object is not converting to numeric after pd.to_numeric

Question

I dont understand why my column allHoldingsFund['ratioBest'] is not converting into numeric?

Initially i replaced empty rows with the number 0 but int 0 and floats numbers did not make the column allHoldingsFund['ratioBest'] easily convertible to numeric (got error TypeError: len() of unsized object). Now I left empty rows as they are and forced the floats to convert to numeric but it did not work- allHoldingsFund['ratioBest'] is is still an object. See below

allHoldingsFund['ratioBest']

Out[170]: 
100                          
144                          
187         0.520704845814978
264                          
386                          
494                          
597                          
635         2.790492957746479

allHoldingsFund['ratioBest']=pd.to_numeric(allHoldingsFund['ratioBest'], errors='coerce')

Name: ratioBest, Length: 1664, dtype: object

Any help is appreciated

Thanks

yes data is confidential .what do you have in mind? pd.__version__ Out[171]: '0.22.0' — SBad, Jul 16 '18 at 13:07
It's hard to replicate your problem without the data. In a mini example of mine your code works. — Wald, Jul 16 '18 at 13:15

jezrael · Accepted Answer · 2018-07-16T14:02:43.403

1

I get your problem - values are not scalar, but numpy.ndarrays:

allHoldingsFund = pd.DataFrame({'ratioBest':[np.array(4.12), '']})
print (allHoldingsFund)
        ratioBest
0      4.12
1

Solution is convert to list and then to 1d array:

arr = np.array(allHoldingsFund['ratioBest'].tolist())
allHoldingsFund['ratioBest'] = (pd.to_numeric(arr, errors='coerce'))
print (allHoldingsFund)
   ratioBest
0       4.12
1        NaN

print (allHoldingsFund['ratioBest'].dtype)
float64

edited Jul 16 '18 at 14:02

answered Jul 16 '18 at 13:22

jezrael

822,522
95
1,334
1,252

thanks for continuous help on this. i get this as error: TypeError: don't know how to convert scalar number to float – SBad Jul 16 '18 at 13:28
1

@SBad - I find problem and added solution. – jezrael Jul 16 '18 at 14:02

object is not converting to numeric after pd.to_numeric

1 Answers1