在熊猫数据帧切片上相乘

Question

I have a dataframe like this:

  Year       A_annex    Arnston     Bachelor         Berg   
  1955      1.625       0.940         NaN            NaN   
  1956      1.219       1.018         NaN            NaN   
  1957      2.090       1.20          NaN            1.190   
  1958      0.950       1.345         NaN            1.090

and I want to multiply everything in [1:,1:] by .404

The code I am trying is:

df=pd.read_csv(r'H:\Sheyenne\Grazing Records\Master_AMU_complete.csv')
hectare=0.404
df=df.iloc[1:,1:]
df=df*hectare

but this returns:

TypeError: Could not operate 0.404686 with block values can't multiply sequence by non-int of type 'float'

printing df.info() says everything after the slice is a non-null object if that helps.

Answer 1

Yes, it is problematic value. You can find these problematic values by function (thanks ajcr):

df = df.convert_objects(convert_numeric=True) 

First NaN are converted to 0 , then apply function above and it return NaN instead of problematic values. So you have to find rows with NaN values and return subset of original df .

print df
   Year  A_annex Arnston  Bachelor  Berg
0  1955    1.625   0.940       NaN   NaN
1  1956    1.219   1.018       NaN   NaN
2  1957    2.090   1.20a       NaN  1.19
3  1958    0.950  1.345a       NaN  1.09

test = df.fillna(0)
test = test.convert_objects(convert_numeric=True)

   Year  A_annex  Arnston  Bachelor  Berg
0  1955    1.625    0.940         0  0.00
1  1956    1.219    1.018         0  0.00
2  1957    2.090      NaN         0  1.19
3  1958    0.950      NaN         0  1.09

test = df[test.isnull().any(axis=1)]

   Year  A_annex Arnston  Bachelor  Berg
2  1957     2.09   1.20a       NaN  1.19
3  1958     0.95  1.345a       NaN  1.09

hectare=0.404
df=df.iloc[1:,1:]
df=df*hectare
print df