Let’s break down why this error is happening and how to fix it quickly.

The Root Cause #

Your code has a classic timing issue here:

• The csv.reader() creates an iterator that relies on the open file handle to read data.
• When your with block finishes, Python automatically closes the Cars.csv file.
• But requests.request() doesn’t actually process the payload (your csv.reader object) until it sends the POST request—after the file is already closed. That’s why you get the "I/O operation on closed file" error.

Solution 1: Keep the File Open During the Request (Best for Large Files) #

The simplest fix is to move your POST request inside the with block, so the file stays open while requests reads it. Even better—you don’t need csv.reader at all here, because requests can directly handle the open file object for CSV uploads:

import requests

url = "https://paneledgesandbox.//API/v3/surveys/id/import-responses"
headers = {
    'X-API-TOKEN': 'zzzz',
    'Content-Type': 'text/csv',
    'Cookie': 'CPSessID=bb5b5cf55ceff2c2b8237810db3ca3a7; XSRF-TOKEN=XSRF_0wG1WMbV3G0ktBb'
}

# Keep the file open while sending the request
with open('Cars.csv', 'r') as file:
    response = requests.request("POST", url, headers=headers, data=file)

print(response.text)

This works because requests streams the file content directly (without loading the whole thing into memory), which is efficient for large CSVs.

Solution 2: Load CSV Data into Memory (For Small Files) #

If you need to modify the CSV data before sending it, you can read all the content into memory first (so it’s no longer tied to the file handle):

import requests
import csv
import io

url = "https://paneledgesandbox.//API/v3/surveys/id/import-responses"
headers = {
    'X-API-TOKEN': 'zzzz',
    'Content-Type': 'text/csv',
    'Cookie': 'CPSessID=bb5b5cf55ceff2c2b8237810db3ca3a7; XSRF-TOKEN=XSRF_0wG1WMbV3G0ktBb'
}

with open('Cars.csv', 'r') as file:
    # Read all CSV rows into a list
    rows = list(csv.reader(file))
    
    # Convert the list back to a CSV-formatted string
    output_buffer = io.StringIO()
    csv_writer = csv.writer(output_buffer)
    csv_writer.writerows(rows)
    payload = output_buffer.getvalue()

# Now the payload is independent of the file, which is safe to send
response = requests.request("POST", url, headers=headers, data=payload)
print(response.text)

Key Takeaways #

• with blocks do close files automatically, but you must ensure any operation that needs the open file happens inside the block.
• csv.reader is lazy—it only reads the file when you iterate over it. requests does this iteration when sending the request, not when you assign data=payload.
• For file uploads, always prefer passing the open file object directly to requests (Solution 1) unless you need to modify the data first.

内容的提问来源于stack exchange，提问作者Aanurag Das