Problem updating bokeh: [WinError 126] The specified module could not be found

1
votes

I tried to update bokeh to 1.4.0 with the command conda update bokeh but encountered the problem of [WinError 126] The specified module could not be found as shown below

(base) D:\Users\jng>conda update bokeh
Collecting package metadata (repodata.json): done
Solving environment: \
Warning: >10 possible package resolutions (only showing differing packages):
  - anaconda/noarch::keras-applications-1.0.8-py_0, anaconda/noarch::keras-preprocessing-1.1.0-py_1, defaults/win-64::h5py-2.8.0-py36hf7173ca_2, defaults/win-64::keras-base-2.2.4-py36_0
  - anaconda/noarch::keras-preprocessing-1.1.0-py_1, defaults/noarch::keras-applications-1.0.8-py_0, defaults/win-64::h5py-2.8.0-py36hf7173ca_2, defaults/win-64::keras-base-2.2.4-py36_0
  - anaconda/noarch::keras-applications-1.0.8-py_0, defaults/noarch::keras-preprocessing-1.1.0-py_1, defaults/win-64::h5py-2.8.0-py36hf7173ca_2, defaults/win-64::keras-base-2.2.4-py36_0
  - defaults/noarch::keras-applications-1.0.8-py_0, defaults/noarch::keras-preprocessing-1.1.0-py_1, defaults/win-64::h5py-2.8.0-py36hf7173ca_2, defaults/win-64::keras-base-2.2.4-py36_0
  - anaconda/noarch::keras-applications-1.0.8-py_0, anaconda/win-64::keras-base-2.2.4-py36_0, defaults/noarch::keras-preprocessing-1.1.0-py_1, defaults/win-64::h5py-2.8.0-py36hf7173ca_2
  - anaconda/win-64::keras-base-2.2.4-py36_0, defaults/noarch::keras-applications-1.0.8-py_0, defaults/noarch::keras-preprocessing-1.1.0-py_1, defaults/win-64::h5py-2.8.0-py36hf7173ca_2
  - anaconda/noarch::keras-applications-1.0.8-py_0, anaconda/noarch::keras-preprocessing-1.1.0-py_1, anaconda/win-64::keras-base-2.2.4-py36_0, defaults/win-64::h5py-2.8.0-py36hf7173ca_2
  - anaconda/noarch::keras-preprocessing-1.1.0-py_1, anaconda/win-64::keras-base-2.2.4-py36_0, defaults/noarch::keras-applications-1.0.8-py_0, defaults/win-64::h5py-2.8.0-py36hf7173ca_2
  - anaconda/noarch::keras-applications-1.0.8-py_0, anaconda/noarch::keras-preprocessing-1.1.0-py_1, anaconda/win-64::h5py-2.8.0-py36hf7173ca_2, anaconda/win-64::keras-base-2.2.4-py36_0
  - anaconda/noarch::keras-applications-1.0.8-py_0, anaconda/win-64::h5py-2.8.0-py36hf7173ca_2, anaconda/win-64::keras-base-2.2.4-py36_0, defaults/noarch::keras-preprocessing-1.1.0-py_1
  ... and othedone

## Package Plan ##

  environment location: D:\Users\jng\AppData\Local\Continuum\anaconda3

  added / updated specs:
    - bokeh


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    h5py-2.8.0                 |   py36hf7173ca_2         822 KB
    keras-base-2.2.4           |           py36_0         458 KB
    ------------------------------------------------------------
                                           Total:         1.3 MB

The following packages will be SUPERSEDED by a higher-priority channel:

  h5py                                             anaconda --> pkgs/main
  keras-base                                       anaconda --> pkgs/main


Proceed ([y]/n)? y


Downloading and Extracting Packages
keras-base-2.2.4     | 458 KB    |                  |   0%
h5py-2.8.0           | 822 KB    |                  |   0%

[WinError 126] The specified module could not be found
[WinError 126] The specified module could not be found

This is so bewildering because I wanted to update bokeh but why conda wanted to download unrelated packages keras-base-2.2.4 and h5py.-2.8.0 (which had already been installed in my conda base)? Also, what did the message [WinError 126] The specified module could not be found mean? Could someone please help me to resolve the issue? Thanks

1
pythoninstallationanacondabokeh
Don't use the base environment, you should create a new, separate one for each project. I'm actually curious to see if it is (partly?) responsible for the current issue. - AMC
Thanks, @AMC. I will be doing that from now on. It's just that I thought using conda (instead of pip) would take care of dependencies which include correct/relevant versions of packages to be installed/updated within an environment. - Nemo
It's just that I thought using conda (instead of pip) would take care of dependencies which include correct/relevant versions of packages to be installed/updated within an environment. What do you mean? - AMC
I meant if conda knew how to handle an environment logically, it would never allow any new installation/update that would break things that previously worked. - Nemo
How is that related to the current issue, though? - AMC

1 Answers