[R-sig-ME] Error in qr.default(X) : too large a matrix for LINPACK

[Ben Bolker]

   We'd love to help, but we can't help much without more information. 
How large is your data set, what analysis are you running?

   Based on the minimal info in the error message, it looks like the 
model is failing in the place where it checks to see whether the model 
matrix is collinear (i.e., here:

https://github.com/glmmTMB/glmmTMB/blob/5939686f39b169b7d5a81512b6a66025f103d598/glmmTMB/R/glmmTMB.R#L1432

) Digging a little bit farther, it seems that LINPACK can't do QR 
decomposition if the total size of your matrix is bigger than 2^31-1:

  if (1 * n * p > 2147483647)
         stop("too large a matrix for LINPACK")

   You should be able to skip this checking step by specifying

control = glmmTMBControl(rank_check = "skip")

  ... but you may run into trouble further down the line if your model 
matrix has more than a billion entries ...

   If you can tell us a little bit more about your problem people may be 
able to suggest strategies.

   Ben Bolker


On 2023-11-21 3:11 p.m., Mohamed Jabir wrote:
> Hi,
> 
> I am trying to use glmmTMB. I am getting this error :
> Error in qr.default(X) : too large a matrix for LINPACK
> Calls: glmmTMB ... .adjustX -> <Anonymous> -> <Anonymous> -> qr.default
> Execution halted
> Can you help?
> 
> Best,
> 
> Mohamed Jabir
> Analyste de donn�es LACED
> Biblioth�que HEC Montr�al (LACED)
> 3000, chemin de la C�te-Sainte-Catherine
> Montr�al (Qu�bec)  H3T 2A7
> 514-340-6934<https://inscription.hec.ca/signature/app/voir?type=minimal>
> HEC Montr�al
> hec.ca<http://www.hec.ca/>
> 
> 
> 	[[alternative HTML version deleted]]
> 
> 
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models