An Introduction To Venv

In this tutorial you’ll learn a bit about the pyvenv command and the venv module that powers it. You’ll learn how to create self-contained Python environments in order to practice safe development and manage package dependency conflicts.

Working with Virtual Environments

For every package
installed in the
system Python, the
gods kill a kitten
- me

Why Virtual Environments?

  • You will need to install packages that aren’t in the Python standard Library
  • You often need to install different versions of the same library for different projects
  • Conflicts arising from having the wrong version of a dependency installed can cause long-term nightmares
  • Use pyvenv ...
  • Always

Creating a Venv

Since version 3.3, Python has come with a built-in venv module.

To use the module, you can run it using your Python 3 executable:

$ python -m venv my_env

On Windows you’ll need something a bit different:

c:\Temp>c:\Python35\python -m venv my_env

Unless you have the Python executable in your path, in which case this:

c:\Temp>python -m venv my_env


Your Python 3 executable may be python3, please substitute that if required

Depending on how you installed Python (and on your operating system) you may also have a pyvenv command available in your PATH. You can use it like so:

$ pyvenv my_env

In any of these command forms, the name of the new virtual environment (my_env) is arbitrary.

I suggest that you name virtual environments to match the project for which the environment is to be used.

I also suggest that you keep your virtual environments in the same directory as the project code you are writing.

Be aware that venv can be sensitive to path names that contain spaces. Please make sure that the entire path to your working directory does not contain any spaces just to be safe.

Let’s make one for demonstration purposes:

$ python -m venv demoenv
$ ls demoenv
bin     include     lib     pyvenv.cfg

When you ran that command, a couple of things took place:

  • A new directory with your requested name was created
  • A new Python executable was created in <ENV>/bin (<ENV>/Scripts on Windows)
  • The new Python was cloned from your system Python (where virtualenv was installed)
  • The new Python was isolated from any libraries installed in the old Python
  • Setuptools was installed so you have easy_install for this new python
  • Pip was installed so you have pip for this new python


Every virtual environment you create contains an executable Python command.

If you do a quick check to see which Python executable is found by your terminal, you’ll see that it is not the one:

$ which python

in powershell:

$ gcm python

You can execute the new Python by explicitly pointing to it:

$ ./demoenv/bin/python -V
Python 3.5.0

But that’s tedious and hard to remember.

Instead, activate your virtual environment using a shell command:

Platform Shell Activation Command
Posix bash/zsh $ source <venv>/bin/activate
fish $ . <venv>/bin/
csh/tcsh $ source <venv>/bin/activate.csh
Windows cmd.exe C:> <venv>/Scripts/activate.bat
powershell PS C:> <venv>/Scripts/Activate.ps1

Notice that when a virtualenv is active you can see it in your command prompt:


So long as the virtualenv is active the python executable that will be used will be the new one in your demoenv.

Installing Packages

Since pip is also installed, the pip that is used to install new software will also be the one in demoenv.

(demoenv)$ which pip

This means that using these tools to install packages will install them into your virtual environment only

The are not installed into the system Python.

Let’s see this in action.

We’ll install a package called docutils

It provides tools for creating documentation using ReStructuredText

Install it using pip (while your virtualenv is active):

(demoenv)$ pip install docutils
Downloading/unpacking docutils
  Downloading docutils-0.11.tar.gz (1.6MB): 1.6MB downloaded
  Running (path:/Users/cewing/demoenv/build/docutils/ egg_info for package docutils
    changing mode of /Users/cewing/demoenv/bin/ to 755
    changing mode of /Users/cewing/demoenv/bin/ to 755
Successfully installed docutils
Cleaning up...

And now, when we fire up our Python interpreter, the docutils package is available to us:

(demoenv)$ python
Python 3.5.0 (default, Sep 16 2015, 10:42:55)
[GCC 4.2.1 Compatible Apple LLVM 6.1.0 (clang-602.0.49)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import docutils
>>> docutils.__path__
>>> ^d

Like some other Python libraries, the docutils package provides a number of executable scripts when it is installed.

You can see these in the bin directory inside your virtualenv:

(demoenv)$ ls ./demoenv/bin

These scripts are set up to execute using the Python with which they were built.

Running these scripts from this location will use the Python executable in your virtualenv, even if that virtualenv is not active!


So you’ve got a virtual environment created and activated so you can work with it.

Eventually you’ll need to stop working with this venv and switch to another

It’s a good idea to keep a separate venv for every project you work on.

When a venv is active, all you have to do is use the deactivate command:

(demoenv)$ deactivate
$ which python

Note that your shell prompt returns to normal, and now the executable Python found when you check python is the system one again.

Cleaning Up

The final advantage that venv offers you as a developer is the ability to easily remove a batch of installed Python software from your system.

Consider a situation where you installed a library that breaks your Python (it happens)

If you are working in your system Python, you now have to figure out what that package installed

You have to figure out where it is

And you have to go clean it out manually.

With venv you simply remove the directory venv created when you started out.

Let’s do that with our demoenv:

$ rm -r demoenv

And that’s it.

The entire environment and all the packages you installed into it are now gone.

There are no traces left to pollute your world.