Problem with read_stan_csv


#1

It might be something silly but I can’t make stan_csv_reader work properly. The example using :

csvfiles <- dir(system.file(‘misc’, package = ‘rstan’), pattern = ‘rstan_doc_ex_[0-9].csv’, full.names = TRUE)
fit <- read_stan_csv(csvfiles)

works but when applied to my Cmdstan 2.18.0 output I get :

Error in [<-(*tmp*, buffer.pointer, , value = scan(con, nlines = 1, :
subscript out of bounds

I don’t see how it could be related to my model. I’m on a Linux computer cluster

Sys.info()
sysname release
“Linux” “3.10.0-862.14.4.el7.x86_64”

and I call the sampler using

./modelA1S sample num_warmup=1000 num_samples=1000
adapt delta=0.9 algorithm=hmc engine=nuts max_depth=10 init=1
data file=data_modelA1S.R output file=S_modelA1S_${SLURM_ARRAY_TASK_ID}.csv refresh=10

Any clue?


#2

Did the model finish running?


#3

Yes it did :)


#4

Are you using rstan 2.18.2 or 2.18.1? An issue like that with 2.18.1 was fixed for 2.18.2.


#5

Oh great catch, it’s now working with rstan 2.18.2. Thanks Ben!