import pandas as pd

import json

import io

def merge_session_cookies(input_files, output_file):

"""

Merges session cookies from multiple files efficiently, handling large datasets.

"""

all_cookies = []

for file in input_files:

try:

with open(file, 'r') as f:

# Assuming each file contains a list of cookie strings, one per line.

cookies = [line.strip() for line in f]

all_cookies.extend(cookies)

except FileNotFoundError:

print(f"Warning: File not found: {file}")

continue # Skip to the next file

# Use pandas for efficient data manipulation, even with large lists

df = pd.DataFrame({'cookie': all_cookies})

# Remove duplicate cookies

df = df.drop_duplicates()

# Save to a new file (e.g., CSV or JSON)

df.to_csv(output_file, index=False) # Or df.to_json(output_file, orient='records')

print(f"Merged cookies saved to {output_file}")

if __name__ == '__main__':

# Example Usage:

input_files = ['cookies1.txt', 'cookies2.txt', 'cookies3.txt'] # Replace with your file names

output_file = 'merged_cookies.csv'

#Create dummy files for testing

with open('cookies1.txt', 'w') as f:

f.write('cookie1\ncookie2\ncookie3\n')

with open('cookies2.txt', 'w') as f:

f.write('cookie2\ncookie4\ncookie5\n')

with open('cookies3.txt', 'w') as f:

f.write('cookie1\ncookie6\ncookie7\n')

merge_session_cookies(input_files, output_file)