How to manage python versions on windows

文章目录

      • [Using the py.exe Launcher](#Using the py.exe Launcher)
        • [Installation and Setup](#Installation and Setup)
        • [Basic Usage](#Basic Usage)
      • [Using pyenv-win](#Using pyenv-win)
      • [Other Considerations](#Other Considerations)

Managing multiple Python versions on Windows can be done through a few reliable methods, depending on your needs for simplicity, flexibility, or project isolation. I'll cover the two most popular approaches: the built-in py.exe launcher (which comes with Python 3 installations) and pyenv-win (a port of the Unix pyenv tool). These allow you to install, switch, and run different versions without conflicts. For project-specific isolation, combine these with Python's built-in venv module to create virtual environments tied to specific versions.

Using the py.exe Launcher

The py.exe launcher is a lightweight tool included with Python 3 installations on Windows. It helps discover installed Python versions, select specific ones for execution, and handle scripts with shebang lines. It's ideal if you just need to run different versions without advanced management.

Installation and Setup
  • When installing Python from the official website (python.org), ensure you select the option to install the launcher during setup. It's optional, so check the box for "py launcher" or similar.
  • If it's not installed, reinstall Python and enable it. The launcher will be added to your PATH automatically.
  • For older versions like Python 2.7, you may need to manually add registry keys for detection (under HKEY_LOCAL_MACHINE\Software\Python\PythonCore or user-specific equivalents) if they're missing.
Basic Usage
  • List installed Python versions : Run py -0 to see all detected versions and architectures (e.g., -3.9-64 * indicates the default). Use py -0p to include full paths to python.exe.

  • Run a specific version : Use flags like py -3.8 to run Python 3.8, or py -2 for the latest Python 2.x. For example:

    复制代码
    py -3.8 --version

    This outputs the version info for Python 3.8.

  • Run scripts : Execute a script with a specific version, e.g., py -3.7 myscript.py. If your script has a shebang (e.g., #!/usr/bin/env python3), running py myscript.py will automatically use the matching version.

  • Default behavior : Without flags, py uses the latest installed version. To check: py --version.

This method doesn't handle installing new versions---you'll need to download and install them manually from python.org---but it's great for switching between existing ones.

Using pyenv-win

For more comprehensive management, including easy installation and switching of versions, use pyenv-win. It's a Windows-specific tool that mirrors the Unix pyenv and lets you install, list, set global/local versions, and uninstall Pythons without affecting your system PATH.

Installation
  1. Open PowerShell (run as administrator if needed).

  2. Download and run the installation script:

    复制代码
    Invoke-WebRequest -UseBasicParsing -Uri "https://raw.githubusercontent.com/pyenv-win/pyenv-win/master/pyenv-win/install-pyenv-win.ps1" -OutFile "./install-pyenv-win.ps1"; &"./install-pyenv-win.ps1"
    • If you get an execution policy error, run Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope LocalMachine first.
  3. Reopen PowerShell.

  4. Verify with pyenv --version.

  5. Update the version database: pyenv update.

  6. (Optional but recommended for Windows 10+): Disable built-in Python app aliases in Settings > Apps > Apps & features > App execution aliases, and turn off Python-related ones.

  7. Add these to your system PATH (via System Properties > Environment Variables) if not added automatically:

    • C:\Users\<YourUsername>\.pyenv\pyenv-win\bin
    • C:\Users\<YourUsername>\.pyenv\pyenv-win\shims
  8. Restart your terminal or IDE.

Basic Usage
  • List available versions to install : pyenv install --list (or pyenv install -l). Filter with pyenv install -l | findstr 3.8 for specific ones.
  • Install a version : pyenv install <version>, e.g., pyenv install 3.13.5. For multiple: pyenv install 3.13.5 3.14.0. Use -q for quiet mode (no prompts).
    • Supports 32-bit with suffixes like 3.8.2-win32.
  • List installed versions : pyenv versions.
  • Check current version : pyenv version.
  • Set versions :
    • Global (system-wide): pyenv global <version>, e.g., pyenv global 3.13.5. Use pyenv global system to revert to your system's default Python.
    • Local (project directory): pyenv local <version>. This creates a .python-version file in the folder for automatic switching.
    • Shell (current session): pyenv shell <version>. Unset with pyenv shell --unset.
  • Uninstall a version : pyenv uninstall <version>.
  • Rehash shims : After installing packages or switching versions, run pyenv rehash to update executable paths.
  • Run commands with a version : pyenv exec python myscript.py uses the current version's Python.
Tips
  • Always run pyenv rehash after changes to ensure tools like pip work correctly.
  • For project isolation, after setting a version, create a virtual environment: python -m venv myenv, then activate it.
  • pyenv-win builds versions from source, so installations can take time but support a wide range (Python 2.4+).

Other Considerations

  • Manual Management : Install versions from python.org into separate directories (e.g., C:\Python38), then call them explicitly like C:\Python38\python.exe myscript.py or adjust PATH temporarily. This is basic but error-prone for frequent switches.
  • Anaconda/Miniconda : If you work with data science, use Miniconda to install different Python versions via environments (e.g., conda create -n py38 python=3.8). It's heavier but includes package management.
  • Best Practices: Avoid modifying your system PATH globally for Pythons to prevent conflicts. Test versions in isolated environments, and keep your default system Python untouched for OS compatibility.

If you run into issues, check your environment variables or restart your terminal. For the latest updates, refer to the official pyenv-win repo or Python docs.

相关推荐
jufeng13079 小时前
【系列:手搓自主 AI Agent:Hermes 架构原理剖析 · 第 8 篇】
python·ai agent·配置系统
circuitsosk9 小时前
跨境电商智能化实战:AI如何赋能客服自动回复、广告智能投放与供应链预测
大数据·人工智能·python·langchain·智能客服
2601_9563198810 小时前
2026年零基础学量化:从看懂示例到写清条件和动作
人工智能·python
知无不研10 小时前
lambda表达式的使用(3)
开发语言·c++·lambda
过期的秋刀鱼!10 小时前
LangChain-D1-模型的工作流程
人工智能·python·langchain
萤火工厂目视化设计10 小时前
智能制造与企业文化目视化浪潮下:中小工厂的机遇与挑战
python·制造
cfm_291410 小时前
SpringAI + Ollama 本地大模型
java·开发语言·人工智能·语言模型
DeepVisionary11 小时前
从GPT-5.6 Sol的Ultrafast模式看推理速度竞赛:750 token/秒背后,Cerebras的晶圆级生意
python·自动化
C++ 老炮儿的技术栈12 小时前
基于Qt实现轻量化本地音乐播放器
开发语言·c++·qt·c·播放器·音乐
qq_4480111612 小时前
C语言中的柔性数组
c语言·开发语言·柔性数组