计算 Pandas 的亏损
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/48344727/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Calculating Drawdown in Pandas
提问by user8491020
I have the following data frame:
我有以下数据框：
Profit Cumulative
Date
1/6/2005 248.8500 248.85
1/12/2005 48.3500 297.20
1/13/2005 29.2900 326.49
1/13/2005 -500.0000 -173.51
1/13/2005 -250.9500 -424.46
1/14/2005 126.6600 -297.80
1/16/2005 58.7400 -239.06
1/19/2005 68.3500 -170.71
1/21/2005 140.0000 -30.71
1/23/2005 200.0000 169.29
1/26/2005 -250.6800 -81.39
1/27/2005 162.5000 81.11
1/27/2005 135.5100 216.62
1/27/2005 -650.0000 -433.38
1/28/2005 96.8800 -336.50
1/28/2005 -1000.0000 -1336.50
1/31/2005 140.0000 -1196.50
2/1/2005 140.0000 -1056.50
The first column are the dollar profits of my portfolio. I have calculated the second column with the following formula:
第一列是我的投资组合的美元利润。我用以下公式计算了第二列：
df['Cumulative'] = df.Profit.cumsum().round(2)
So, is there a formula where I can calculate the dollar (not %) drawdons of my portfolio? The column should look like this:
那么，是否有一个公式可以计算我的投资组合的美元（而不是 %）收益？该列应如下所示：
Drawdonwn
0.00
0.00
0.00
-500.00
-750.95
-624.29
-565.55
-497.20
-357.20
-157.20
-407.88
-245.38
-109.87
-759.87
-662.99
-1,662.99
-1,522.99
-1,382.99
-1,382.99
回答by user8491020
df['Cumulative'] = df.Profit.cumsum().round(2)
df['HighValue'] = df['Cumulative'].cummax()
df['Drawdown'] = df['Cumulative'] - df['HighVal']
That is the simplest solution I can found.
这是我能找到的最简单的解决方案。
回答by pault
From what I remember, a drawdown is the amount by which your portfolio profit is less than the high. (In the future, you should explicitly define your desired output and don't assume people know what you are asking.)
据我所知，回撤是指您的投资组合利润低于高点的金额。（将来，您应该明确定义您想要的输出，不要假设人们知道您在问什么。）
There may be better ways, but you can calculate this in pandasusing a itertuples():
可能有更好的方法，但您可以pandas使用以下方法计算itertuples()：
import pandas as pd
from StringIO import StringIO
# read the data
df = pd.DataFrame.from_csv(StringIO("""Date Profit Cumulative
1/6/2005 248.8500 248.85
1/12/2005 48.3500 297.20
1/13/2005 29.2900 326.49
1/13/2005 -500.0000 -173.51
1/13/2005 -250.9500 -424.46
1/14/2005 126.6600 -297.80
1/16/2005 58.7400 -239.06
1/19/2005 68.3500 -170.71
1/21/2005 140.0000 -30.71
1/23/2005 200.0000 169.29
1/26/2005 -250.6800 -81.39
1/27/2005 162.5000 81.11
1/27/2005 135.5100 216.62
1/27/2005 -650.0000 -433.38
1/28/2005 96.8800 -336.50
1/28/2005 -1000.0000 -1336.50
1/31/2005 140.0000 -1196.50
2/1/2005 140.0000 -1056.50"""), sep="\s+").reset_index()
# calculate drawdown
prev_high = 0
for i, date, profit, cumulative in df.itertuples():
prev_high = max(prev_high, cumulative)
dd = cumulative - prev_high
df.loc[i, 'Drawdown'] = dd if dd < 0 else 0
The resultant dataframe:
结果数据框：
>>> print(df)
Date Profit Cumulative Drawdown
0 2005-01-06 248.85 248.85 0.00
1 2005-01-12 48.35 297.20 0.00
2 2005-01-13 29.29 326.49 0.00
3 2005-01-13 -500.00 -173.51 -500.00
4 2005-01-13 -250.95 -424.46 -750.95
5 2005-01-14 126.66 -297.80 -624.29
6 2005-01-16 58.74 -239.06 -565.55
7 2005-01-19 68.35 -170.71 -497.20
8 2005-01-21 140.00 -30.71 -357.20
9 2005-01-23 200.00 169.29 -157.20
10 2005-01-26 -250.68 -81.39 -407.88
11 2005-01-27 162.50 81.11 -245.38
12 2005-01-27 135.51 216.62 -109.87
13 2005-01-27 -650.00 -433.38 -759.87
14 2005-01-28 96.88 -336.50 -662.99
15 2005-01-28 -1000.00 -1336.50 -1662.99
16 2005-01-31 140.00 -1196.50 -1522.99
17 2005-02-01 140.00 -1056.50 -1382.99
Investopedia definition of drawdown.
Investopedia 对缩编的定义。
