python - 替换重复值并在列中重复组合

Question

我在数据框中有“状态”列，其值为“Buy”、“Sell”和“NaN”：

No.Row	状态	价格
1	钠	10
2	卖	20
3	买	10
4	钠	5
5	买	6
6	买	30
7	卖	50
8	卖	25
9	买	40
10	买	35
11	钠	10
12	卖	5

我想有第一个“买入”，然后是“卖出”，分别（买入，卖出）。然后是“卖出”（现在）-“买入”（之前）的“减少”列的值。例如，在第 7 行，我们有 50（第 3 卷）-10（第 7 卷）= 40
理想输出：

No.Row	状态	减少
1
2
3	买
4
5
6
7	卖	40
8
9	买
10
11
12	卖	-35

标签： pythondataframenumpy

Answer 1

一种方法是使用一个global跟踪last buy状态的变量。

然后使用减少方法来计算每次到达某个Sell状态时的成本。

global last_buy
last_buy = None 


def reduction(row):
    global last_buy
    if row.state == "Sell" and last_buy is not None: 
        cost = row.price - last_buy
        last_buy = None
        return cost
    
    if row.state == "Buy" and last_buy is None: 
        last_buy = row.price


df["reduction"] = df.apply(reduction, axis=1)

输出：

   state  price  reduction
0   NaN      10        NaN
1   Sell     20        NaN
2    Buy     10        NaN
3   NaN       5        NaN
4    Buy      6        NaN
5    Buy     30        NaN
6   Sell     50       40.0
7   Sell     25        NaN
8    Buy     40        NaN
9    Buy     35        NaN
10  NaN      10        NaN
11  Sell      5      -35.0