The inla result collection failed

[Virginia Morera Pujol]

Hi,

I've been teaching the use of inlabru to a group of students using the gorillas example following this code, and in about 4 computers out of the ~15 involved they encountered an error when running a very basic model. This happened all in windows machines, newest R version 4.5 and several 4.4, INLA newest stable and also testing version, and presence and absence of updated RTools, so I can't really pin point the settings or session in which the error happens. I'm sorry this is not a very good reproducible example, but this did not happen in my machine. However, I ran the code with options = list(bru_verbose = 4) in one of them and got this output. It says absolutely nothing to me but maybe it makes more sense to someone here? I have seen in another thread here this "result collection failed" can be due to an antivirus blocking inla from saving or recovering files? Might that be the issue here? Or are other issues that might cause this?

> m1 <- bru(components = form1, +          data = gorillas_sf$plotsample$nests, +          family = "cp", +          domain =  list(geometry = mesh_better), +          options = list(bru_max_iter = 20, +                         bru_verbose = 4)) Aviso en input_eval.bru_input(subcomp$input, data = lh$data, env = env,  :  Model input 'covars$elev' for 'Eff.elevation' returned some NA values. Attempting to fill in spatially by nearest available value. To avoid this basic covariate imputation, supply complete data.

iinla: Evaluate component inputs

Aviso en input_eval.bru_input(component[[part]]$input, data, env = component$env,  :  Model input 'covars$elev' for 'Eff.elevation' returned some NA values. Attempting to fill in spatially by nearest available value. To avoid this basic covariate imputation, supply complete data.

iinla: Evaluate component linearisations Linearise component 'Intercept' Linearise component 'Eff.elevation' iinla: Evaluate component simplifications Simplify component 'Intercept' Simplify component 'Eff.elevation' iinla: Evaluate predictor linearisation Linearise with respect to component 'Intercept' Linearise with respect to component 'Eff.elevation' iinla: Construct inla stack iinla: Model initialisation completed iinla: Iteration 1 [max:1]

Aviso en iinla(model = info[["model"]], lhoods = info[["lhoods"]], options = info[["options"]]) :  iinla: Problem in inla: Error in inla.core.safe(formula = formula, family = family, contrasts = contrasts,  :  The inla result collection failed. The inla result collection failed. iinla: Problem in inla: 1: bru(components = form1, data = gorillas_sf$plotsample$nests, family = "cp" [...] 2: iinla(model = info[["model"]], lhoods = info[["lhoods"]], options = info[[ [...] 3: fm_try_callstack(...) 4: do.call(INLA::inla, inla.options.merged, envir = environment(model$effects)) 5: (function (formula = NULL, family = "gaussian", contrasts = NULL, data = N [...] 6: inla.core.safe(formula = formula, family = family, contrasts = contrasts,  [...] 7: stop(paste0(r$message, "\n", "The inla result collection failed."))

iinla: Giving up and returning last successfully obtained result for diagnostic purposes.

[Finn Lindgren]

Hi,

Unlikely to an inla specific issue, but need more details on what package versions etc are involved; need full sessionInfo() at the very least. Likely an issue with sf and/or terra. And the inlabru and fmesher versions. Please post a report on the inlabru github instead of here.

Finn

On 14 May 2025, at 21:17, Virginia Morera Pujol <morera....@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "R-inla discussion group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to r-inla-discussion...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/r-inla-discussion-group/ff5f184b-092a-4b8b-9e7f-3987c2d85891n%40googlegroups.com.

[Finn Lindgren]

You’re missing the samplers argument.

That, in addition to using a different mesh than that stored with the data likely means You’re trying to construct a point process likelihood outside the available covariate values; that’s what the initial NA messages are indicating.

Not sure why that would lead to an inla result collection error though…

Finn

On 14 May 2025, at 21:28, Finn Lindgren <finn.l...@gmail.com> wrote:

Hi,

[Ruiling Zhang]

Hi,

I am replicating the results of this paper: https://github.com/DenisRustand/Bayes-surv-INLA/blob/main/Bayes-surv-INLA.R

I can library all the packages properly. But when I # Fit the model (line 23), it went to the error :

 inla.core.safe(formula = formula, family = family, contrasts = contrasts, :

  The inla result collection failed.
The inla result collection failed.

What is the problem?

Thank you!

[Denis Rustand]

Hi,

This could deserve a separate thread as it's likely a different issue. The code runs fine for me, can you try with the latest testing version of INLA (www.r-inla.org/download-install) and the latest GitHub version of INLAjoint (github.com/DenisRustand/INLAjoint)?

Best,

Denis Rustand