CODEWITHSUNDEEP
pythonnumpymean
MerleauT
MerleauT

Reputation: 103

numpy nanmean () 'float' object has no attribute 'dtype' error

I need some help handling the above error. So far my searches didn't return any solution. Code is below. Works for some datasets but raises an error for others.

a = np.array(df_cols)
aver = np.nanmean(a)

    File "…\Continuum\anaconda3\lib\site-packages\numpy\lib\nanfunctions.py", line 916, in nanmean

    avg = _divide_by_count(tot, cnt, out=out)

    File "…\Continuum\anaconda3\lib\site-packages\numpy\lib\nanfunctions.py", line 190, in _divide_by_count

    return a.dtype.type(a / b)

    AttributeError: 'float' object has no attribute 'dtype'

I am using Spyder 3.3.4 Python 3.7.3 64-bit | Qt 5.9.6 | PyQt5 5.9.2 | Windows 10

Thank you for your help.

Upvotes: 1

Views: 9706

Answers (3)

Aadil Rashid
Aadil Rashid

Reputation: 779

This worked for me

 a = np.array(df_cols)
 aver = np.nanmean(a,dtype='float32')

Upvotes: 1

MerleauT
MerleauT

Reputation: 103

I solved this using Pandas mean() after converting my df_cols to series. Pandas mean considers all entries as object and takes care of NaN.

Upvotes: 2

hpaulj
hpaulj

Reputation: 231605

df_cols in the problem cases is probably object dtype. pandas freely uses this dtype, as for strings or None contents.

In [117]: np.nanmean(np.array([1.2,np.nan]))      # a float array                                         
Out[117]: 1.2

In [118]: np.nanmean(np.array([1.2,np.nan], object))   # object dtype array                    
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-118-ca706fd2a20e> in <module>
----> 1 np.nanmean(np.array([1.2,np.nan], object))

/usr/local/lib/python3.6/dist-packages/numpy/lib/nanfunctions.py in nanmean(a, axis, dtype, out, keepdims)
    914     cnt = np.sum(~mask, axis=axis, dtype=np.intp, keepdims=keepdims)
    915     tot = np.sum(arr, axis=axis, dtype=dtype, out=out, keepdims=keepdims)
--> 916     avg = _divide_by_count(tot, cnt, out=out)
    917 
    918     isbad = (cnt == 0)

/usr/local/lib/python3.6/dist-packages/numpy/lib/nanfunctions.py in _divide_by_count(a, b, out)
    188         else:
    189             if out is None:
--> 190                 return a.dtype.type(a / b)
    191             else:
    192                 # This is questionable, but currently a numpy scalar can

AttributeError: 'float' object has no attribute 'dtype'

Upvotes: 0

Related Questions