Should there be a catch for this particular error that gets produced in step_dummy() when it tries to calculate too big of a matrix? I realise this will eventually get fixed by #420, but is vector memory exhausted (limit reached?) enough of a hint that x has too many levels/too many rows
library(recipes)
dat <- data.frame(x = as.character(1:100000))
recipe(~ ., data = dat) %>%
step_dummy(x) %>%
prep()
#> Error: vector memory exhausted (limit reached?)