满足两个条件时减去前一行

如何解决满足两个条件时减去前一行

我需要帮助创建一个数据框，该数据框减去前一行且满足 3 个条件的行：

如果列 Same Driver 为 FALSE，则结果列间隔应为 0

如果 Same Driver 列为 TRUE 且 Same Trip 列也为 TRUE，则结果应为 0

如果列 Same Driver 为 TRUE 但列 Same Trip 为 FALSE，则应应用df['Interval'] = df['Data']-df['Data'].shift()

import pandas as pd

date = ["04/04/2021","01/04/2021","11/04/2021","15/04/2021","07/04/2021","09/04/2021","09/04/2021"]

date = pd.to_datetime(date,dayfirst=True)
df_date = pd.DataFrame(date,columns=['Data'])

df = pd.DataFrame({
          'Same Driver': [False,False,True,True],'Same Trip': [False,'Desired Interval': [0,4,2,0]
          })

df = pd.concat([df,df_date],axis=1)

df['Interval'] = df['Data']-df['Data'].shift()

您可以在 Desired Interval 列中看到所需的输出

解决方法

让我们尝试类似的事情：

import pandas as pd

date = ["04/04/2021","01/04/2021","11/04/2021","15/04/2021","07/04/2021","09/04/2021","09/04/2021"]

date = pd.to_datetime(date,dayfirst=True)
df_date = pd.DataFrame(date,columns=['Data'])

df = pd.DataFrame({
    'Same Driver': [False,False,True,True],'Same Trip': [False,True]
})

df = pd.concat([df,df_date],axis=1)
g = df.groupby(
    # Delimit Groups by True False Rows
    (df['Same Driver'] & ~df['Same Trip'])
        .cumsum()
        .shift()  # Shift So that last row is in next group
        .fillna(0)
)
# Set Rows equal to size of previous group (-1 to counter the previous shift)
df.loc[g.apply(pd.DataFrame.last_valid_index),'Interval'] = g.size().values - 1
# Fill Na with 0 and convert back to int
df['Interval'] = df['Interval'].fillna(0).astype(int)

print(df.to_string(index=False))

输出：

Same Driver  Same Trip       Data  Interval
      False      False 2021-04-04         0
      False      False 2021-04-01         0
       True       True 2021-04-11         0
       True       True 2021-04-11         0
       True      False 2021-04-15         4
       True       True 2021-04-15         0
      False      False 2021-04-07         0
       True      False 2021-04-09         2
       True       True 2021-04-09         0