In Windows, I created a Conda virtual environment with the command
conda create -n test python=2.7 pandas scipy matplotlib numpy
Once it is
For my case, I have also the import numpy DLL error in anaconda3, and Decrayer (Nov 28) provided the best solution: I had the same error except for Anaconda3: just added the path \Anaconda3\Library\bin to your Windows 10 path variable and then it worked.
I have been trying to install anaconda3.7.1 or 3.7.0 without luck, going to full length of restarting the PC after the key steps to make sure the DLL is no longer in use in memory. Ultimately, thanks to decrayer, now numpy works for me!
Unlike @Rafael, for me, libiomp5md.dll
wasn't the issue. I installed Dependency Walker to investigate what was going on. Even though the dll versions were different, but Dependency Walker said it was okay.
What was wrong though, was that mkl_intel_thread.dll
had warnings (red icon). If you're using Win 8++, ignore the api-win
and ext-ms
issues as Dependency Walker wasn't updated for new Windows versions and doesn't recognise Windows new APIs.
My solution is to copy all mkl_*.dlls
from the former to the latter:
I was able to import numpy and sklearn after that.
Uninstall and install numpy again.
pip uninstall numpy
pip install numpy
Then try import again, it should work. That is what I did
For the record, I had the same error here (Python 3.5 64-bit on Windows 10), and this page helped me find the solution. The problem was a conflict with libiomp5md.dll
, which existed on two locations:
Python was trying to use the version in System32 folder, which was an old version. I removed it (renamed) and now it uses the correct version, on Anaconda3 folder, and now I can import numpy without the import error.
For me this was solved by adding the following paths to my system path variable.
C:\Users\UserName\Anaconda3\
C:\Users\UserName\Anaconda3\bin
C:\Users\UserName\Anaconda3\Scripts
C:\Users\UserName\Anaconda3\Library\mingw-w64\bin (not strictly necessary)
C:\Users\UserName\Anaconda3\Library\bin
Before doing this, you could test this by adding these paths manually:
base_path = r"C:\Users\UserName\Anaconda3"
path = os.pathsep.join([os.path.join(base_path, i) for i in [r"", r"bin", r"Scripts", r"Library\mingw-w64\bin", r"Library\bin"]])
os.environ["PATH"]+=os.pathsep+path
Thanks to this post on PyCharm support.
It seems the proper way to fix this is to do:
conda install msvc_runtime
If you are in a virtual environment, add this package there.