r){  a <- 0.05 d <- seq(-20,20,2)  power <- rep(NA,length(d)) p.lab <- rep(NA,length(d))  for (j in 1:length(d)){    mu1 <- 110   mu2 <- mu1-d[j]      reject <- rep(NA,r)    for (i in 1:r){      sample1 <- rnorm(n,mu1,sigma)     sample2 <- rnorm(n,mu2,sigma)      sample.t <- t.test(sample1,sample2)     p.val <- sample.t[3]      reject[i] <- p.val<a       power[j] <- sum(reject)/length(reject)     p.lab[j] <- paste('d=',d[j],sep='')    }}  d.power <- cbind(d,power)  return(d.power)  p.plot <- plot(d.power[,1], d.power[,2], type="l", xlab=bquote(H[a]), ylab="Power", main="Power Calculations for Two Sample T Test")  print(p.plot) return(p.plot)  } 

Any ideas?

Answer 1

This works. As suggested, you can save more than one object by using a list:

powerc.fun <- function(n,sigma,r){  a <- 0.05 d <- seq(-20,20,2)  power <- rep(NA,length(d)) p.lab <- rep(NA,length(d))  for (j in 1:length(d)){    mu1 <- 110   mu2 <- mu1-d[j]      reject <- rep(NA,r)    for (i in 1:r){      sample1 <- rnorm(n,mu1,sigma)     sample2 <- rnorm(n,mu2,sigma)      sample.t <- t.test(sample1,sample2)     p.val <- sample.t[3]      reject[i] <- p.val<a       power[j] <- sum(reject)/length(reject)     p.lab[j] <- paste('d=',d[j],sep='')    }}  d.power <- cbind(d,power) p.plot <- plot(d.power[,1], d.power[,2], type="l", xlab=bquote(H[a]), ylab="Power", main="Power Calculations for Two Sample T Test")  return(list(p.plot, d.power))  }  # prints the plot and saves d.power values output <- powerc.fun(100,0.1,10)  # d.power values output[[2]] 

But probably you prefer just to save d.power and then call it for plotting the graph:

powerc.fun <- function(n,sigma,r){  a <- 0.05 d <- seq(-20,20,2)  power <- rep(NA,length(d)) p.lab <- rep(NA,length(d))  for (j in 1:length(d)){    mu1 <- 110   mu2 <- mu1-d[j]      reject <- rep(NA,r)    for (i in 1:r){      sample1 <- rnorm(n,mu1,sigma)     sample2 <- rnorm(n,mu2,sigma)      sample.t <- t.test(sample1,sample2)     p.val <- sample.t[3]      reject[i] <- p.val<a       power[j] <- sum(reject)/length(reject)     p.lab[j] <- paste('d=',d[j],sep='')    }}  d.power <- cbind(d,power)  return(d.power)  }  # saves d.power output <- powerc.fun(100,0.1,10)  # plot p.plot <- plot(output[,1], output[,2], type="l", xlab=bquote(H[a]), ylab="Power", main="Power Calculations for Two Sample T Test") 

Answer 2

A call to "return()" ends the function call, so anything after it is ignored

Try getting rid of

return (d.power) 

or moving it to after you define and print the plot.

On a related note, you can't return two objects from one function. Pick one or put them in a list and return the list.