Can't run nosetest #3711
Comments
|
do you have |
|
Yes, It seems like the problem has to do with hashtable ERROR: Failure: ImportError (cannot import name hashtable)Traceback (most recent call last): Ran 1 test in 0.003s On Wed, May 29, 2013 at 8:12 PM, Phillip Cloud notifications@github.comwrote:
|
|
oh that traceback is brutal. try to make sure you 1) put in monospaced font (you can still type github flavored markdown from email) and 2) separate it from prose. i only get hashtable import errors when i'm ignoring the linters and i've edited some code and made a syntax error. obligatory question: what version are you running? |
|
The thing is, I'm getting this error without even making any changes to the On Wed, May 29, 2013 at 8:39 PM, Phillip Cloud notifications@github.comwrote:
|
|
on linux, did you build?
where you cloned it? |
|
yes On Wed, May 29, 2013 at 8:50 PM, jreback notifications@github.com wrote:
|
do anything? (if that works), then the following should too make sure you are in the same dir that you built in, call it
do anything? |
|
should be |
|
oh! i think i know what it is: i get this error when i blindly use the command line history to (re)install pandas after a merge/rebase/change and i end up doing |
|
@jeffreyrosenbluth u might want to do something like pip uninstall --yes pandas
while [ $? -eq 0 ]; do
pip uninstall --yes pandas
doneto remove all the versions of pandas that you have installed. you probably shouldn't use this if you aren't in a virtualenv |
|
Python 2.7.3 (v2.7.3:70274d53c1dd, Apr 9 2012, 20:52:43)
On Wed, May 29, 2013 at 8:56 PM, jreback notifications@github.com wrote:
|
|
Is hashtable built? Do find . -name '*hashtable*'in the pandas directory and look for files with the so extension. |
|
yes it's there and buitl On Thu, May 30, 2013 at 9:33 AM, Phillip Cloud notifications@github.comwrote:
|
|
and you're sure you only have one version of pandas installed? |
|
I have a production build installed, but in the directory I'm working I am On Thu, May 30, 2013 at 10:49 AM, Phillip Cloud notifications@github.comwrote:
|
|
the |
|
Sure, I'll set up a virtualenv and give it a go. On Thu, May 30, 2013 at 10:58 AM, Phillip Cloud notifications@github.comwrote:
|
|
Any progress here? |
|
Sorry, I haven't had a chance to try the virualenv idea yet, I've been j. On Tue, Jun 4, 2013 at 11:25 AM, Phillip Cloud notifications@github.comwrote:
|
|
Finally got a chance to try again under a virtual env and had same problem.
======================================================================
ERROR: Failure: ValueError (numpy.dtype has the wrong size, try recompiling)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/Library/Python/2.7/site-packages/nose/loader.py", line 413, in
loadTestsFromName
addr.filename, addr.module)
File "/Library/Python/2.7/site-packages/nose/importer.py", line 47, in
importFromPath
return self.importFromDir(dir_path, fqname)
File "/Library/Python/2.7/site-packages/nose/importer.py", line 94, in
importFromDir
mod = load_module(part_fqname, fh, filename, desc)
File "/Users/jeffreyrosenbluth/pandas_env/pandas/pandas/__init__.py",
line 6, in <module>
from . import hashtable, tslib, lib
File "numpy.pxd", line 157, in init pandas.hashtable
(pandas/hashtable.c:19547)
ValueError: numpy.dtype has the wrong size, try recompiling
----------------------------------------------------------------------
Ran 1 test in 0.001s
FAILED (errors=1) |
|
you have another version numpy installed somewhere (that's why you're getting the dtype size error). looks like ur on a mac, but i can reproduce this on an ubuntu machine if i do mkvirtualenv pandas
sudo apt-get build-dep scipy
workon pandas
pip install cython
pip install numpy
git clone git://github.com/pydata/pandas.git
cd pandas
python setup.py build_ext --inplace
nosetests pandassince it will find the numpy that was installed by ubuntu. The fix here is to add |
|
so if ur using |
|
I started from scratch and tried again as you recommended with On Fri, Jun 7, 2013 at 12:21 PM, Phillip Cloud notifications@github.com wrote:
|
|
By the way, I tried the same procedure on my ubuntu setup and I get On Fri, Jun 7, 2013 at 10:40 PM, Jeffrey Rosenbluth
|
|
ok. i'm going to try the following on an ubuntu 64-bit precise (12.04 i think) # install the build dependencies
mkvirtualenv pandas
workon pandas
pip install cython
pip install numpy
pip install python-dateutil
pip install pytz
git clone git://github.com/pydata/pandas.git
cd pandas
python setup.py build_ext --inplace
nosetests pandas |
|
@jeffreyrosenbluth i can repro your very first error (the one at the top) if i forget to install |
|
I don't see how thats possible if i'm installing numpy fresh in the virtual On Sat, Jun 8, 2013 at 12:42 AM, Phillip Cloud notifications@github.comwrote:
|
|
did u try the above? i don't actually have a separate machine sitting around with ubuntu i actually use something called vagrant which let's you set virtualbox or vmware bare systems (thanks @jreback). i use it mostly for testing. in this case it might be useful to see if you can get it running on that machine...i can walk you through setting one up if you want... |
|
thanks, I have to go out for a few hours. If you around when we get back we On Sat, Jun 8, 2013 at 9:16 AM, Phillip Cloud notifications@github.comwrote:
|
|
resolved? |
|
i'm not sure what else to try here.... |
|
No, I was never able to get it to work. On Wed, Jul 10, 2013 at 9:23 AM, jreback notifications@github.com wrote:
|
|
there must be another version of numpy somewhere. there's no way that i know of to get those when you do |
|
/src/scipy/scipy/io/matlab/numpy_rephrasing.h On Wed, Jul 10, 2013 at 9:46 AM, Phillip Cloud notifications@github.comwrote:
|
|
actually i'm mistaken is all over, see attached On Wed, Jul 10, 2013 at 9:46 AM, Phillip Cloud notifications@github.comwrote:
|
|
i think therein lies the issue, leave what you must but see if you can get rid of some of the extra numpy installs .e.g, don't unintsll the version that comes with the devtools probably |
|
@jeffreyrosenbluth have you gotten to the bottom of this? |
|
haven't had a chance to root out stray numpy instances. thanks for staying on top of it. On Sat, Jul 20, 2013 at 7:44 PM, Phillip Cloud notifications@github.comwrote:
|
|
I hit this error also, not really sure the cause, however, I was able to resolve by develop install of pandas repo. Never did seem to resolve via pip uninstall. If I find out more, I'll make sure and post here! |
Trying to contribute but get the following error when trying to run nosetests with "nosetests pandas". I get the same error on both OSX 10.8.3 and Linux Ubuntu:
ERROR: Failure: ImportError (No module named dateutil.parser)
Traceback (most recent call last):
File "/usr/lib/python2.7/dist-packages/nose/loader.py", line 390, in loadTestsFromName
addr.filename, addr.module)
File "/usr/lib/python2.7/dist-packages/nose/importer.py", line 39, in importFromPath
return self.importFromDir(dir_path, fqname)
File "/usr/lib/python2.7/dist-packages/nose/importer.py", line 86, in importFromDir
mod = load_module(part_fqname, fh, filename, desc)
File "/home/reedrosenbluth/pandas/pandas/init.py", line 6, in
from . import hashtable, tslib, lib
File "tslib.pyx", line 31, in init pandas.tslib (pandas/tslib.c:48137)
ImportError: No module named dateutil.parser
Ran 1 test in 0.001s
The text was updated successfully, but these errors were encountered: