CODEWITHSUNDEEP
r
Travis Gerke
Travis Gerke

Reputation: 344

avoid hyphen as minus in as.formula

I often need to convert large sets of variables from data frames into formulaic expressions. These variables sometimes have dashes/hyphens in them, and this gives as.formula trouble.

I am aware that I could just gsub these out of the data frame names, but that's undesirable as I need to map back to other annotation files which use the hyphens.

I have one workaround that leverages backticks (`) and paste0 below, but it feels clunky (i.e. is hard to read and may be less efficient than something in base that I've missed).

Is there a simpler expression that will get this done?

dat <- data.frame(rnorm(100), rnorm(100))
names(dat) <- c("y", "x-1")

form <- as.formula(paste("y~", names(dat)[2]))
form #treats hyphen as minus
lm(form, data = dat) #error

# is this the only workaround?
form <- as.formula(paste("y~", paste0("`", names(dat)[2], "`")))
form
lm(form, data = dat)

Upvotes: 2

Views: 818

Answers (1)

Roland
Roland

Reputation: 132736

I'd use .:

f <- y ~ .
lm(f, dat)

Subset the data.frame to only contain variables of interest.

Upvotes: 3

Related Questions