I currently need to create 26 new columns based on 26*2 other columns in my dataframe in R. I don’t think using mutate is an option because the function is a little bit complicated. An example dataframe with fewer columns is

test_dat <- tibble(post1=c(1,3,2),pre1=c(0,3,3),post2=c(3,1,4),pre2=c(3,3,3),post3=c(2,4,1),pre3=c(1,3,2))

I need to generate 3 new columns based on the current columns. Right now, the function I have written to try to accomplish what I need is:

normalized_change <- function(df,post,pre,newcol){
  df <- df %>% add_column(newcol = NA)
  for(i in 1:nrow(df)){
    if(post[i]>pre[i]){
      df$newcol[i] <- (post[i]-pre[i])/(4-pre[i])
    }
    if(post[i]==pre[i]){
      df$newcol[i] <- 0
    }
    if(post[i]<pre[i]){
      df$newcol[i] <- (post[i]-pre[i])/pre[i]
    }
  }
  return(df)
}

In order to achieve the desired result, I will have to call the function a total of 3 times, once for each new column I need to generate.

test_dat <- normalized_change(test_dat,test_dat$post1,test_dat$pre1,newcol="norm_change1")
test_dat <- normalized_change(test_dat,test_dat$post2,test_dat$pre2,newcol="norm_change2")
test_dat <- normalized_change(test_dat,test_dat$post3,test_dat$pre3,newcol="norm_change3")

My final test_dat should like like:

tibble(post1=c(1,3,2),
       pre1=c(0,3,3),
       post2=c(3,1,4),
       pre2=c(3,3,3),
       post3=c(2,4,1),
       pre3=c(1,3,2),
       norm_change1=c(0.25,0,-0.33),
       norm_change2=c(0,-0.66,1),
       norm_change3=c(0.33,1,-0.5)
)

However, the issue is that whenever I run the function normalized_change, the new column is given the name newcol instead of “norm_change1.” In my actual data, I will have to make 26 new columns based on 26*2 current columns. If anyone can give advice on how I can make the function actually generate a column with my preferred name, that would be great. However, if anyone has a more efficient way of doing this that wouldn’t require me calling my function 26 times, that would be even better!

I have tried a number of different things. Mutate functions don’t work because I don’t think they can accept such a complicated function. In order to make my current function work, I have tried things like forcing newcol to be a character which didn’t work.