import csv
import datetime
import numpy as np
import random

# Remove existing 'retail_transactions.csv' file, if any
! rm -f /p/a/t/h retail_transactions.csv

# Set the no of transactions and othet configs
no_of_iterations = 1000000
data = []
csvFile = 'retail_transactions.csv'

# Open a file in write mode
with open(csvFile, 'w', newline='') as f:

   fieldnames = ['orderID', 'customerID', 'productID', 'state', 'paymentMthd', 'totalAmt', 'invoiceTime']
   writer = csv.DictWriter(f, fieldnames=fieldnames)
   writer.writeheader()

   for num in range(no_of_iterations):
     # Create a transaction record with random values
     new_txn = {
     'orderID': num,
     'customerID': random.choice([100, random.randint(1, 100000)]),
     'productID': np.random.randint(10000, size=random.randint(1, 5)).tolist(),
     'state': random.choice(['CA', 'TX', 'FL', 'NY', 'PA', 'OTHERS']),
     'paymentMthd': random.choice(['Credit card', 'Debit card', 'Digital wallet', 'Cash on delivery', 'Cryptocurrency']),
     'totalAmt': round(random.random() * 5000, 2),
     'invoiceTime': datetime.datetime.now().isoformat()
     }

     data.append(new_txn)

  writer.writerows(data)