Modify all files in directory using Python (Loop)

Question

I am having a lot of troubles to make a loop for my script. Originally the script is modifying one CSV file but I have edited it to loop thrue one folder "CSVtoGD" and modify every csv in this folder. So far with no luck. The modified script is:

from pathlib import Path 
import pandas as pd
dir = r'/users/krzysztofpaszta/CSVtoGD' 
csv_files = [f for f in Path(dir).glob('*.csv')] 



for csv in csv_files: #iterate list
   
    df = pd.read_csv(csv, sep=": \s+", engine='python', names=['dane', 'wartosc'])
    # creating columns with names: ścieżka_do_pliku:czcionka.ttf 
    df['dana_czcionka'] = df['dane'].str.split(':').str[0]

    print('\n--- df ---\n')
    print(df.to_string())

    with open('csv', 'w') as f_out:
        writer = csv.writer(f_out)
    
# sorting data by columns: ścieżka_do_pliku:czcionka.ttf 
        for name, data in df.groupby('dana_czcionka'):
            print('\n---', name, '---\n')
        
            headers = (data['dane'] + ":").to_list()
            print(headers)
    
            values = data['wartosc'].to_list()
            print(values)
            values.insert(0, name) # - Adding name (path) to every row
            values.insert(0, name)
            #writer.writerow(headers) 
            writer.writerow(values)
            
# showing results in terminal, saving to file

    print(f'{csv.name} saved.')

I receive this error:

AttributeError                            Traceback (most recent call last)
/var/folders/zw/12ns4dw96zb34ktc_vfn0zp80000gp/T/ipykernel_49714/1288759270.py in <module>
     16 
     17     with open('csv', 'w') as f_out:
---> 18         writer = csv.writer(f_out)
     19 
     20 # grupowanie danych według kolumn ścieżka_do_pliku:czcionka.ttf

AttributeError: 'PosixPath' object has no attribute 'writer'

I was trying to achevie this by modifying the 'writer' but I guess my knowledge is to little for now. I think an easy loop would get results but I have no idea if my loop is build wrong or just what is the problem..

Original script, without loop is working correctly. Original script (modifying one csv) look like this:

import pandas as pd
import csv
df = pd.read_csv('TTF-Projects-INFO.csv', sep=": \s+", engine='python', names=['dane', 'wartosc'])

# creating columns with names like: ścieżka_do_pliku:czcionka.ttf 
df['dana_czcionka'] = df['dane'].str.split(':').str[0]

print('\n--- df ---\n')
print(df.to_string())

with open('newTTF-Projects-INFO.csv', 'w') as f_out:
    writer = csv.writer(f_out)
    
# sorting data by columns: ścieżka_do_pliku:czcionka.ttf 
    for name, data in df.groupby('dana_czcionka'):
        print('\n---', name, '---\n')
        
        headers = (data['dane'] + ":").to_list()
        print(headers)
    
        values = data['wartosc'].to_list()
        print(values)
        values.insert(0, name) # - add name (path) to every row with data
        #writer.writerow(headers) 
        writer.writerow(values)
            
# showing effect in terminal, saving to file

print('\n--- file ---\n')
print(open('newTTF-Projects-INFO.csv').read())

Answer 1

Your problem is a name clash. The original code uses import csv to import the library. But you also use csv as a variable in your loop

for csv in csv_files: #iterate list

So change your variable name to something like that:

for csv_file in csv_files: #iterate list

and the following occurrences of the variable:

...
for csv_file in csv_files: #iterate list
    df = pd.read_csv(csv_file, sep=": \s+", engine='python', names=['dane', 'wartosc'])
    # creating columns with names: ścieżka_do_pliku:czcionka.ttf 
    df['dana_czcionka'] = df['dane'].str.split(':').str[0]

    print('\n--- df ---\n')
    print(df.to_string())

    with open(csv_file, 'w') as f_out:
        writer = csv.writer(f_out)
...
print(f'{csv_file} saved.')

Finally, add the line import csv at the beginning.
Now your code should work as expected.

---

Alternatively you can use the line

import csv as cs

and change the writer line to

writer = cs.writer(f_out)

In this case you can stay with the csv variable name.
But IMHO this is less clear in relation to naming conventions.