how to use mutate to assign a list of vectors to a list of new var names in r

Question

I'm building a function that you provide a dataframe and a vector of variable names that are numeric - and i want it to return a dataframe with a corresponding variable for each numeric vector that splits it to intervals.

I know how to do it with one variable, but with multiple var names I get "Error: The LHS of := can't be spliced with !!! ".

Here is a reproducible example:

library(tidyverse)

vars <- c("wt", "qsec")
new_vars <- map_chr(vars, ~ paste0(.x, "_bin"))
bins <- map(vars, ~ quantile(mtcars[,.x], na.rm = T))
result <- mtcars %>% 
  mutate(!!!new_vars := map2(vars, bins, ~ cut(mtcars[,.x], breaks = .y)))

What I want to get is mtcars but with 2 extra columns. 1. wt_bin - which will be the quantile interval "wt" matches to 2. qsec_bin - which will be the quantile interval "qsec" matches to

I know it has something to do with !!!new_vars := but I can't seem to understand exactly the problem.

Thanks!

Are you trying to create a single column?. `mutate` creates single variable unless it is an expression — akrun, Apr 16 '19 at 15:15
I'm trying to create multiple columns. In this example i'm trying to create 2 new columns. — guyx1992, Apr 16 '19 at 15:35

score 3 · Accepted Answer · answered Apr 16 '19 at 15:40

3

You could do it without mutate and dodge the nse too:

map2_dfc(vars, bins, ~ cut(mtcars[, .x], .y, labels = F)) %>% 
  set_names(new_vars) %>% 
  bind_cols(mtcars, .)

answered Apr 16 '19 at 15:40

Nate

10,361
3
33
40

akrun · Answer 2 · 2019-04-16T15:56:53.397

1

We can use := with transmute

library(tidyverse)
map2_dfc(vars, bins, ~
                 mtcars %>%
                       transmute(!! paste0(.x, "_bin") := 
                           cut(!! rlang::sym(.x), breaks = .y))) %>% 
                 bind_cols(mtcars,.)

edited Apr 16 '19 at 15:56

answered Apr 16 '19 at 15:47

akrun

874,273
37
540
662

how to use mutate to assign a list of vectors to a list of new var names in r

2 Answers2