Assigning the result of an across(starts_with(), mean) to the R dataframe

Question

Instead of writing manually an average of

df$average<- (df$a + df$b + df$c)/3

I wanted to use across, but I can't manage to attach it because then creates a 1x1 dataframe I was trying variations of this without success.

df$`mean trust` <- df%>% 
    summarise(across(starts_with("Trust"), mean, na.rm = TRUE))

and also

df$`mean trust` <- df%>%
  summarise(across(starts_with("Trust"), 
                   mean, na.rm=TRUE,
                   .names = "`mean trust`"))

I also prefer to use %<>% to assign, if that's possible at all. Any suggestions?

Answer 1

We can use rowMeans and instead of summarise, it would be mutate

library(dplyr)
df %>% 
     mutate(meantrust = select(., starts_with("Trust")) %>%
                     rowMeans( na.rm = TRUE))

---

With c_across, we can use mean (but it is not vectorized)

df %>%
    rowwise %>% 
    mutate(meantrust = mean(c_across(starts_with("Trust")), na.rm = TRUE))

A reproducible example with iris

data(iris)
head(iris) %>%
    rowwise %>%
     mutate(meanSepal = mean(c_across(starts_with("Sepal")), na.rm = TRUE))