[R-sig-ME] glmer: Downdated X'X is not positive definite [related to nAGQ] fixed?

[Ross Boylan]

On 2/27/2013 5:37 PM, Ross Boylan wrote:
> On 2/27/2013 3:31 PM, Ross Boylan wrote:
>
> I discovered that omitting the nAGQ argument is sufficient to avoid 
> the error.  I also tried nAGQ=5 and removing the missing data in 
> advance; neither eliminated the downdating error.
>
>> r2 <- glmer(sexActs~(1|id), data=sexpartner, family=poisson())
>
>> r2 <- glmer(sexActs~(1|id), data=sexpartner, family=poisson(), nAGQ=5)
>
> Error in mer_finalize(ans) : Downdated X'X is not positive definite, 1.
>
> The error message sounds as if it is just a function of the input 
> covariates, which obviously don't change when I change nAGQ.  My 
> interpretation of X in the error message may be wrong.
>
> This is with the version of lme4 from CRAN.
>
I have some notes that was with lme4 0.999999-0.

I do not get the error when I use MS Windows version 0.999902344-0 (more 
recent, but a lower version number?):

   There is a binary version available (and will be installed) but the
   source version is later:
             binary        source
lme4 0.999902344-0 0.999902345-0

trying URL 'http://cran.cnr.berkeley.edu/bin/windows/contrib/2.15/minqa_1.2.1.zip'
Content type 'application/zip' length 606807 bytes (592 Kb)
opened URL
downloaded 592 Kb

trying URL 'http://lme4.r-forge.r-project.org/repos/bin/windows/contrib/2.15/lme4_0.999902344-0.zip'
Content type 'application/zip' length 2027695 bytes (1.9 Mb)
opened URL

The problem may have had more to do with minqa.  Even after updating 
lme4 only, I got

> library(lme4)
Loading required package: lattice
Loading required package: Matrix
r2 <- glmer(sexActs~(1|id), data=sexpartner, family=poisson(), nAGQ=5)
Error in inDL(x, as.logical(local), as.logical(now), ...) :
   unable to load shared object 'c:/Users/rdboylan/Documents/R/R-2.15.2/site-library/minqa/libs/i386/minqa.dll':
   LoadLibrary failure:  Invalid access to memory location.

Error: package/namespace load failed for 'lme4'
> Error: could not find function "glmer"

It was only after I removed both packages (lme4 and minqa) and 
reinstalled that I was able to run the analysis.

The March 2 announcement said:
> I have just pushed updated source, Windows 64-bit binary, and MacOS
> binaries of the very most current development version of lme4,
> 0.99999911-1, to the repository at
>
>   http://lme4.r-forge.r-project.org/repos/src/contrib/  ;
apparently that needs to be read literally, since I did not pick up a 32 
bit windows binary for that version.