如何通过Anaconda切换Python 3.x/3.y版本？HDP集群多版本共存方案

Absolutely, running both Python 3.5 and 3.6 side-by-side with Anaconda is not only possible—it’s a standard practice for managing multiple Python versions without conflicts. Let’s walk through your questions and the general workflow:

Yes, 100%. Conda’s environment system is designed exactly for this scenario: each environment is an isolated sandbox with its own Python version, packages, and dependencies. This means your existing Python 3.5 setup (and any HDP-related integrations relying on it) will remain completely untouched while you set up and use Python 3.6 in a separate environment.

Since you’re working on an HDP cluster, just ensure you have the necessary permissions to create conda environments (most clusters allow user-specific environments in home directories, even if the base Anaconda is installed system-wide).

Assuming your current Python 3.5 is running in the default base conda environment (or a custom environment you’ve already set up), here’s how to add Python 3.6 and toggle between them:

• First, create a dedicated Python 3.6 environment
  Run this command in your cluster’s terminal:

  conda create -n py36 python=3.6
  

  If your team relies on specific packages (like pyspark, numpy, or pandas for HDP workflows), include them in the command to install them alongside Python 3.6:

  conda create -n py36 python=3.6 pyspark numpy pandas
  

• Activate the Python 3.6 environment
  For Linux-based HDP clusters, use:

  conda activate py36
  

  If you’re using an older version of conda that doesn’t support conda activate, use:

  source activate py36
  

  Once activated, run python --version to confirm you’re on Python 3.6.

• Switch back to Python 3.5
  To return to your original Python 3.5 environment (e.g., the base environment), run:

  conda activate base
  

  If your Python 3.5 is in a custom-named environment (like py35), replace base with that environment name.

This workflow applies to any pair of Python 3 versions, not just 3.5 and 3.6:

1. Create an environment for your target Python version
   Replace <version> with your desired Python release (e.g., 3.7, 3.9):

   conda create -n py<version> python=<version>
   

   Example: conda create -n py38 python=3.8

2. Activate the target environment

   • Linux/macOS: conda activate py<version>
   • Windows (Command Prompt/PowerShell): conda activate py<version>

3. Verify the active version
   Confirm you’re using the correct Python with:

   python --version
   

   You can also check the path to ensure it’s pointing to the conda environment:

   # Linux/macOS
   which python
   # Windows
   where python
   

4. Switch to another environment
   Simply activate the environment you want to use:

   conda activate <environment-name>
   

   This could be the base environment, or another custom environment you’ve created.

Quick Tips for HDP Clusters #

• Avoid modifying the base conda environment’s Python version—this can break core Anaconda functionality and cause conflicts with HDP tools.
• If your team needs to share a consistent environment, export the environment config:

  conda env export -n py36 > hdp_py36_env.yml
  

  Team members can then recreate the environment with:

  conda env create -f hdp_py36_env.yml
  

• When submitting Spark jobs, specify the conda environment’s Python path to ensure Spark uses the correct version:

  spark-submit --conf spark.pyspark.python=/home/your-user/anaconda3/envs/py36/bin/python your_script.py
  

内容的提问来源于stack exchange，提问作者HISI