Basic Conda Guide¶

What is Conda¶

It is short for Anaconda, the most popular Python data science platform. It is a Python and R distribution. It aims to provide almost everything you need for data science as it makes it easy to install Python plus a number of it's most often used 3rd party libraries in a flexible way. Personally, I prefer it over Vanilla and other virtual environments.

How to Install Conda¶

Official guide to installation for different OS

1. Download graphical installer according to your OS and choose your Python version.
   macOS
   Windows
   For Linux, only command-line install is supported. Check the official guide above.
2. Open the installer and follow the instructions

Verify the Installation¶

1. Open Anaconda Prompt (or Terminal on Linux or macOS):

   • Windows: Open the Anaconda Prompt (Click Start, select Anaconda Prompt)
   • macOS: Open Launchpad, then open Terminal.
   • Linux–CentOS: Open Applications - System Tools - Terminal.
   • Linux–Ubuntu: Open the Dash by clicking the upper left Ubuntu icon, then type “terminal”.

2. Enter conda list. If Anaconda is installed and working, this will display a list of installed packages and their versions.

Getting Started¶

• Update conda to the current version
  conda update conda
• Create a new environment, assign a name to it, and specify its Python version
  conda create --name <envname> python=<version:2.7/3.5>
• Activate this environment. If successfull, (envname) should show up in front of the prompt in the terminal
  conda activate <envname>
• Use regular installation commands to install packages after environment succesfully activated
  pip install -r requirements.txt
• Check if a package is available from the Anaconda repo
  conda search <packagename>
• Show a list of all the packages installed in this environment with an asterisk after the name of the active environment
  conda list
• Export a list of all installed packages to txt
  conda list -e > requirements.txt
• Show a list of all your conda environments
  conda info --envs

NOTE¶

• When using regular pip install for large numbers of packages, the entire installation process could freeze or fail due to a single error in the numerous packages. In such case, run the following command instead to retain successful installations.
  while read requirement; do conda install --yes $requirement; done < requirements.txt

Official Cheat Sheet¶