Python 比较和减去日期

我正在寻找一种方法来确定某列中的某个时间是否在同一列中另一个日期的7天内

假设这是我的数据帧-

dic = {'firstname':['Rick','Rick','Rick','John','John','John','David',
                    'David','David','Steve','Steve','Steve','Jim','Jim',
                    'Jim'],
       'lastname':['Smith','Smith','Smith','Jones','Jones','Jones',
                   'Wilson','Wilson','Wilson','Johnson','Johnson',
                   'Johnson','Miller','Miller','Miller'],
       'company':['CFA','CFA','CFA','WND','WND','WND','INO','INO','INO',
                  'CHP','CHP','CHP','MCD','MCD','MCD'],
       'faveday':['2020-03-16','2020-03-11','2020-03-25','2020-04-30',
                  '2020-05-22','2020-05-03','2020-01-31','2020-01-13',
                  '2020-01-10','2020-10-22','2020-10-28','2020-10-22',
                  '2020-10-13','2020-10-28','2020-10-20']}
df = pd.DataFrame(dic)
df['faveday'] = pd.to_datetime(df['faveday'])
print(df)

有输出-

   firstname lastname company    faveday
0       Rick    Smith     CFA 2020-03-16
1       Rick    Smith     CFA 2020-03-11
2       Rick    Smith     CFA 2020-03-25
3       John    Jones     WND 2020-04-30
4       John    Jones     WND 2020-05-22
5       John    Jones     WND 2020-05-03
6      David   Wilson     INO 2020-01-31
7      David   Wilson     INO 2020-01-13
8      David   Wilson     INO 2020-01-10
9      Steve  Johnson     CHP 2020-10-22
10     Steve  Johnson     CHP 2020-10-28
11     Steve  Johnson     CHP 2020-10-22
12       Jim   Miller     MCD 2020-10-13
13       Jim   Miller     MCD 2020-10-28
14       Jim   Miller     MCD 2020-10-20

然后我用-

df = df.sort_values(['firstname','lastname','company','faveday'])
print(df)

得到-

   firstname lastname company    faveday
8      David   Wilson     INO 2020-01-10
7      David   Wilson     INO 2020-01-13
6      David   Wilson     INO 2020-01-31
12       Jim   Miller     MCD 2020-10-13
14       Jim   Miller     MCD 2020-10-20
13       Jim   Miller     MCD 2020-10-28
3       John    Jones     WND 2020-04-30
5       John    Jones     WND 2020-05-03
4       John    Jones     WND 2020-05-22
1       Rick    Smith     CFA 2020-03-11
0       Rick    Smith     CFA 2020-03-16
2       Rick    Smith     CFA 2020-03-25
9      Steve  Johnson     CHP 2020-10-22
11     Steve  Johnson     CHP 2020-10-22
10     Steve  Johnson     CHP 2020-10-28

假设我想知道当前的顺序（索引8，然后是7，6，12等等），一个日期是否在另一个日期的7天之内。（因此，索引8和7都将产生true，但索引6不会）

但我也希望按名称对其进行分组。（因此，吉姆·米勒小组的指数12和14是正确的，而吉姆·米勒小组的指数13不是正确的，但史蒂夫·约翰逊小组的指数9、11和10都是正确的）

是否有一种方法可以减去每组中的日期，然后创建一个列，根据它是否在另一天的7天内来判断是真是假

我正在寻找这样的输出-

   firstname lastname company    faveday seven_days
8      David   Wilson     INO 2020-01-10       TRUE
7      David   Wilson     INO 2020-01-13       TRUE
6      David   Wilson     INO 2020-01-31      FALSE
12       Jim   Miller     MCD 2020-10-13       TRUE
14       Jim   Miller     MCD 2020-10-20       TRUE
13       Jim   Miller     MCD 2020-10-28      FALSE
3       John    Jones     WND 2020-04-30       TRUE
5       John    Jones     WND 2020-05-03       TRUE
4       John    Jones     WND 2020-05-22      FALSE
1       Rick    Smith     CFA 2020-03-11       TRUE
0       Rick    Smith     CFA 2020-03-16       TRUE
2       Rick    Smith     CFA 2020-03-25      FALSE
9      Steve  Johnson     CHP 2020-10-22       TRUE
11     Steve  Johnson     CHP 2020-10-22       TRUE
10     Steve  Johnson     CHP 2020-10-28       TRUE

---

让我们尝试使用

numpy

broadcast自定义一个函数

def sefd (x): 
    return np.sum((np.abs(x.values-x.values[:,None])/np.timedelta64(1, 'D'))<=7,axis=1)>=2
s=df.groupby(['firstname', 'lastname', 'company'])['faveday'].transform(sefd)
Out[301]: 
0      True
1      True
2     False
3      True
4     False
5      True
6     False
7      True
8      True
9      True
10     True
11     True
12     True
13    False
14     True
Name: faveday, dtype: bool
df['seven_days']=s

def sefd（x）：
返回np.sum（（np.abs（x.values-x.values[：，None]）/np.timedelta64（1，'D'））=2
s=df.groupby（['firstname'，'lastname'，'company']）['faveday'].transform（sefd）
Out[301]：
0对
1正确
2错误
3正确
4错误
5对
6错误
7正确
8正确
9正确
10对
11对
12对
13错误
14对
姓名：faveday，数据类型：bool
df['seven_days']=s

您可以试试这个

from datetime import timedelta

m = (df.groupby(['firstname','lastname']).
        apply(lambda x: x['faveday'].sub(x['faveday'].shift()).bfill()).
        reset_index(level=[0,1],drop=True))
df['seven_days'] = m.le(timedelta(days=7))

   firstname lastname company    faveday  seven_days
8      David   Wilson     INO 2020-01-10        True
7      David   Wilson     INO 2020-01-13        True
6      David   Wilson     INO 2020-01-31       False
12       Jim   Miller     MCD 2020-10-13        True
14       Jim   Miller     MCD 2020-10-20        True
13       Jim   Miller     MCD 2020-10-28       False
3       John    Jones     WND 2020-04-30        True
5       John    Jones     WND 2020-05-03        True
4       John    Jones     WND 2020-05-22       False
1       Rick    Smith     CFA 2020-03-11        True
0       Rick    Smith     CFA 2020-03-16        True
2       Rick    Smith     CFA 2020-03-25       False
9      Steve  Johnson     CHP 2020-10-22        True
11     Steve  Johnson     CHP 2020-10-22        True
10     Steve  Johnson     CHP 2020-10-28        True

---

这太棒了，谢谢！现在我想删除新的七天专栏中包含False的索引。最有效的方法是什么？@NicholasDelaTorre df=df[s]