I want to relevel the factors in a dataset, however I'm really struggling with the fct_relevel syntax and using it with mutate_at. I get a series of errors about my data not being a factor.
The solution must allow me to relevel multiple factors (the actual dataset has 20-odd factors to relevel in different ways)
This answer seems like it should work, but I'm clearly not picking up the syntax properly. Where am I going wrong?
Here's an example:
library(tidyverse)
dat <- tibble (x1 = c("b", "b", "a", "c", "b"),
x2 = c("c", "b", "c", "a", "a"),
y = c(10, 5, 12, 3, 4)) %>%
mutate_at(.vars = vars(x1:x2), factor)
I'm definitely dealing with factors
sapply(dat, class)
But I can't relevel x1, I receive the following error: f must be a factor (or character vector))
dat %>% fct_relevel(x1, "c", "b", "a")
And this is what I ideally want to be able to do
dat2 <- dat %>%
mutate_at(.vars = vars (x1:x2),
.funs = fct_relevel("c", "b", "a"))
At the moment that final set is giving me the following errors:
Error: Can't create call to non-callable object
Call rlang::last_error() to see a backtrace
In addition: Warning message:
Unknown levels in f: b, a
I'd be really grateful for anyone pointing out what I'm sure is an obvious mistake.
This should work
library(dplyr)
library(forcats)
dat <- dat %>% mutate_at(vars(x1:x2), ~fct_relevel(., c("c", "b", "a")))
dat$x1
#[1] b b a c b
#Levels: c b a
dat$x2
#[1] c b c a a
#Levels: c b a