I am trying to convert all the elements of two pandas series from a pandas data frame, which aren't ascii characters to ascii. Simply apply the function to the relevant columns doesnt work. Python only shows an attribute error stating that 'series' object has no attribute encode.
import pandas as pd
import numpy as np
from unidecode import unidecode
try_data=pd.DataFrame({
'Units': np.array([3,4,5,6,10],dtype='int32'),
'Description_PD': pd.Categorical(['VEIJA 5 TRIÂNGULOS 200','QUEIJO BOLA','QJ BOLA GRD','VEIJO A VACA TRIÂNGULOS 100','HEITE GORDO TERRA']),
'Description_Externa' : pd.Categorical(['SQP 4 porções', 'Bola', ' SIESTA BOLA', 'SQP 16 porções', 'TERRA NOSTRA'])
})
try_data[['Description_PD','Description_Externa']].apply(unidecode)
Iterate over the col list and in the loop call apply, for some reason your attempt didn't work but it should have:
In[47]:
for col in ['Description_PD','Description_Externa']:
try_data[col] = try_data[col].apply(unidecode)
try_data
Out[47]:
Description_Externa Description_PD Units
0 SQP 4 porcoes VEIJA 5 TRIANGULOS 200 3
1 Bola QUEIJO BOLA 4
2 SIESTA BOLA QJ BOLA GRD 5
3 SQP 16 porcoes VEIJO A VACA TRIANGULOS 100 6
4 TERRA NOSTRA HEITE GORDO TERRA 10
For instance calling apply on a single column works fine:
In[49]:
try_data['Description_Externa'].apply(unidecode)
Out[49]:
0 SQP 4 porcoes
1 Bola
2 SIESTA BOLA
3 SQP 16 porcoes
4 TERRA NOSTRA
Name: Description_Externa, dtype: category
Categories (5, object): [SIESTA BOLA, Bola, SQP 16 porcoes, SQP 4 porcoes, TERRA NOSTRA]