初学者从 CSV 中删除列（无熊猫）

Question

I've just started coding, I'm trying to remove certain columns from a CSV for a project, we aren't supposed to use pandas. For instance, one of the fields I have to delete is called DwTm , but there's about 15 columns I have to get rid of; I only want the first few, Here's what I've gotten:

import csv
FTemp = "D:/tempfile.csv"
FOut = "D:/NewFile.csv"


with open(FTemp, 'r') as csv_file:
    csv_reader = csv.reader(csv_file)

    with open(FOut, 'w') as new_file:
        fieldnames = ['Stn_Name', 'Lat', 'Long', 'Prov', 'Tm']
        csv_writer = csv.DictWriter(new_file, fieldnames=fieldnames)

        for line in csv_reader:
            del line['DwTm']
            csv_writer.writerow(line)

When I run this, I get the error

del line['DwTm']
TypeError: list indices must be integers or slices, not str

This is the only method I've found to almost work without using pandas. Any ideas?

Answer 1

The easiest way around this is to use a DictReader to read the file. Like DictWriter, which you are using to write the file, DictReader uses dictionaries for rows, so your approach of deleting keys from the old row then writing to the new file will work as you expect.

import csv
FTemp = "D:/tempfile.csv"
FOut = "D:/NewFile.csv"


with open(FTemp, 'r') as csv_file:

    # Adjust the list to be have the correct order
    old_fieldnames = ['Stn_Name', 'Lat', 'Long', 'Prov', 'Tm', 'DwTm']
    csv_reader = csv.DictReader(csv_file, fieldnames=old_fieldnames)

    with open(FOut, 'w') as new_file:
        fieldnames = ['Stn_Name', 'Lat', 'Long', 'Prov', 'Tm']
        csv_writer = csv.DictWriter(new_file, fieldnames=fieldnames)

        for line in csv_reader:
            del line['DwTm']
            csv_writer.writerow(line)

Answer 2

Below

import csv

# We only want to read the 'department' field 
# We are not interested in 'name' and 'birthday month'

# Make sure the list items are in ascending order
NON_INTERESTING_FIELDS_IDX = [2,0]
rows = []
with open('example.csv') as csv_file:
    csv_reader = csv.reader(csv_file, delimiter=',')
    for row in csv_reader:
        for idx in NON_INTERESTING_FIELDS_IDX:
          del row[idx]
        rows.append(','.join(row))
with open('example_out.csv','w') as out:
  for row in rows:
    out.write(row + '\n')

example.csv

name,department,birthday month
John Smith,Accounting,November
Erica Meyers,IT,March

example_out.csv

department
Accounting
IT

Answer 3

It's possible to simultaneously open the file to read from and the file to write to. Let's say you know the indices of the columns you want to keep, say, 0,2, and 4:

good_cols = (0,2,4)
with open(Ftemp, 'r') as fin, open(Fout, 'w') as fout:
    for line in fin:
        line = line.rstrip()        #clean up newlines
        temp = line.split(',')      #make a list from the line
        data = [temp[x] for x in range(len(temp)) if x in good_cols]
        fout.write(','.join(data) + '\n')

The list comprehension (data) pulls only the columns you want to keep out of each row and immediately writes line-by-line to your new file, using the join method (plus tacking on an endline for each new row).

If you only know the names of the fields you want to keep/remove it's a bit more involved, you have to extract the indices from the first line of the csv file, but it's not much more difficult.