python - 无法在 jupyter 笔记本单元中第二次加入 pandas 数据框

Question

我使用 pandas.DataFrame.join 函数向我的数据帧添加两个新的 SMA 列。

第一次工作正常，但如果我再次运行该单元，我会收到以下错误:

ValueError:列重叠但未指定后缀:Index(['SMA_LONG'], dtype='object')

以下是我的代码。

#cell 1
import numpy as np
import pandas as pd

#cell 2
df = pd.DataFrame({
   'close': np.random.uniform(0.1,0.9, 100),
})

#cell 3
SMA_long = 12
SMA_short = 7
sma_long = df['close'].rolling(window=SMA_long, min_periods=SMA_long - 1).mean()
df = df.join(sma_long.to_frame('SMA_LONG'))

sma_short = df['close'].rolling(window=SMA_short, min_periods=SMA_short - 1).mean()
df = df.join(sma_short.to_frame('SMA_SHORT'))

df.tail()

我该如何解决这个问题？

谢谢。

Answer 1

您的错误意味着 df = df.join(sma_long.to_frame('SMA_LONG')) 之前已经存在 SMA_LONG。

<小时/>

对我来说，您的示例数据解决方案工作得很好，但我认为您可以通过分配到新列来简化它:

SMA_long = 12
SMA_short = 7
df['SMA_LONG'] = df['close'].rolling(window=SMA_long, min_periods=SMA_long - 1).mean()
df['SMA_SHORT'] = df['close'].rolling(window=SMA_short, min_periods=SMA_short - 1).mean()

print (df)
       close  SMA_LONG  SMA_SHORT
0   0.649439       NaN        NaN
1   0.332926       NaN        NaN
2   0.492527       NaN        NaN
3   0.500444       NaN        NaN
4   0.583334       NaN        NaN
..       ...       ...        ...
95  0.775169  0.532850   0.577613
96  0.470479  0.524123   0.594733
97  0.237417  0.517054   0.512506
98  0.753701  0.536595   0.554372
99  0.133795  0.526603   0.475595

[100 rows x 3 columns]

或者:

sma_long = df['close'].rolling(window=SMA_long, min_periods=SMA_long - 1).mean()
sma_short = df['close'].rolling(window=SMA_short, min_periods=SMA_short - 1).mean()

df = df.assign(SMA_LONG=sma_long, SMA_SHORT=sma_short)