导入pandas出现ImportError:无法确定numexpr版本的问题求助

问题详情 #

• 导入命令：

import pandas as pd

• 错误信息：

Traceback (most recent call last):
 
  Cell In[54], line 1
    import pandas as pd

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\__init__.py:48
    from pandas.core.api import (

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\api.py:27
    from pandas.core.arrays import Categorical

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\arrays\__init__.py:1
    from pandas.core.arrays.arrow import ArrowExtensionArray

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\arrays\arrow\__init__.py:1
    from pandas.core.arrays.arrow.array import ArrowExtensionArray

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\arrays\arrow\array.py:60
    from pandas.core.arraylike import OpsMixin

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\arraylike.py:21
    from pandas.core.ops.common import unpack_zerodim_and_defer

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\ops\__init__.py:38
    from pandas.core.ops.array_ops import (

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\ops\array_ops.py:57
    from pandas.core.computation import expressions

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\computation\expressions.py:20
    from pandas.core.computation.check import NUMEXPR_INSTALLED

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\core\computation\check.py:5
    ne = import_optional_dependency("numexpr", errors="warn")

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\compat\_optional.py:157 in import_optional_dependency
    version = get_version(module_to_get)

  File ~\AppData\Local\anaconda3\lib\site-packages\pandas\compat\_optional.py:84 in get_version
    raise ImportError(f"Can't determine version for {module.__name__}")

ImportError: Can't determine version for numexpr

• Python版本：

Python 3.10.10 | packaged by Anaconda, Inc. | (main, Mar 21 2023, 18:39:17) [MSC v.1916 64 bit (AMD64)]

• 环境限制：需通过VPN远程访问服务器，仅登录VPN时可访问程序

解决方案 #

• 重新安装numexpr包
  登录VPN后，打开Anaconda Prompt执行以下命令，卸载并重新安装numexpr以确保包完整性：

  conda remove numexpr
  conda install numexpr
  

  若conda安装失败，可改用pip：

  pip uninstall numexpr
  pip install numexpr
  

• 指定兼容版本安装
  若自动安装的版本仍有问题，可指定与Python 3.10和当前Pandas版本兼容的numexpr版本（如2.8.4）：

  conda install numexpr=2.8.4
  

  或pip：

  pip install numexpr==2.8.4
  

• 修复Anaconda环境
  若环境损坏导致版本识别异常，执行以下命令清理并更新环境（需登录VPN）：

  conda clean --all
  conda update --all
  

• 临时禁用numexpr（应急方案）
  若无法立即重新安装，可在导入Pandas前设置环境变量跳过numexpr检查（会禁用Pandas的numexpr加速功能）：

  import os
  os.environ['PANDAS_NUMEXPR_ENABLED'] = '0'
  import pandas as pd
  

内容的提问来源于stack exchange，提问作者Victor Nielsen