Python Forum

i want to read a csv file into a python list. sometimes, it produces an

Error:
IOError: [Errno 0]

i have searched a lot and tried to use flush() or seek(0,1) functions. i am using the csv built-in module in python. also, i used pandas.read_csv. but always, the list in not complete sometimes. i am working on windows 10 machine, with python2.7 64-bit.
here is my code:

import csv
from io import open

j = 0
k=0
csv_filename = 'G:\work and courses\Medrar\heirarchies\hier1.csv'
directories_lst = []

def fn():
    global j
    global k        
    global directories_lst
    try:
        csvFile = open(csv_filename,'r')      
        directories_lst=[row for row in map(list, csv.reader(csvFile))[1:]]
        csvFile.close()
        for i in directories_lst:
            try:
                print i    
            except:
                k = k+1
    except:
        j = j+1        
        fn()

when that function ran for about 1000 times, it produces like 300 exceptions, at the body of the for loop. everybody say that it is a problem with python io module implementation in windows. i need a workaround to solve that problem.
that all the content of my .csv file. (it is just a sample)

(Mar-24-2018, 05:36 PM)socertis Wrote: [ -> ]everybody say that it is a problem with python io module implementation in windows

Who is everybody? I don't believe that the bug is in the (heavily tested) standard library. There is no mention of unicode in your program. What is the csv file's encoding? are there unicode characters outside the ascii range?

everbody on the internet i mean (stackoverflow and similar websites).
the file is ascii encoded.

Have you tried with python 3?

(Mar-24-2018, 05:36 PM)socertis Wrote: [ -> ]when that function ran for about 1000 times, it produces like 300 exceptions, at the body of the for loop

I find that hard to believe, as the file has been fully read and closed by the time you're at the for loop. How are you running it 1000 times?