Trouble updating to Spyder 4.0.0

(Spyder maintainer here) Other users have reported that you need to run

conda update anaconda

and then

conda install spyder=4

to successfully update to version 4.

Hope that helps.

I used the following:

conda update anaconda

and then

conda update spyderconda install spyder=4.0.1

to successfully update to version 4.0.1

Just an additional note here for those trying to upgrade to Spyder 4 and use conda environments in Windows 10. I was wanting to use Spyder 4 but also getting the error inside Spyder saying that it could not find the spyder-kernels. This happened regardless of (1) whether or not I installed Spyder inside a new environment or the base environment, and (2) even if I subsequently installed spyder-kernels through conda.

I tried to follow the advice given above and already had Anaconda installed (via the windows installer) but when I tried

conda update anaconda

It could not find Anaconda inside the base environment. So I installed Anaconda using conda anyway (even though this seems kind of recursive to me)

conda install anaconda

After that I was able to update Anaconda by:

conda update anaconda

Once I did that I was able to run Spyder 4 from the base environment (already installed) and I was getting a working Ipython kernel.

So then I could open Spyder from the base environment and using the Spyder menu, Tools -> Preferences -> Python Interpreter, I can set the conda environment I want to be in, eg,C:\Users\\Anaconda3\envs\myenv\python.exe

Then, to allow Spyder to create a Ipython Kernel inside the conda environment that I would like to work in, I have to install spyder-kernels into this environment, so, I started up a new Anaconda prompt, and:

conda activate myenvconda install spyder-kernels

Finally, back in Spyder, I restarted my kernel (Ctrl+.). This gave me a functioning Ipython kernel , and looking at the status bar at the bottom of the Spyder window, I could see that I was in the myenv environment.

So now I can use the Spyder IDE (starting it from the base environment) but be able to work inside previously created conda environments. It also avoids having to install the full Spyder into each environment I create (but does mean having to install spyder-kernels into each environment).

I know this is a long comment, but I spent quite a while looking around before I could solve this exact problem, so I am hoping it will be useful for others who are also facing the same issue.