Univariate Gaussian MLX produces error upon readSample

[jonasallgeier]

I am running sampling_univariate_gaussian_distribution_via_paradram.mlx on Matlab R2021b and get an error message on line 51 (sample = pmpd.readSample();):

Error using ParaDRAM_class/readTabular
Too many output arguments.
Error in ParaDRAM_class/readSample (line 176)
        sampleList = self.readTabular(callerName,file,delimiter); 

Is that a known problem? How can it be fixed?

[shahmoradi]

Hi @jonasallgeier, what version of MATLAB are you using?

[shahmoradi]

I just tested the latest release of the library with MATLAB 2021a and the software and example work as intended. A few possible remedies:

1. check if your MATLAB version is supported by ParaMonte. That means any version of MATLAB that is newer than 2018a. MATLAB 2018a should also work with ParaMonte, albeit on Windows you will have to install the patch that MATLAB released after MATLAB 2018a release. Without the MATLAB patch, ParaMonte won't work on MATLAB 2018a on Windows (Linux and macOS are fine).
2. Make sure you are using the latest ParaMonte version for MATLAB: https://www.cdslab.org/paramonte/notes/installation/matlab/#the-fast-way
3. Let me know if the problem persists.

[jonasallgeier]

Thank you very much for the quick reply. I use R2021b on Ubuntu 20.04. Apparently the ParaMonte version was the issue: Previously I used ParaDram (obtained from the first link in the description of https://www.mathworks.com/matlabcentral/fileexchange/80866-matdram-delayed-rejection-adaptive-metropolis-mcmc). With the ParaMonte version of your link it works.