Andrey Dugin

from pathlib import Path

max_id = max(int(path.stem) for path in Path('path_to_folder').glob('*.txt'))

'CWYCWKAY'.replace('W', 'H').replace('Y', 'N').replace('N', 'Y', 1)

'CWYCWKAY'.translate(str.maketrans({'W': 'H', 'Y': 'N'})).replace('N', 'Y', 1)

'CWYCWKAY'.translate({87: 'H', 89: 'N'}).replace('N', 'Y', 1)

def replace(word='CWYCWKAY'):
    return 'CHYCHKAN'

import pandas as pd
from random choices, randint
from string import ascii_uppercase

items = [''.join(choices(ascii_uppercase, k=randint(4, 11))) for i in range(28000)]

def generate():
    for receipt_id in range(3700000):
        if receipt_id % 100000 == 0:
            print(receipt_id)
        for item in sample(items, randint(2, 15)):
            yield receipt_id, item
            
data = pd.DataFrame(generate(), columns=['receipt_id', 'item_name'])

from collections import Counter
from itertools import combinations
# from sys import intern

# В эмуляции все строки в DataFrame уже интернированы по умолчанию
# data['item_name'] = data.item_name.map(intern) 

statistics = Counter(
    pair_of_items  # Для экономии памяти можно взять hash(pair_of_items)
        for items_in_receipt in data.groupby('receipt_id', sort=False).item_name.agg(sorted)
            for pair_of_items in combinations(items_in_receipt, 2)  # сортировка сохраняется
)