Why is Matlab 2019B not using python 3.6?

I am attempting to use a Matlab wrapper for a Python library, one that requires >= Python 3.5. When I use the `pyenv` command for Python 2, it shows something like this, with a library path

Version: "2.7"

Executable: "/usr/bin/python2.7"

Library: "library path"

Home: "/usr"

...

But when I use the same command after "loading" Python 3.6, the library path is empty. As a result, python is not actually loaded, so any py command (e.g. `py.list`) returns an error in the form `unable to resolve the name py.commandname`. If I check my terminal shell, I know that I have this version installed however. Also it is in --enable -shared upon calling RPM.

I feel like I've reached a dead end at this point, and due to the configuration of the computer I am working on, a solution without using anaconda would be preferable.

Answer 1

[the cyclist]

The following worked for me, per instructions in the documentation for pyenv.

First, I ran

pyenv

with no argument, which told me 2.7 was loaded. Then I ran

py.list

just to verify that it worked.

Then, I restarted MATLAB, which according to the first note on that page is required to change the interpreter.

Then I ran

pyenv('Version','/usr/local/bin/python3')

using the full path to the executable on my machine. (Your path may differ. I found it by typing "which python3" in a Mac terminal.)

I hope that helps.

[Brian Barry]

This is what I have been attempting to begin with, however the library path is still empty. Weird.

[the cyclist]

That is weird. You could try contacting Mathworks support and see if they can help out.