Numpy: ModuleNotFoundError: No module named 'numpy.core._multiarray_umath'

Hi,

I got a problem similar to #11871. I recently installed numpy 1.16.0 but then I needed to downgrade and now I keep getting ModuleNotFoundError: No module named 'numpy.core._multiarray_umath'.
The example below was working fine a week ago or so.

I actually ended up removing ALL Python packages (instructions here) and reinstalling them _one by one_ using apt-get (Ubuntu 18.04.1 LTS), except for ipyparallel and openbabel, which were installed using pip (I'm pretty sure both are unrelated to the problem). The example below was run _after_ all this clean-up.

It seems that a lot of people are having this issue lately, with different packages:

• ModuleNotFoundError: No module named 'numpy.core._multiarray_umath'
• Encountered error when python3 import tensorflow - No module named 'numpy.core._multiarray_umath'

Reproducing code example and error message:

Below is a IPython session that shows the error:

% ipython3
Python 3.6.7 (default, Oct 22 2018, 11:32:17) 
Type "copyright", "credits" or "license" for more information.

IPython 5.5.0 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]: import pandas as pd

In [2]: pd.read_hdf("data.h5")
---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-2-122d644764c1> in <module>()
----> 1 pd.read_hdf("data.h5")

/usr/lib/python3/dist-packages/pandas/io/pytables.py in read_hdf(path_or_buf, key, mode, **kwargs)
    368                                      'contains multiple datasets.')
    369             key = candidate_only_group._v_pathname
--> 370         return store.select(key, auto_close=auto_close, **kwargs)
    371     except:
    372         # if there is an error, close the store

/usr/lib/python3/dist-packages/pandas/io/pytables.py in select(self, key, where, start, stop, columns, iterator, chunksize, auto_close, **kwargs)
    715                            chunksize=chunksize, auto_close=auto_close)
    716 
--> 717         return it.get_result()
    718 
    719     def select_as_coordinates(

/usr/lib/python3/dist-packages/pandas/io/pytables.py in get_result(self, coordinates)
   1455 
   1456         # directly return the result
-> 1457         results = self.func(self.start, self.stop, where)
   1458         self.close()
   1459         return results

/usr/lib/python3/dist-packages/pandas/io/pytables.py in func(_start, _stop, _where)
    708             return s.read(start=_start, stop=_stop,
    709                           where=_where,
--> 710                           columns=columns, **kwargs)
    711 
    712         # create the iterator

/usr/lib/python3/dist-packages/pandas/io/pytables.py in read(self, start, stop, **kwargs)
   2902             blk_items = self.read_index('block%d_items' % i)
   2903             values = self.read_array('block%d_values' % i,
-> 2904                                      start=_start, stop=_stop)
   2905             blk = make_block(values,
   2906                              placement=items.get_indexer(blk_items))

/usr/lib/python3/dist-packages/pandas/io/pytables.py in read_array(self, key, start, stop)
   2457 
   2458         if isinstance(node, tables.VLArray):
-> 2459             ret = node[0][start:stop]
   2460         else:
   2461             dtype = getattr(attrs, 'value_type', None)

/usr/lib/python3/dist-packages/tables/vlarray.py in __getitem__(self, key)
    669                 key += self.nrows
    670             (start, stop, step) = self._process_range(key, key + 1, 1)
--> 671             return self.read(start, stop, step)[0]
    672         elif isinstance(key, slice):
    673             start, stop, step = self._process_range(

/usr/lib/python3/dist-packages/tables/vlarray.py in read(self, start, stop, step)
    813         atom = self.atom
    814         if not hasattr(atom, 'size'):  # it is a pseudo-atom
--> 815             outlistarr = [atom.fromarray(arr) for arr in listarr]
    816         else:
    817             # Convert the list to the right flavor

/usr/lib/python3/dist-packages/tables/vlarray.py in <listcomp>(.0)
    813         atom = self.atom
    814         if not hasattr(atom, 'size'):  # it is a pseudo-atom
--> 815             outlistarr = [atom.fromarray(arr) for arr in listarr]
    816         else:
    817             # Convert the list to the right flavor

/usr/lib/python3/dist-packages/tables/atom.py in fromarray(self, array)
   1226         if array.size == 0:
   1227             return None
-> 1228         return six.moves.cPickle.loads(array.tostring())

ModuleNotFoundError: No module named 'numpy.core._multiarray_umath'

Some relevant files and versions:

In [4]: import pandas as pd

In [5]: pd.__file__
Out[5]: '/usr/lib/python3/dist-packages/pandas/__init__.py'

In [6]: pd.__version__
Out[6]: '0.22.0'

In [7]: pd.__path__
Out[7]: ['/usr/lib/python3/dist-packages/pandas']

In [8]: import tables

In [9]: tables.__file__
Out[9]: '/usr/lib/python3/dist-packages/tables/__init__.py'

In [10]: tables.__version__
Out[10]: '3.4.2'

In [11]: tables.__path__
Out[11]: ['/usr/lib/python3/dist-packages/tables']

In [12]: import six

In [13]: six.__file__
Out[13]: '/usr/lib/python3/dist-packages/six.py'

In [14]: six.__version__
Out[14]: '1.11.0'

In [15]: six.__path__
Out[15]: []

Numpy/Python version information:

In [1]: import sys, numpy; print(numpy.__version__, sys.version)
1.13.3 3.6.7 (default, Oct 22 2018, 11:32:17) 
[GCC 8.2.0]

In [2]: import pandas; pandas.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 3.6.7.final.0
python-bits: 64
OS: Linux
OS-release: 4.15.0-43-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_GB.UTF-8
LOCALE: pt_BR.UTF-8

pandas: 0.22.0
pytest: None
pip: 9.0.1
setuptools: 40.6.3
Cython: None
numpy: 1.13.3
scipy: 0.19.1
pyarrow: None
xarray: None
IPython: 5.5.0
sphinx: None
patsy: 0.4.1+dev
dateutil: 2.6.1
pytz: 2018.3
blosc: None
bottleneck: None
tables: 3.4.2
numexpr: 2.6.4
feather: None
matplotlib: 2.1.1
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.6.0
html5lib: 0.999999999
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

I'm using Ubuntu 18.04.1 LTS:

% uname -a
Linux mothership 4.15.0-43-generic #46-Ubuntu SMP Thu Dec 6 14:45:28 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

In #11871 @mattip mentioned that

The c-extension module _multiarray_umath is new to numpy 1.16.0, and may indicate a mix of versions, or version expectations, somewhere. Note this issue is closed, if you believe there is a problem with numpy please open a new issue detailing exactly what you are doing.

How come I get this error even though I have numpy 1.13.3? Thanks for all the help!