python - Pandas sum column with scalar results in zeros -
i summing column of pandas dataframe scalar , result column full of zeros. weird thing return zeros , won't. if tell me wouldn't believe it, that's why made video showing result:
https://dl.dropboxusercontent.com/u/15853805/pandas%20bug.mp4
can explain me black magic? loosing trust on pandas , need work.
ok code can similar error:
import pandas pd import numpy np pdb import set_trace pdb  _ in xrange(100):     data = np.random.randint(1,100000,1000000)     df = pd.dataframe(data,columns=['column'])     scalar = np.random.randint(1,100000)     df.column += scalar     if df.column.max()==0:         pdb()   my data variable gets full zeros. memory issues? if data smaller doesn't happen, need process big data in safe way!
>> print pandas.__version__ 0.17.1 >> print numpy.__version__ 1.10.1  python 2.7.11 |anaconda 2.4.1 (64-bit)      
the problem version of numexpr=2.4.4 updating numexpr=2.4.6 fixes problem. github issue: https://github.com/pydata/pandas/issues/12167
Comments
Post a Comment