Skip to content

Conda Guide

Conda is a powerful package and environment management system, particularly popular in data science and scientific computing. It can manage both Python packages and system-level dependencies.

What is Conda?

Conda is an open-source package management system and environment management system that runs on Windows, macOS, and Linux. It can:

  • Install packages and their dependencies
  • Create isolated environments
  • Manage multiple Python versions
  • Handle binary packages (not just Python)
  • Work with pip and other package managers

Installation

Miniconda is a minimal installer that includes conda, Python, and a few essential packages:

# Download Miniconda for macOS (Apple Silicon)
cd ~/Downloads
curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-arm64.sh

# Download Miniconda for macOS (Intel)
curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh

# Run installer
bash Miniconda3-latest-MacOSX-arm64.sh  # or x86_64.sh for Intel

# Follow prompts:
# - Press Enter to review license
# - Type 'yes' to accept
# - Press Enter to confirm installation location
# - Type 'yes' to initialize conda

Install Anaconda (Full Distribution)

Anaconda includes conda plus 150+ pre-installed packages:

# Download from: https://www.anaconda.com/products/distribution
# Or use Homebrew:
brew install --cask anaconda

Initialize Conda

After installation, restart your terminal or run:

source ~/.zshrc  # or ~/.bash_profile for bash

Verify Installation

conda --version
conda info

Basic Usage

Creating Environments

# Create environment with Python 3.10
conda create -n myenv python=3.10

# Create environment with specific packages
conda create -n myenv python=3.10 numpy pandas matplotlib

# Create environment from file
conda env create -f environment.yml

Activating and Deactivating

# Activate environment
conda activate myenv

# Deactivate environment
conda deactivate

# Activate base environment
conda activate base

Installing Packages

# Install a package
conda install numpy

# Install multiple packages
conda install numpy pandas matplotlib

# Install from specific channel
conda install -c conda-forge package-name

# Install specific version
conda install numpy=1.24.0

Listing and Searching

# List all environments
conda env list
# or
conda info --envs

# List packages in current environment
conda list

# List packages in specific environment
conda list -n myenv

# Search for packages
conda search numpy

# Search in specific channel
conda search -c conda-forge package-name

Removing Environments and Packages

# Remove a package
conda remove numpy

# Remove environment
conda env remove -n myenv

# Remove all unused packages
conda clean --all

Best Practices

1. Use Environments for Each Project

Always create a separate environment for each project:

# Good: Project-specific environment
conda create -n project1 python=3.10
conda activate project1
conda install project-dependencies

# Bad: Installing everything in base
conda install everything

2. Use Environment Files

Create environment.yml files for reproducible environments:

name: my-project
channels:
  - conda-forge
  - bioconda
  - defaults
dependencies:
  - python=3.10
  - numpy>=1.20.0
  - pandas>=1.3.0
  - matplotlib>=3.4.0
  - pip
  - pip:
    - some-pip-only-package

Usage: 

# Create from file
conda env create -f environment.yml

# Update existing environment
conda env update -f environment.yml --prune

# Export current environment
conda env export > environment.yml

# Export without build numbers (more portable)
conda env export --no-builds > environment.yml

3. Use Channels Wisely

Channels are repositories where conda looks for packages:

# Add channels (order matters - first checked first)
conda config --add channels conda-forge
conda config --add channels bioconda

# List channels
conda config --show channels

# Install from specific channel
conda install -c conda-forge package-name

# Set channel priority
conda config --set channel_priority strict

Recommended channel order: 1. conda-forge - Community-maintained packages 2. bioconda - Bioinformatics packages 3. defaults - Anaconda's default channel

4. Keep Base Environment Clean

Don't install packages in base environment:

# Good: Create new environment
conda create -n myproject python=3.10
conda activate myproject

# Bad: Installing in base
conda activate base
conda install packages  # Don't do this!

5. Use Mamba for Faster Installs

Mamba is a faster drop-in replacement for conda:

# Install mamba
conda install mamba -n base -c conda-forge

# Use mamba instead of conda
mamba create -n myenv python=3.10
mamba install numpy pandas
mamba env update -f environment.yml

6. Regular Maintenance

# Update conda
conda update conda

# Update all packages in environment
conda update --all

# Clean cache and unused packages
conda clean --all

# Remove unused environments
conda env remove -n old-env

Working with pip

When to Use pip vs conda

Use conda for: - ✅ Packages available in conda (numpy, pandas, scipy, etc.) - ✅ Packages with C/C++ dependencies - ✅ System-level dependencies

Use pip for: - ✅ Packages only available on PyPI - ✅ Development versions - ✅ Packages not in conda channels

Best Practice: Install conda packages first, then pip

# Create environment
conda create -n myenv python=3.10

# Install conda packages first
conda install numpy pandas matplotlib

# Then install pip-only packages
pip install some-pip-only-package

Including pip packages in environment.yml

name: my-project
dependencies:
  - python=3.10
  - numpy
  - pip
  - pip:
    - package-only-on-pypi
    - another-pip-package

Common Workflows

Workflow 1: New Project Setup

# 1. Create environment
conda create -n myproject python=3.10

# 2. Activate
conda activate myproject

# 3. Install packages
conda install numpy pandas matplotlib jupyter

# 4. Install pip-only packages
pip install some-package

# 5. Export environment
conda env export > environment.yml

# 6. Add to git
git add environment.yml

Workflow 2: Sharing Environment

# Export environment
conda env export --no-builds > environment.yml

# Share file with team

# Others can recreate:
conda env create -f environment.yml
conda activate myproject

Workflow 3: Updating Environment

# Update specific package
conda update numpy

# Update all packages
conda update --all

# Update from environment file
conda env update -f environment.yml --prune

Workflow 4: Multiple Python Versions

# Create environments with different Python versions
conda create -n py39 python=3.9
conda create -n py310 python=3.10
conda create -n py311 python=3.11

# Switch between them
conda activate py39
# work with Python 3.9

conda activate py310
# work with Python 3.10

Advanced Usage

Cloning Environments

# Clone an environment
conda create --name newenv --clone oldenv

# Useful for testing updates
conda create --name test-update --clone production
conda activate test-update
conda update --all

Managing Channels

# Show channel configuration
conda config --show channels

# Add channel
conda config --add channels conda-forge

# Remove channel
conda config --remove channels channel-name

# Set channel priority
conda config --set channel_priority strict  # Prefer higher priority
conda config --set channel_priority flexible  # Try all channels

Environment Variables

# Set environment variable in environment
conda env config vars set MY_VAR=value -n myenv

# List environment variables
conda env config vars list -n myenv

# Unset variable
conda env config vars unset MY_VAR -n myenv

Installing from Git

# Install package from git
pip install git+https://github.com/user/repo.git

# Install editable (development mode)
pip install -e git+https://github.com/user/repo.git#egg=package

Troubleshooting

Issue: Environment Not Activating

Problem: conda activate doesn't work

Solution: 

# Initialize conda
conda init zsh  # or bash

# Restart terminal or source
source ~/.zshrc

Issue: Package Not Found

Problem: PackagesNotFoundError

Solution: 

# Search for package
conda search package-name

# Try different channels
conda install -c conda-forge package-name
conda install -c bioconda package-name

# Use pip if not in conda
pip install package-name

Issue: Dependency Conflicts

Problem: Conflicting package versions

Solution: 

# Let conda solve dependencies
conda install package-name --solver=libmamba

# Or use mamba (faster solver)
mamba install package-name

# Create fresh environment
conda create -n fresh python=3.10 package-name

Issue: Slow Installation

Problem: Conda is slow

Solution: 

# Use mamba (much faster)
conda install mamba -n base -c conda-forge
mamba install packages

# Use conda-forge channel (often faster)
conda install -c conda-forge package-name

Issue: Environment Takes Too Much Space

Problem: Environments are large

Solution: 

# Clean unused packages
conda clean --all

# Remove old environments
conda env remove -n old-env

# Use miniconda instead of anaconda

Conda vs Other Tools

Conda vs pip

Feature Conda pip
Python packages ✅ ✅
Non-Python dependencies ✅ ❌
Environment management ✅ ⚠️ (venv)
Binary packages ✅ ⚠️ (wheels)
Dependency solving ✅ ⚠️ (basic)

Conda vs venv

Feature Conda venv
Multiple Python versions ✅ ❌
Non-Python packages ✅ ❌
Cross-platform ✅ ✅
Lightweight ⚠️ ✅

Best Practice: Use conda for scientific computing, use venv for simple Python projects.

Quick Reference

# Environments
conda create -n envname python=3.10
conda activate envname
conda deactivate
conda env list
conda env remove -n envname

# Packages
conda install package
conda install -c conda-forge package
conda list
conda search package
conda update package
conda remove package

# Environment files
conda env export > environment.yml
conda env create -f environment.yml
conda env update -f environment.yml

# Maintenance
conda update conda
conda update --all
conda clean --all

Last updated: November 23, 2025