R: could not find function "function (object, ...) \nobject" in glm() and bestglm()

Question

I was running a bestglm analysis from the package bestglm:

res.bestglm <-
  bestglm(Xy = offermodel,
          family = binomial,
          IC = "BIC",
          method = "exhaustive")

summary(res.bestglm$BestModel) 
Morgan-Tatar search since family is non-gaussian.

and I started getting this error:

Error in model.frame.default(formula = y ~ 1, weights = weights, drop.unused.levels = TRUE) : 
  could not find function "function (object, ...) \nobject"

I inspected my model data and it looks good -- the right classes, no missingness, variables I've used before in this same way with no problem.

After unsucessfully troubleshooting and web searching the error, I tried a regular glm with the data and started getting a nearly identical error:

> glm.mba <- glm(y ~  female + DNC + SE_region + enr_before_offer + FA,
+                family = binomial(link = "probit"), data = offermodel)
Error in model.frame.default(formula = y ~ female + DNC + SE_region +  : 
  could not find function "function (object, ...) \nobject"

I'm not sure how to make a reproducible example for this problem. Here's the structure of my data:

> str(offermodel)
'data.frame':   2559 obs. of  10 variables:
 $ online_chan     : num  1 1 1 1 1 1 1 1 1 1 ...
 $ SE_region       : num  1 0 0 0 1 0 0 0 0 0 ...
 $ recruited       : num  0 0 0 0 0 0 0 0 0 0 ...
 $ referral        : num  0 0 0 0 0 0 0 0 0 0 ...
 $ FA              : num  1 1 0 0 1 1 1 1 1 1 ...
 $ female          : num  0 0 0 0 0 0 0 0 0 0 ...
 $ enr_before_offer: num  0 0 0 0 0 0 0 0 0 0 ...
 $ reg_lag_high    : num  0 0 0 0 0 0 0 0 0 0 ...
 $ DNC             : num  0 0 0 0 1 0 0 0 0 0 ...
 $ y               : num  1 1 1 0 0 1 1 1 1 1 ...

This problem persists regardless of my model specification and if I remove columns from the dataframes, so I do not think that it's a data issue.

> traceback()
5: model.frame.default(formula = y ~ female + DNC + SE_region + 
       enr_before_offer + FA, data = offermodel, drop.unused.levels = TRUE)
4: stats::model.frame(formula = y ~ female + DNC + SE_region + enr_before_offer + 
       FA, data = offermodel, drop.unused.levels = TRUE)
3: eval(expr, envir, enclos)
2: eval(mf, parent.frame())
1: glm(y ~ female + DNC + SE_region + enr_before_offer + FA, family = binomial(link = "probit"), 
       data = offermodel)

At the suggestion of @BenBolker, I tried some debugging options and got a little more info:

> glm.mba <- glm(y ~  female + DNC + SE_region + enr_before_offer + FA, data = offermodel)
Error in model.frame.default(formula = y ~ female + DNC + SE_region +  : 
  could not find function "function (object, ...) \nobject"

Enter a frame number, or 0 to exit   

1: glm(y ~ female + DNC + SE_region + enr_before_offer + FA, data = offermodel)
2: eval(mf, parent.frame())
3: eval(expr, envir, enclos)
4: stats::model.frame(formula = y ~ female + DNC + SE_region + enr_before_offer + FA, data = offermodel, drop.unu
5: model.frame.default(formula = y ~ female + DNC + SE_region + enr_before_offer + FA, data = offermodel, drop.un

