CODEWITHSUNDEEP
python-3.xpandasnumpytypeerror
aks85
aks85

Reputation: 805

python pandas error - ufunc 'subtract' did not contain a loop with signature matching types dtype

I'm trying to create quartile groups of a variable in a new variable. I'm getting an error message and I'm not sure why.

I wrote: 

df.describe().popularity
count    10865.000000
mean         0.646446
std          1.000231
min          0.000065
25%          0.207575
50%          0.383831
75%          0.713857
max         32.985763
Name: popularity, dtype: float64

Then: 

bin_edges = ['0.000065', '0.207575','0.383831','0.713857','32.985763']
bin_names = ['low','mod_low','medium','high']
df['popularity_levels']= pd.cut(df['popularity'], bin_edges, labels=bin_names)
df.head()

I'm getting the following error: 

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-49-b6e8c834de1b> in <module>()
----> 1 df['popularity_levels']= pd.cut(df['popularity'], bin_edges, labels=bin_names)
      2 df.head()

/opt/conda/lib/python3.6/site-packages/pandas/core/reshape/tile.py in cut(x, bins, right, labels, retbins, precision, include_lowest)
    128         bins = np.asarray(bins)
    129         bins = _convert_bin_to_numeric_type(bins, dtype)
--> 130         if (np.diff(bins) < 0).any():
    131             raise ValueError('bins must increase monotonically.')
    132 

/opt/conda/lib/python3.6/site-packages/numpy/lib/function_base.py in diff(a, n, axis)
   1766         return diff(a[slice1]-a[slice2], n-1, axis=axis)
   1767     else:
-> 1768         return a[slice1]-a[slice2]
   1769 
   1770 

TypeError: ufunc 'subtract' did not contain a loop with signature matching types dtype('<U9') dtype('<U9') dtype('<U9')

What does the error mean? I think it may have to do with defining the data type of the new variable as a float... Is that right How can I fix it?

Upvotes: 1

Views: 3038

Answers (1)

Andy Hayden
Andy Hayden

Reputation: 375495

The bin_edges should be floats:

bin_edges = ['0.000065', '0.207575','0.383831','0.713857','32.985763']
# should instead be
bin_edges = [0.000065, 0.207575, 0.383831, 0.713857, 32.985763]

The error occurs since this list is converted to a numpy array: 

In [11]: np.array(['0.000065', '0.207575','0.383831','0.713857','32.985763'])
Out[11]:
array(['0.000065', '0.207575', '0.383831', '0.713857', '32.985763'],
      dtype='<U9')

(Here dtype='<U9' means 9 character unicode.)

In [12]: np.array(['0.000065', '0.207575','0.383831','0.713857','32.985763']) - 1
TypeError: ufunc 'subtract' did not contain a loop with signature matching types dtype('<U9') dtype('<U9') dtype('<U9')

Upvotes: 2

Related Questions