Question

我有一个数据框，其中包含一个稀疏填充的列。大多数值都是空白的;唯一的其他价值是“买入”和“卖出”。如果最后一个非空白值为“买入”，或者如果最后一个非空白值为“卖出”，则我希望将空白值替换为“long”。我可以在一个循环中轻松地做到这一点，但我想知道是否有一种非循环的方式来完成这个？

Answer 1

您可以使用fillna或combine_first将帮助None创建的df替换为replace和ffill（fillna使用方法ffill - 转发填充NaN和None s）：

np.random.seed(12)
df = pd.DataFrame({'A':np.random.choice(['Buy','Sell', None], 10, p=(.2,.2,.6)),
                   'B':np.random.choice(['Buy','Sell', None], 10, p=(.2,.2,.6)),
                   'C':np.random.choice(['Buy','Sell', None], 10, p=(.2,.2,.6))})

print (df)

      A     B     C
0   Buy  Sell  None
1  None  None   Buy
2  Sell  None   Buy
3  None  None   Buy
4   Buy   Buy  Sell
5  None  None  None
6  None  None  None
7   Buy  None  None
8  None  None  Sell
9   Buy   Buy  None

df = df.fillna(df.replace({'Sell':'short', 'Buy':'long'}).ffill())
#alternative solution
#df = df.combine_first(df.replace({'Sell':'short', 'Buy':'long'}).ffill())
print (df)
       A      B      C
0    Buy   Sell   None
1   long  short    Buy
2   Sell  short    Buy
3  short  short    Buy
4    Buy    Buy   Sell
5   long   long  short
6   long   long  short
7    Buy   long  short
8   long   long   Sell
9    Buy    Buy  short

说明：

print (df.replace({'Sell':'short', 'Buy':'long'}))
       A      B      C
0   long  short   None
1   None   None   long
2  short   None   long
3   None   None   long
4   long   long  short
5   None   None   None
6   None   None   None
7   long   None   None
8   None   None  short
9   long   long   None

print (df.replace({'Sell':'short', 'Buy':'long'}).ffill())
       A      B      C
0   long  short   None
1   long  short   long
2  short  short   long
3  short  short   long
4   long   long  short
5   long   long  short
6   long   long  short
7   long   long  short
8   long   long  short
9   long   long  short

有条件地替换pandas数据帧中的空白值

1 个答案: