Error in 'getParameterOrder'

[jcpiquet]

Hi everyone,
After downloading the latest version of TMB package, I tried to re-run a code on glmmTMB that was working just fine until this morning, and I got this error:
'Error in .Call("getParameterOrder", data, parameters, new.env(), NULL, :
Incorrect number of arguments (4), expecting 3 for 'getParameterOrder'
If I use the same code for lme4 it works perfectly fine. Any idea on what is happening and what can I do to fix this?
If needed, I would have no problem sharing the code or anything else.
Thanks!
Julien

[mebrooks]

Try re-installing glmmTMB (from source) after TMB. We're working on preventing this in the future.

[bbolker]

I will point out that you should have gotten a warning about "Package version inconsistency detected" when you loaded glmmTMB. Did you?

[jcpiquet]

Yes!! I saw the message right now. Thank you both for the rapid response. After re-installing glmmTMb from source it worked again.

[antigosa]

Hi all,

Sorry to open this again, but I have a related issue, I think. When I try to simply see what's inside a saved model that I have run using glmmTMB, I got this message:

Error in .Call("MakeDoubleFunObject", data, parameters, reportenv, PACKAGE = DLL) : 
Incorrect number of arguments (3), expecting 4 for 'MakeDoubleFunObject'

Try re-installing glmmTMB (from source) after TMB. We're working on preventing this in the future.

I tried this, but did not work unfortunately.

I will point out that you should have gotten a warning about "Package version inconsistency detected" when you loaded glmmTMB. Did you?

And I don't get this

Any idea? Do you need more info?

Thanks!
Raffa

[kaskr]

@antigosa You might be able to run up2date(oldfit) on your old fit...

[antigosa]

@antigosa You might be able to run up2date(oldfit) on your old fit...

It works! Great, thanks @kaskr

[bbolker]

We should make sure this is documented appropriately ...

[Thom-Teears]

Hello,
I get a similar error (shown below) as well when using the TMB package. I tried reinstalling TMB and glmmTMB from source and the error persists. I don't get any messages regarding package version inconsistency. Any help you can offer would be greatly appreciated.

Thanks,
Thom

Error in .Call("getParameterOrder", data, parameters, reportenv, control, :
Incorrect number of arguments (4), expecting 3 for 'getParameterOrder'

[bbolker]

OK, the usual questions:

• are you loading a stored fit from an older session (i.e., fitted with an older version of glmmTMB)? (In which case, run up2date() on your model.)
• have you installed TMB and glmmTMB from source in that order?

I don't have many ideas beyond that, but it would be helpful if you can give the results of sessionInfo() from a clean R session (i.e., only TMB and glmmTMB loaded)

[Thom-Teears]

Thank you so much for your response. I appreciate it.

I uninstalled R, RStudio, and Rtools and then reinstalled them as per the instructions at http://github.com/kaskr/adcomp/wiki/Download.

Then I installed TMB and glmmTMB from source in that order. The error still persists from a clean R session (i.e., only TMB and glmmTMB loaded). The results of sessionInfo() from a clean R session are below.

Thanks,
Thom

sessionInfo()
R version 4.2.0 (2022-04-22 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19044)

Matrix products: default

locale:
[1] LC_COLLATE=English_Australia.utf8 LC_CTYPE=English_Australia.utf8 LC_MONETARY=English_Australia.utf8
[4] LC_NUMERIC=C LC_TIME=English_Australia.utf8

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] glmmTMB_1.1.7 TMB_1.9.6

loaded via a namespace (and not attached):
[1] minqa_1.2.6 MASS_7.3-56 compiler_4.2.0 Matrix_1.4-1 tools_4.2.0
[6] rstudioapi_0.15.0 Rcpp_1.0.11 splines_4.2.0 nlme_3.1-157 grid_4.2.0
[11] nloptr_2.0.3 boot_1.3-28 numDeriv_2016.8-1.1 lme4_1.1-34 lattice_0.20-45

[Thom-Teears]

I should also mention that I am not trying to load a stored fit from an older session.

Thanks,
Thom

[bbolker]

So far I haven't been able to reproduce this. Do you get this with any glmmTMB fit (e.g.

library(glmmTMB)
data("sleepstudy", package = "lme4")
glmmTMB(Reaction ~ Days + (Days|Subject), sleepstudy)

) or is it only with a specific model? (Note that you might as well update now to glmmTMB 1.1.8, which just went up on CRAN, although I doubt it will make a difference ...)

If it's with a specific model, can you create a reproducible example for us?

[Thom-Teears]

Running the code you provided above worked without any problems. The issue I was having was related to the specific model demo that I was trying to run. I was able to get the model demo to run without any errors by replacing the .cpp file with another that was correctly specified. Thanks again for your assistance.

[bbolker]

OK, closing this, but if you can provide us a reproducible example that's causing a problem we can try to help diagnose (or, send it to r-sig-mixed-models@r-project.org)

[Thom-Teears]

I was able to correct the error by replacing the .cpp file and the model ran without error so there is no longer a need for diagnosis. It seems that maybe the .cpp file I was using had an error. Thanks again for your assistance. I appreciate it.

[amynang]

Sorry to resurrect this issue but I have similar problems after updating to R 4.4.1.
I have installed Rtools44.
I have reinstalled the Matrix package and then installed TMB and glmmTMB from source as suggested here.
This made the "Package version inconsistency" warning go away but what has not changed is the following behavior:

library(glmmTMB)
set.seed(42)
d = data.frame(s = c(rep(letters, times = sample(5:50, 26, replace = T))),
               x = NA,
               y = NA)
d$x = runif(nrow(d), -5,5)
for (i in letters) {
 d$y[d$s == i] = runif(1, 8,12) + (runif(1, -2,4) + rnorm(nrow(d[d$s == i,]), 0, 5))*d$x[d$s == i]
}

mcp = glmmTMB(y ~ x, data = d) # no problems
summary(mcp)
mnp = glmmTMB(y ~ 0 + s + x:s, data = d) # throws error:

Error in getParameterOrder(data, parameters, new.env(), DLL = DLL) :
Error when reading the variable: 'Z'. Please check data and parameters.
In addition: Warning message:
In getParameterOrder(data, parameters, new.env(), DLL = DLL) :
Expected sparse matrix of class 'dgTMatrix'.

mpp = glmmTMB(y ~ x + (x|s), data = d) # throws warning:

Warning message:
In augReTrms$theta[barpos] <- reTrms$theta :
number of items to replace is not a multiple of replacement length

ranef(mpp) # error

Error in if (ub2 == length(levs[[i]])) levs[[i]] else seq(ub2) :
argument is of length zero

This example is pre-existing code that was running without issues.
I there anything else I can try?

(perhaps irrelevant: I also had to troubleshoot to make brms, cmdstanr and cmdstan to work together. In that case it was a matter of getting the cmdstan toolchain to use rtools44 instead of rtools42)