巨蟒之战

Question

I have learned Python long ago and when I was just installing Python (I might have done some mistakes on installing it), I have upgraded from 3.6 (32-bit) to 3.7.4 (64-bit). One day I have found that I have (by mistake?) installed 3.7 (64-bit), too.

Now I am working on some problem and sometimes I need to install some new libraries, for example, OpenCV for detecting and processing images. But when I install something with 'pip', it installs it in the directory Python36-32. I have Python IDLE 3.7.4, so there are constantly appearing errors between different Pythons. I have also deleted Python 3.6 (32-bit), but then I couldn't use 'pip' command anymore. I have also tried to change environmental variables from Python36-32 to Python37-64. I was chroming a lot but without success.

I am not very new in Python as programming, but I am just a newbie on a real understanding of how Python works (I mean there for example: what are the differences between different Pythons, what are environmental variables ...)

In general : When I was a newbie on Python programming, I have (by mistake) installed 3 Python versions: 3.6 (32-bit), 3.7 (64-bit), and 3.7.4 (64-bit). My IDLE works in 3.7.4 (64-bit) and my 'pip' installs all the libraries into Python 3.6 (32-bit). My computer is the battlefield for a lot of Pythons. I have no idea what to do. Should I delete some version, change something ...

Thank you very much for solving the problem!

PS I have realized that for me this problem (battle of Pythons) is much harder than the problem on which I am doing.

Answer 1

I have a similar experience when I first using python before grad school. For me, it is because the OS I was using at the time, windows of course. Even now, I still feel windows is not a good OS for programming. One library management solution ana(conda) helped me a lot. Right now conda has GUI interface now, and it provides a much new-programmer friendly entry point for using python. Another huge leg-up during that time is the Jupyter notebook (used to called Ipython), it allows have a semi-console and WYSIWYG programming. Right now, we all use debuggers, but still it is a very intuitive way to programming.

Hope this helps.

Answer 2

Different versions of Python can coexist on your machine simultaneously.

Almost all operating systems include a system version of Python which is installed alongside the operating system and which should not be changed; if you run the command python in your terminal out of the box, this is the version of Python you will get. This system Python should not be modified, as doing so has the potential of breaking OS-installed tools and libraries that depend on it.

Installing Python on your machine (via the Python website) sets up a new Python on your machine which has priority over system Python. However, this is an "prepend" process, not a "replace" process. Each new version of Python installed will have priority over the previous versions. But , if you try to import a package, and that package doesn't exist in your highest priority Python version, your computer will rifle through the old Python's module file-paths to try to find them. This is bad because all of the sudden you are using packages from different versions of Python, which may break due to changes between language versions.

Best practice is to only ever install a single version of Python on your machine (besides the system Python), and then use virtualenv or conda to manage setting up new work environment. So you should install all versions of Python except for latest, and then learn how to use these tools to avoid environment problems.

Answer 3

This is the reason why virtual environments are encouraged these days. Many modern ide like pycharm make it simple to have each project made in an virtual environment. According to the python version you could install libraries which would not disturb any other projects or global environment/setting.

So the best way to stay away from these kind of issues is the use of virtual environment. Hope this helps in further projects.

This article was a good start for me https://realpython.com/python-virtual-environments-a-primer/

Would edit more links if I do find