import pandas as pd

DATA_FOLDER = '../readonly/final_project_data/'
transactions = pd.read_csv(os.path.join(DATA_FOLDER, 'sales_train.csv.gz'), )

transactions['date'] = pd.to_datetime(transactions['date'])
grouped_sum = transactions[(transactions['date'].dt.year == 2013) & (transactions['date'].dt.month == 9)].groupby('item_id').sum()

def calc(df):
    return df['item_price'] * df['item_cnt_day']

grouped_sum['total_revenue'] = grouped_sum.apply(calc, axis=1)
max_revenue = grouped_sum.sort_values('total_revenue', ascending=False).iloc[0,4]