CODEWITHSUNDEEP
python
hmhsing
hmhsing

Reputation: 1

AttributeError: '_io.TextIOWrapper' object has no attribute 'decode'

I'm trying read multiple text files, doing word segmentation (use jieba) and then save the results to CSV files respectively. It shows 

AttributeError: '_io.TextIOWrapper' object has no attribute 'decode'

Thanks for anyone's help.

The python code is:

import jieba
import csv
import glob

list_of_files = glob.glob('C:/Users/user/Desktop/speech./*.txt')           
for file_name in list_of_files:

    FI = open(file_name, 'r')
    FO = open(file_name, 'w')
    seglist = jieba.cut(FI, cut_all=False)
    w = csv.writer(FO)  
    w.writerows(seglist)  
    FI.close()
    FO.close()

Upvotes: 0

Views: 4154

Answers (2)

Yoav Glazner
Yoav Glazner

Reputation: 8066

It seems that you need to send bytes to cut and not a file object

try this code instead:

list_of_files = glob.glob('C:/Users/user/Desktop/speech./*.txt')           
for file_name in list_of_files:
    with  open(file_name, 'rb') as f:
        text = f.read()


    seglist = jieba.cut(text, cut_all=False)

    with open(file_name, 'w') as f:
        w = csv.writer(f)  
        w.writerows(seglist)

Upvotes: 1

Arount
Arount

Reputation: 10401

From what I read from source code as example and jieba.cut definition it seems jieba.cut needs string as parameter.

But you are giving an instance of file.

seglist = jieba.cut(FI.read(), cut_all=False)

Fixed the issue from what I saw. (FI.read() is the fix).

Btw, do not call variables like FI / FO, this is valid name for constants or maybe classes, but not variable.

Explicit is better than implicit: prefere something like: file_output & file_input.

Upvotes: 0

Related Questions