I wanted to know what is the best way to rename multiple columns after using nest() and map().
I have a sample code of what I have done at the moment to achieve this using the iris dataset.
iris_names <- colnames(iris[, 1:4])
iris_sqrt <- iris %>%
nest(-Species) %>%
mutate(square_root = map(data, sqrt)) %>%
unnest(square_root)
names(iris_sqrt)[3:ncol(iris_sqrt)] <- paste0(iris_names, ".sd")
Here, I make a vector of the column names I want to rename before creating iris_sqrt and then renaming using paste0. The drawback of this method is the column name has to appear in the same order as the iris_names vector to rename correctly.
Is there a tidyverse/dplyr way of doing this?
You can use mutate inside the map call. Use across on everything, apply sqrt, using the .names argument you can change the names of the new columns, and use .keep = "unused" in mutate to remove the columns that were used during the calculation:
iris %>%
nest(data = -Species) %>%
mutate(square_root = map(data, ~ .x %>%
mutate(across(everything(), sqrt,
.names = "{.col}.sd"),
.keep = "unused"))) %>%
unnest(square_root)
output
# A tibble: 150 × 6
Species data Sepal.Length.sd Sepal.Width.sd Petal.Length.sd Petal.Width.sd
<fct> <list> <dbl> <dbl> <dbl> <dbl>
1 setosa <tibble [50 × 4]> 2.26 1.87 1.18 0.447
2 setosa <tibble [50 × 4]> 2.21 1.73 1.18 0.447
3 setosa <tibble [50 × 4]> 2.17 1.79 1.14 0.447
4 setosa <tibble [50 × 4]> 2.14 1.76 1.22 0.447
5 setosa <tibble [50 × 4]> 2.24 1.90 1.18 0.447
6 setosa <tibble [50 × 4]> 2.32 1.97 1.30 0.632
7 setosa <tibble [50 × 4]> 2.14 1.84 1.18 0.548
8 setosa <tibble [50 × 4]> 2.24 1.84 1.22 0.447
9 setosa <tibble [50 × 4]> 2.10 1.70 1.18 0.447
10 setosa <tibble [50 × 4]> 2.21 1.76 1.22 0.316
# … with 140 more rows
# ℹ Use `print(n = ...)` to see more rows