Selection: 5
Called from: (function () 
{
    if (.isMethodsDispatchOn()) {
        tState <- tracingState(FALSE)
        on.exit(tracingState(tState))
    }
    calls <- sys.calls()
    from <- 0L
    n <- length(calls)
    if (identical(sys.function(n), recover)) 
        n <- n - 1L
    for (i in rev(seq_len(n))) {
        calli <- calls[[i]]
        fname <- calli[[1L]]
        if (!is.na(match(deparse(fname)[1L], c("methods::.doTrace", 
            ".doTrace")))) {
            from <- i - 1L
            break
        }
    }
    if (from == 0L) 
        for (i in rev(seq_len(n))) {
            calli <- calls[[i]]
            fname <- calli[[1L]]
            if (!is.name(fname) || is.na(match(as.character(fname), 
                c("recover", "stop", "Stop")))) {
                from <- i
                break
            }
        }
    if (from > 0L) {
        if (!interactive()) {
            try(dump.frames())
            cat(gettext("recover called non-interactively; frames dumped, use debugger() to view\n"))
            return(NULL)
        }
        else if (identical(getOption("show.error.messages"), 
            FALSE)) 
            return(NULL)
        calls <- limitedLabels(calls[1L:from])
        repeat {
            which <- menu(calls, title = "\nEnter a frame number, or 0 to exit  ")
            if (which) 
                eval(substitute(browser(skipCalls = skip), list(skip = 7 - 
                  which)), envir = sys.frame(which))
            else break
        }
    }
    else cat(gettext("No suitable frames for recover()\n"))
})()

I also tried:

rm(res.bestglm) # Because of a message you can see in the comments
install.packages("bestglm")
require(bestglm)

try starting from a clean (`--vanilla`) session? you probably have something like an object called `function` that's causing trouble — Ben Bolker, Dec 31 '14 at 17:03
@BondedDust: sure it is. `d <- data.frame(y=c(1,1,0)); glm(y~1,data=d,family=binomial(link="probit"))` (the link function only needs to be monotonic, continuous, and differentiable, I think ...) — Ben Bolker, Dec 31 '14 at 17:05
@BondedDust True -- one is a probit, one isn't. `bestglm` does not support probits or links. The families were `binomial`. — Hack-R, Dec 31 '14 at 17:05
another guess is that you might have overwritten `binomial`? what are the results of `traceback()` ? — Ben Bolker, Dec 31 '14 at 17:07
@BenBolker Trying that now, will take a couple minutes due to big data, thanks — Hack-R, Dec 31 '14 at 17:07
have you checked http://stackoverflow.com/questions/4442518/general-suggestions-for-debugging-r ? — Ben Bolker, Dec 31 '14 at 17:20
I think I just got a big clue as to what the problem is. I started a vanilla session then loaded a `.Rdata` file and got this message: `Warning: namespace ‘bestglm’ is not available and has been replaced by .GlobalEnv when processing object ‘res.bestglm’` — Hack-R, Dec 31 '14 at 17:59
@BenBolker that's a good link and I'm trying to debug with the suggestions but so far still no luck. — Hack-R, Dec 31 '14 at 18:07
`?model.frame.default` tells me that `drop.unused.levels` defaults to `FALSE` but appears `TRUE` in my errors... it shouldn't matter since these are numeric not factor, but maybe it's a clue... — Hack-R, Dec 31 '14 at 18:27

score 3 · Accepted Answer · answered Dec 31 '14 at 18:48

3

Many thanks to the folks in the comments.

I'm not really sure what caused this error, but here's what fixed it:

Removed a problematic object, res.bestglm
Re-installed bestglm
Saved image and closed the project
Re-opened the project and loaded packages

I'm not sure why the object in #1 was problematic, but it was somehow cluttering the Global Environment. Moreover, R could not find bestglm after I launched a vanilla session without removing the object first and re-installing the package.

answered Dec 31 '14 at 18:48

Hack-R

22,422
14
75
131

Restarting the R session works for me as well even though `rm(list=ls())` doesn't. – Heisenberg Mar 09 '15 at 21:52
1

Experienced the same problem. `rm(list=ls())` and subsequent session restart was enough for me to fix the symptoms though. Don't have an idea yet what causes the problem. Something really causes confusion when passing formulas that also exists in the global env. – Matt Bannert May 31 '15 at 08:43
Were you using RStudio? – Carlos Cinelli May 17 '17 at 23:48
@CarlosCinelli yes indeed – Hack-R May 18 '17 at 00:30
2

@Hack-R not sure yet, but I'm suspecting this is an RStudio's bug. – Carlos Cinelli May 19 '17 at 23:06
@CarlosCinelli I agree. I continue to come across this error 4 years later. – Hack-R Oct 15 '18 at 23:52
1

I ran into the same error using just `stats::lm()`. Switching out of RStudio worked for me. – coletl Feb 07 '20 at 21:52

R: could not find function "function (object, ...) \nobject" in glm() and bestglm()

1 Answers1

Linked