10000 ValueError: numpy.dtype has the wrong size, try recompiling · Issue #11383 · pandas-dev/pandas · GitHub
[go: up one dir, main page]

Skip to content
ValueError: numpy.dtype has the wrong size, try recompiling #11383
@anentropic

Description

@anentropic

I have been setting up a new CI server and I'm hitting a problem when running our tests on the new server - tests which use Pandas are giving this error:

File "/root/.virtualenvs/myproject/local/lib/python2.7/site-packages/pandas/__init__.py", line 7, in <module>
    from . import hashtable, tslib, lib
  File "pandas/src/numpy.pxd", line 157, in init pandas.lib (pandas/lib.c:78660)
ValueError: numpy.dtype has the wrong size, try recompiling

I have read other threads about the cause of this problem but they don't seem to apply in my case.

We are running in a virtualenv, inside a Docker container (Debian Stretch).

There is no 'system' numpy, pandas or cython installed:

root@e00c968153d1:/# pip freeze | grep -i numpy
root@e00c968153d1:/# pip freeze | grep -i pandas
root@e00c968153d1:/# pip freeze | grep -i cython

In our requirements.txt we have:

numpy==1.9.1
pandas==0.15.2

but...

$ workon myproject
$ ipython
In [1]: import numpy
In [2]: numpy.version.full_version
Out[2]: '1.9.1'
In [3]: import pandas
numpy.dtype has the wrong size, try recompiling

Now, I have found a manual fix, which is to re-install pandas:

$ pip uninstall pandas
$ pip install --no-cache-dir pandas==0.15.2

The need for --no-cache-dir suggests something about the problem, but I don't understand what exactly. I am installing in a 'virgin' docker container, there is no pip cache before I first install requirements.txt:

$ ls -l /root/.cache/pip/wheels
$ ls: cannot access /root/.cache/pip/wheels: No such file or directory

After pip install -r requirements.txt:

$ locate pandas
/root/.cache/pip/wheels/f7/7b/7c/6de5614135da1e25d864beec01ed40bb43fcafdeda870a05f6/pandas-0.15.2-cp27-none-linux_x86_64.whl
/root/.virtualenvs/myproject/lib/python2.7/site-packages/pandas
...
(only one installation of pandas found)

The manual fix is not acceptable as I need to automate the build.

Appreciate if you can suggest other avenues to explore towards a solution.

Metadata

Metadata

Assignees

No one assigned

    Labels

    BuildLibrary building on various platforms

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

      0