user defined R function that use column names as an argument

April 11, 2023

set.seed(2023)
pid <- 1:30
y1 <- rnorm(30, 10, 10)
y2 <- rnorm(30, 10, 10)
x1 <- rnorm(30, 5, 2)
x2 <- rnorm(30, 5, 2)
x3 <- rnorm(30, 300, 3)

This is minimal reproducible data.
I want to make a function as below.

try <- function(data, a, b) {
  
  data <- data %>%
    mutate(paste0(a, b, "manual_mean") = ({{a}} + {{b}}) /2
           ) 

return(data)
}

so, a and b will be column names from data.


t<-try(data2,y1,y2)

and I tried this try() function.
I expect that I have a new column named as "y1y2manualmean"
but it throws an error.

>Solution :

Use the := operator – try is already a function name. So, use a different name as well

f1<- function(data, a, b) {
  
   data %>%
    mutate("{{a}}_{{b}}_manual_mean" := ({{a}} + {{b}}) /2
           ) 
}

-testing

> f1(data2, y1, y2)
    pid         y1          y2       x1         x2       x3 y1_y2_manual_mean
1    1  9.1621564   7.2512294 6.976368  2.2217230 298.5748          8.206693
2    2  0.1705625  22.7665407 4.431874  5.8573400 301.1042         11.468552
3    3 -8.7506732   1.8901966 1.155465  4.4124372 300.3106         -3.430238
4    4  8.1385534   9.5507722 2.652560  9.5599976 301.0390          8.844663
...