# [R] Specifying formula inside a function

Erik Iverson eriki at ccbr.umn.edu
Thu Jun 10 06:44:35 CEST 2010

```Hello,

> How does one specify a formula to lm inside a function (with variable
> names not known in advance) and have the formula appear explicitly in
> the output?
>
> For example,
>
> f <- function(d) {
>   in.model <- sample(c(0,1), ncol(d)-1, replace=T)
>   current.model <- lm(paste(names(d), "~",
> paste(names(d[2:ncol(d)])[which(in.model == 1)], collapse= "+")),
> data=d)  #***
>   return(current.model)
> }
> x1 <- rnorm(50,0,1)
> x2 <- rnorm(50,0,1)
> x3 <- rnorm(50,0,1)
> x4 <- rnorm(50,0,1)
> y <- rnorm(50,0,1)
> d <- data.frame(y, x1, x2, x3, x4)
> f(d)
>
> Call:
> lm(formula = paste(names(d), "~",
> paste(names(d[2:ncol(d)])[which(in.model ==     1)], collapse = "+")),
> data = d)
>
> Coefficients:
> (Intercept)           x3           x4
>     -0.1087       0.2830       0.1024
>
> How can I specify the formula in the line marked *** so that the
> output will show "formula = y ~ x3 + x4" instead of "formula =
> paste..."?
>

Well, there could very well be some tricks you can pull with ?substitute, or
others, but I can't seem to figure it out.  Instead of tricks, it might be
easier/clearer to assign a class to your object, say 'test', and write a print
method for that based on print.lm.  There very well may be drawbacks to this
that I am not realizing :).

f <- function(d) {
in.model <- sample(c(0,1), ncol(d)-1, replace=TRUE)
current.model <- lm(paste(names(d), "~",
paste(names(d[2:ncol(d)])[which(in.model == 1)],
collapse= "+")),
data=d)
class(current.model) <- c("test", "lm")
current.model
}

# slight modification to print.lm
print.test <- function (x, digits = max(3, getOption("digits") - 3), ...)
{
cat("\nCall:\n", deparse(x\$terms), "\n\n", sep = "")
if (length(coef(x))) {
cat("Coefficients:\n")
print.default(format(coef(x), digits = digits), print.gap = 2,
quote = FALSE)
}
else cat("No coefficients\n")
cat("\n")
invisible(x)
}

# should give you what you want...
f(d)

```