R: Reference column name with paste

user_14252096 0
Viewed 169

I am attempting to reference a column in my data frame with paste given a variable name and the column is not being recognized. Below is an example data frame and variable names that I am attempting to work with:

arm <- "ARM"

ex_df <- data.frame(col1 = c("A","B","C"), col2 = c("X","Y","Z"), ARM_1 = c(1,2,3), ARM_2 = c(4,5,6,), ARM_3 = c(7,8,9), ARM_4 = c(0,0,0))

ex_df %>%
    transmute(col1 = col1, col2 = str_to_title(col2), paste0(arm,"_",4) = paste0(arm,"_",1) + paste0(arm,"_",2) + paste0(arm,"_",3)) %>%
    arrange(col1,col2)

Error: unexpected '=' in:
    "col2 = str_to_title(col2),
    paste0(arm,"_",4) = "

The desired output should read ARM_4 = ARM_1 + ARM_2 + ARM_3 and output the sum of each columns values.

Thank you for your help.

dplyr paste dataframe r
1 Answers

We can use := with !! - along with rowSums on the columns that starts_with the value stored in arm

library(dplyr)
library(stringr)
ex_df %>%
    transmute(col1 = col1, col2 = str_to_title(col2), 
    !!paste0(arm,"_",4) := rowSums(across(starts_with(arm)), na.rm = TRUE))

-output

 col1 col2 ARM_4
1    A    X    12
2    B    Y    15
3    C    Z    18

NOTE: rowSums would be better compared to + - 1) take care of NA elements with na.rm = TRUE, 2) compact option when there are more than 10 'ARM' columns

In case we want to get the value of the column with paste

ex_df %>%
    transmute(col1 = col1, col2 = str_to_title(col2), 
   !!paste0(arm,"_",4) := .data[[paste0(arm,"_",1)]] + 
                         .data[[paste0(arm,"_",2)]] + 
                        .data[[paste0(arm,"_",3)]]) %>%
    arrange(col1,col2)

-output

 col1 col2 ARM_4
1    A    X    12
2    B    Y    15
3    C    Z    18
user_37322710
Related