Home / Questions / 106156
Accepted
Alejandro Carrera
Asked: 2020-09-30 14:32:33 +0800 CST 2020-09-30 14:32:33 +0800 CST

How to get the source code of a function in R? (car package)

  • 772

The "car" package in R has an interesting function called outlierTest and I want to analyze its source code to understand how it works. I tried the following and got an error:

getMethod("outlierTest")

    Error in getMethod("outlierTest") : 
     no generic function found for 'outlierTest'

This seems to get me a response but it redirects me to a page with no source code displayed:

methods(outlierTest)
    [1] outlierTest.lm*
    see '?methods' for accessing help and source code

Finally, when typing only "outlierTest" in console I get no response:

outlierTest
    function (model, ...) 
    {
          UseMethod("outlierTest")
    }
<bytecode: 0x000000000cf4d3d0>
<environment: namespace:car>

I appreciate any guidance on this.

r

2 Answers

  1. Best Answer

    You can do the following, at least with this package it works:

    We look for the methods:

    > methods("outlierTest")
[1] outlierTest.lm*
see '?methods' for accessing help and source code

    Particularly the method outlierTest.lmthat has a *according to the documentation:

    The S3 method name is followed by an asterisk * if the method definition is not exported from the package namespace in which the method is defined

    To retrieve in this case your code we can try this:

    getAnywhere('outlierTest.lm')

    If not, the alternative is to directly access the project's github and, on time, you can retrieve this function at this link

    • 1

  2. An alternative is that you can access "unexported" functions (not visible in your namespace) with triple ":"

    Remember that as you can see, the function outlierTestis generic and to access that in particular it applies to objects of the class lmyou must add a point: functiongenerica.clase:

    car:::outlierTest.lm


function (model, cutoff = 0.05, n.max = 10, order = TRUE, labels = names(rstudent), 
    ...) 
{
    rstudent <- rstudent(model)
    if (length(rstudent) != length(labels)) 
        stop("Number of labels does not correspond to number of residuals.")
    else names(rstudent) <- labels
    df <- df.residual(model) - 1
    rstudent <- rstudent[!is.na(rstudent)]
    n <- length(rstudent)
    p <- if (class(model)[1] == "glm") 
        2 * (pnorm(abs(rstudent), lower.tail = FALSE))
    else 2 * (pt(abs(rstudent), df, lower.tail = FALSE))
    bp <- n * p
    ord <- if (order) 
        order(bp)
    else 1:n
    ord <- ord[bp[ord] <= cutoff]
    result <- if (length(ord) == 0) {
        which <- which.max(abs(rstudent))
        list(rstudent = rstudent[which], p = p[which], bonf.p = bp[which], 
            signif = FALSE, cutoff = cutoff)
    }
    else {
        if (length(ord) > n.max) 
            ord <- ord[1:n.max]
        result <- list(rstudent = rstudent[ord], p = p[ord], 
            bonf.p = bp[ord], signif = TRUE, cutoff = cutoff)
    }
    class(result) <- "outlierTest"
    result
}
<bytecode: 0x000000001e2258a0>
<environment: namespace:car>
    • 1