具有条件的新列的操作。

Question

我正在尝试根据另一列的条件创建一个新列。例如:我想根据事件计算一个“折扣价格”列。如果‘诗’申请10%的折扣，如果‘剧院’申请15%的折扣，否则，返回正常的价格。

# importing pandas as pd 
import pandas as pd 

# Creating the DataFrame 
df = pd.DataFrame({'Date':['10/2/2011', '11/2/2011', '12/2/2011', '13/2/2011'], 
                    'Event':['Music', 'Poetry', 'Theater', 'Comedy'], 
                    'Cost':[10000, 5000, 15000, 2000]}) 

# Create a new column 'Discounted_Price' after applying 
# 10% discount on the existing 'Cost' column. 

df['Discounted_Price'] = df['Cost'] - (0.1 * df['Cost']) 

在我看来，这就像是(忽略语法)：

df['Discounted_Price'] = if(df['Event']== 'Poetry') then df['Cost']*0.9,
                         elif (df['Event']== 'Theater') then df['Cost']*0.85

诸若此类。

对如何做这个手术有什么建议吗？

Answer 1

IIUC，您可以使用：

df：

       Date     Event   Cost
0   10/2/2011   Music   10000
1   11/2/2011   Poetry  5000
2   12/2/2011   Theater 15000
3   13/2/2011   Comedy  2000

您还可以使用np.where在一行中满足所有需求。

df['Discounted_Price'] = np.where(df.Event == 'Poetry', df['Cost']*0.9, 
                       np.where(df.Event == 'Theater', df['Cost']*0.85, df['Cost']))

现在的产出是：

       Date     Event   Cost    Discounted_Price
0   10/2/2011   Music   10000   10000.0
1   11/2/2011   Poetry  5000    4500.0
2   12/2/2011   Theater 15000   12750.0
3   13/2/2011   Comedy  2000    2000.0

因此，它满足了else的要求，而不是获得NaNs，而是得到了所有的成本

df.loc[df['Event']=='Poetry','Discounted_Price']=df['Cost']*0.9
df.loc[df['Event']=='Theater','Discounted_Price']=df['Cost']*0.85

输出：

        Date    Event   Cost    Discounted_Price
0   10/2/2011   Music   10000   NaN
1   11/2/2011   Poetry  5000    4500.0
2   12/2/2011   Theater 15000   12750.0
3   13/2/2011   Comedy  2000    NaN

编辑

您还可以使用select(如@sammmywammy建议的那样)来解决手头的问题。我设计了以下语句，可以帮助您解决多列上的多个条件。

conditions = [
    (df['Event'] == 'Music'),
    (df['Event'] == 'Theater'),
    (df['Event'] == 'Poetry'),
    (df['Event'] == 'Comedy')]
choices = [(df['Cost']*0.9),(df['Cost']*0.85),(df['Cost']*0.7), (df['Cost']*0.5)]

df['Discounted'] = np.select(conditions, choices, default='null')

对于多列的单次计算，上述查询转换为

conditions = [
    (df['Event'] == 'Music'),
    ((df['Event'] == 'Theater') | (df['Event'] == 'Comedy')),
    (df['Event'] == 'Poetry')]

choices = [(df['Cost']*0.9),(df['Cost']*0.85),(df['Cost']*0.7), ]

df['Discounted'] = np.select(conditions, choices, default='null')