How to dynamically create columns with tidy evaluation?

I'd like to create variables using a specific pattern. I need to include the variable name as a tidy evaluation in the mutate function. I made a custom example below:

  iris %>%
  rename(Sepal = Sepal.Width, Petal = Petal.Width) %>%
  mutate_at(c('Sepal', 'Petal'), list(test = ~ . / !!sym(paste0(., '.Length'))))

Solution

In the newer version of dplyr, we can use across

library(dplyr)
library(stringr)
iris %>%
   rename(Sepal = Sepal.Width, Petal = Petal.Width) %>%
   mutate(across(c('Sepal', 'Petal'), ~ 
              ./get(str_c(cur_column(), '.Length')), .names = '{.col}_test'))

-output

#    Sepal.Length Sepal Petal.Length Petal    Species Sepal_test Petal_test
#1            5.1   3.5          1.4   0.2     setosa  0.6862745 0.14285714
#2            4.9   3.0          1.4   0.2     setosa  0.6122449 0.14285714
#3            4.7   3.2          1.3   0.2     setosa  0.6808511 0.15384615
#4            4.6   3.1          1.5   0.2     setosa  0.6739130 0.13333333
#5            5.0   3.6          1.4   0.2     setosa  0.7200000 0.14285714
#6            5.4   3.9          1.7   0.4     setosa  0.7222222 0.23529412
#7            4.6   3.4          1.4   0.3     setosa  0.7391304 0.21428571
#8            5.0   3.4          1.5   0.2     setosa  0.6800000 0.13333333

# ...

We don't need to rename just for doing the division. It can also be achieved by keeping the original column names

iris %>%
  mutate(across(ends_with('Width'), ~
              ./get(str_replace(cur_column(), 'Width', 'Length')),
         .names = '{.col}_test'))

The . returns the value and not the column name. So, paste0(., '.Length') would be pasting substring .Length with the corresponding column values