pandas：取每一行的差异

Question

我有一个看起来像这样的数据框

                       price    signal
time                    
2017-10-28 13:54:00   293.18    buy
2017-11-02 17:13:04   285.30    sell
2017-11-02 18:40:04   283.24    buy
2017-11-03 08:40:06   291.23    sell
2017-11-03 12:38:06   294.26    buy
2017-11-05 13:33:06   296.52    sell

我想创建一个利润列，填充前一个 buy/sell 货币对的差值

                       price    signal   profit
time                    
2017-10-28 13:54:00   293.18    buy
2017-11-02 17:13:04   285.30    sell     -7.88
2017-11-02 18:40:04   283.24    buy
2017-11-03 08:40:06   291.23    sell      7.99
2017-11-03 12:38:06   294.26    buy
2017-11-05 13:33:06   296.52    sell      2.29

我最好怎么做？
也欢迎使用替代或更有效的方法来计算利润

Answer 1

我可能会这样做：

    df['profit'] = df.price - df.price.shift()

你总是在开仓前平仓吗？

我会建议一个 trade_id 专栏，用于合并交易。

Answer 2

IIUC 使用 cumsum 创建组密钥，然后 diff

df.groupby(df.signal.eq('buy').cumsum())['price'].diff()

pandas：取每一行的差异

pandas: taking the diff of every second row

python

finance

pandas