Piping with `magrittr`

I have just spent a semester teaching multilevel modeling, and in the R codes I provided, I usually use the pipe operator (%>%). For example, to compute the cluster means, we can do

library(tidyverse)
data("Hsb82", package = "mlmRev")
Hsb82 <- Hsb82 %>% 
  group_by(school) %>% 
  mutate(ses_cm = mean(ses)) %>% 
  ungroup()

However, it’s kind of embarassing that I only recently found out the assignment pipe (%<>%) operator, as discussed here. For example,

library(magrittr)
set.seed(123)
x <- rnorm(10)
mean(x)

## [1] 0.07462564

# Add 1 to x
x %<>% magrittr::add(1)
mean(x)

## [1] 1.074626

# The above is equivalent to 
# x <- x + 1

For the cluster mean example, we can do

Hsb82 %<>% 
  group_by(school) %>% 
  mutate(ses_cm2 = mean(ses)) %>% 
  ungroup()
select(Hsb82, ses_cm, ses_cm2)

## # A tibble: 7,185 x 2
##    ses_cm ses_cm2
##     <dbl>   <dbl>
##  1 -0.434  -0.434
##  2 -0.434  -0.434
##  3 -0.434  -0.434
##  4 -0.434  -0.434
##  5 -0.434  -0.434
##  6 -0.434  -0.434
##  7 -0.434  -0.434
##  8 -0.434  -0.434
##  9 -0.434  -0.434
## 10 -0.434  -0.434
## # … with 7,175 more rows

which saves the additional typing of Hsb82 <- Hsb82 %>%. That said, the %<>% is not commonly seen when reading other people’s code, so perhaps the R community still prefer just using the %>% operator. But it’s at least good to know there is a potentially more convenient way. There is also the %$% and %T>% operator, as discussed in this vignette.