13

I was trying to install jupyter package for anaconda in my current environment but constantly getting the following error.

Preparing transaction: done Verifying transaction: failed

CondaVerificationError: The package for ipython located at D:\Anaconda\pkgs\ipython-7.1.1-py36h39e3cac_0 appears to be corrupted. The path 'Lib/site-packages/IPython/lib/tests/test.wav' specified in the package manifest cannot be found.

CondaVerificationError: The package for notebook located at D:\Anaconda\pkgs\notebook-5.7.2-py36_0 appears to be corrupted. The path 'Lib/site-packages/notebook/static/components/MathJax/extensions/a11y/invalid_keypress.mp3' specified in the package manifest cannot be found.

I have Googled many Q&A for this problem, mostly I found:

conda clean --packages --tarballs 
conda clean --all

After trying both I tried the installation again, but none worked...

Any solution for this?

Desmond
  • 307
  • 1
  • 2
  • 8

2 Answers2

6

First try this:

conda remove {failing_packages}
conda install {failing_packages}

Where {failing_packages} is/are the package(s) for which an error is reported.

If that fails, you could try forcing an over-install (this solution is very handy and will likely fix many issues):

conda install -f {failing_packages}

So, for example, for the original poster of this question, it would be;

conda remove ipython notebook
conda install -f ipython notebook

Note that in future releases of Conda, the -f will be removed and you may have to use --clobber or --force-reinstall flags.

I personally like to use the --force-reinstall flag, and worst case scenario I will remove the failing package's directory (i.e. ~/miniconda3/pkgs/{failing_package}...) manually before using the --force-reinstall to re-install it. That always works really well to fix any package issue.

If all that fails, try these steps, one by one, in order, till it works:

conda clean --packages --tarballs
conda clean --all
conda update conda

conda activate base
conda install --revision 0

Please note that each command has implication as to your environment.

Roel Van de Paar
  • 205
  • 2
  • 13
2

It seems that problem was anyhow with conda installation. I tried pip installation for jupyter and it worked fine.

python -m pip install jupyter
Desmond
  • 307
  • 1
  • 2
  • 8