如果条件为true，则在pandas中创建一个新列，并从现有列中分配值

问题描述

我有15分钟的蜡烛数据，并且有一个简短的信号-我想创建一个新的柱止损，如果signal = 0，则止损=下一个蜡烛的第二个高点，即（df ['high'] .shift（-2））

                        open     high      low    close  signal
date                                                           
2020-01-01 09:15:00  1452.50  1457.00  1449.20  1452.50     NaN
2020-01-01 09:30:00  1452.30  1454.40  1450.00  1451.45     NaN
2020-01-01 09:45:00  1450.50  1454.80  1450.00  1453.75     NaN
2020-01-01 10:00:00  1453.70  1453.70  1450.10  1450.70     0.0
2020-01-01 10:15:00  1450.70  1453.00  1450.50  1452.20     NaN
2020-01-01 10:30:00  1452.00  1452.00  1446.75  1446.85     NaN
2020-01-01 10:45:00  1447.60  1449.00  1445.50  1447.10     NaN
2020-01-01 11:00:00  1446.75  1449.00  1446.55  1447.65     NaN

在此示例中： 2020-01-01 10:00:00的短期信号止损将为1452.00
这是2020-01-01 10:30:00的最高点

解决方法

让我们尝试np.where(condition,answer if condition is true,answer if condition is false)

df['stop-loss']=np.where(df.signal==0,df.high.shift(-2),'')

在这种情况下，您没有指定为false时的条件，所以我将其放在''

                        open    high      low    close  signal stop-loss
date                                                                    
2020-01-01 09:15:00  1452.50  1457.0  1449.20  1452.50     NaN          
2020-01-01 09:30:00  1452.30  1454.4  1450.00  1451.45     NaN          
2020-01-01 09:45:00  1450.50  1454.8  1450.00  1453.75     NaN          
2020-01-01 10:00:00  1453.70  1453.7  1450.10  1450.70     0.0    1452.0
2020-01-01 10:15:00  1450.70  1453.0  1450.50  1452.20     NaN          
2020-01-01 10:30:00  1452.00  1452.0  1446.75  1446.85     NaN          
2020-01-01 10:45:00  1447.60  1449.0  1445.50  1447.10     NaN          
2020-01-01 11:00:00  1446.75  1449.0  1446.55  1447.65     NaN          

在评论中回答您的其他问题。假设数据框是

                  open    high      low    close  signal
date                                                          
2020-01-01 09:15:00  1452.50  1457.0  1449.20  1452.50     NaN
2020-01-01 09:30:00  1452.30  1454.4  1450.00  1451.45     NaN
2020-01-01 09:45:00  1450.50  1454.8  1450.00  1453.75     NaN
2020-01-01 10:00:00  1453.70  1453.7  1450.10  1450.70     0.0
2020-01-01 10:15:00  1450.70  1453.0  1450.50  1452.20     NaN
2020-01-01 10:30:00  1452.00  1452.0  1446.75  1446.85     1.0
2020-01-01 10:45:00  1447.60  1449.0  1445.50  1447.10     NaN
2020-01-01 11:00:00  1446.75  1449.0  1446.55  1447.65     NaN

使用np.select([conditons],[choices],alternative)

conditions=[df.signal==0,df.signal==1]
choices=[df.high.shift(-2),df.low.shift(-2)]
df['stop-loss']=np.select(conditions,choices,'')



                   open    high      low    close  signal stop-loss
date                                                                    
2020-01-01 09:15:00  1452.50  1457.0  1449.20  1452.50     NaN          
2020-01-01 09:30:00  1452.30  1454.4  1450.00  1451.45     NaN          
2020-01-01 09:45:00  1450.50  1454.8  1450.00  1453.75     NaN          
2020-01-01 10:00:00  1453.70  1453.7  1450.10  1450.70     0.0    1452.0
2020-01-01 10:15:00  1450.70  1453.0  1450.50  1452.20     NaN          
2020-01-01 10:30:00  1452.00  1452.0  1446.75  1446.85     1.0   1446.55
2020-01-01 10:45:00  1447.60  1449.0  1445.50  1447.10     NaN          
2020-01-01 11:00:00  1446.75  1449.0  1446.55  1447.65     NaN