Anaconda3 (8.6) is installed at /Users/myusername/anaconda3. Over the years I have installed non-standard packages using PyPi and conda-forge; these are visible using ‘conda list’. This is (was) my base(root) environment. I want to keep this configuration but not as the base(root), so I saved this as a .yaml file in an offline directory intending to import once the newer version was up and running. My objective is to have the current version of anaconda as base(root) and have the older version, with its modified packages, as a second environment. I attempted to update to the latest version using “conda” commands, but the process was still running after two weeks so I aborted. I restored the old configuration using a TimeMachine backup. The documentation says anaconda’s default installation directory is at /opt/anaconda3, but then I would have two directories of anaconda. Seeking to avoid this, I attempted to install, for myself, at /users/myusername, believing it would overwrite the existing directory. It didn’t; it created an /opt/anaconda3 in myusername directory. So I inadvertently have two directories. The old uses 32 GB the new uses 9 GB. This is bad.
Launching the new anaconda-navigator, I see that the new setup is reading from both directories, the base(root) is the current version and there’s an ‘anaconda3’ environment which resides in the old directory.
Question: How do I clean this up? Just one directory. What’s the proper way to remove the older directory without adversely affecting the new one? Do I delete the navigator’s secondary environment, thus removing the pointers to the older directory, and then, if the older directory still exists, use. ‘rm -rf’ at the old directory to remove those files? Do I then import from the .yaml to generate an environment at ~/opt/anaconda/envs/ ? How do I avoid straggling .xxx files pointing to nowhere and similar configuration misfits?