1. import csv
  3. def split_cookie_data(input_file, output_prefix, chunk_size=10000):
  4. """
  5. Splits a cookie data CSV file into smaller files to reduce memory usage.
  7. Args:
  8. input_file (str): Path to the input CSV file.
  9. output_prefix (str): Prefix for the output file names.
  10. chunk_size (int): Number of rows per output file.
  11. """
  12. try:
  13. with open(input_file, 'r', newline='') as infile:
  14. reader = csv.reader(infile)
  15. header = next(reader, None) # Read the header row
  17. chunk_num = 1
  18. row_count = 0
  19. outfile = None
  20. writer = None
  22. for row in reader:
  23. if row_count % chunk_size == 0:
  24. if outfile:
  25. outfile.close()
  26. output_file = f"{output_prefix}_part{chunk_num}.csv"
  27. outfile = open(output_file, 'w', newline='')
  28. writer = csv.writer(outfile)
  29. if header:
  30. writer.writerow(header)
  31. chunk_num += 1
  32. writer.writerow(row)
  33. row_count += 1
  35. if outfile:
  36. outfile.close()
  38. except FileNotFoundError:
  39. print(f"Error: Input file '{input_file}' not found.")
  40. except Exception as e:
  41. print(f"An error occurred: {e}")
  44. if __name__ == '__main__':
  45. # Example usage:
  46. split_cookie_data('cookies.csv', 'cookie_data', chunk_size=5000)

Add your comment