在python-3.4 unix环境中导入熊猫时出错-未定义符号：pyobject_cmp

Question

I was using Pandas version 0.16.1 and it was working fine but I wanted to update Pandas to version 0.16.2 so I uninstalled it and reinstalled it using

pip uninstall pandas
pip install pandas

The installation seemed to work fine. However, when I attempt to import pandas

import pandas

I get the error:

Traceback (most recent call last):
File "path_to_python/python3.4/site-packages/pandas/__init__.py", line 7, in <module>
from pandas import hashtable, tslib, lib
ImportError: path_to_python/python3.4/site-packages/pandas/hashtable.cpython-34m.so: undefined symbol: pyobject_cmp

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "path_to_python/python3.4/site-packages/pandas/__init__.py", line 13, in <module>
"extensions first.".format(module))
ImportError: C extension: path_to_python/python3.4/site-packages/pandas/hashtable.cpython-34m.so:
undefined symbol: pyobject_cmp not built. If you want to import pandas from the source directory, you may need to run 'python setup.py build_ext --inplace' to build the C extensions first.

However:

cd path_to_python/python3.4/site-packages/pandas/
python setup.py build_ext --inplace

doesn't work either as there is no setup.py to be found. The error is:

Fatal Python error: Py_Initialize: can't initialize sys standard streams
AttributeError: 'module' object has no attribute 'OpenWrapper'
Aborted (core dumped)

Any advice or insight into what could be going wrong would be greatly appreciated!

UPDATE:

I tried to install it from source by downloading the package:

python setup.py install
python setup.py build_ext --inplace

This time the second command generated a lot of output and successfully completed. However, I still get the same error when I try to import pandas

Answer 1

如果有人感兴趣，我无法让pip安装熊猫，但我改用miniconda并安装了它，它工作正常。

Answer 2

I was unable to build pandas using clang due to compiler flags. Clang using C99 model by default which leads to improper compilation of inline code used in pandas.

So in my case CFLAGS=-fgnu89-inline pip install pandas --global-option=build_ext did the trick.