统计DataFrame中买卖事件对的盈利与亏损次数

问题分析 #

已知包含buy/sell事件的DataFrame，每个buy对应后续一个sell，需统计盈利（sell价>buy价）和亏损（sell价<buy价）的次数。

实现步骤与代码 #

1. 导入依赖并构造示例数据 #

import pandas as pd

# 构造示例DataFrame
data = {
    'event': ['buy', pd.NA, 'sell', 'buy', pd.NA, pd.NA, pd.NA, 'sell'],
    'b': [4, pd.NA, 5, 3, pd.NA, pd.NA, pd.NA, 9]
}
df = pd.DataFrame(data)

2. 提取有效交易点并配对计算 #

# 过滤出仅含交易事件的行
trade_records = df.dropna(subset=['event']).reset_index(drop=True)

# 分离买入价和卖出价（保证顺序一一对应）
buy_prices = trade_records[trade_records['event'] == 'buy']['b'].tolist()
sell_prices = trade_records[trade_records['event'] == 'sell']['b'].tolist()

# 统计盈亏次数
win_count = 0
lost_count = 0
for buy, sell in zip(buy_prices, sell_prices):
    if sell > buy:
        win_count += 1
    elif sell < buy:
        lost_count += 1

# 输出结果
print(f'total wins={win_count}')
print(f'total lost={lost_count}')

运行结果 #

total wins=2
total lost=0

内容的提问来源于stack exchange，提问作者gotiredofcoding