按日期对我的数据框进行排序（d/m/y + 小时：分钟：秒）

Question

我正在尝试根据日期（d/m/y + 小时: 分钟: 秒）对列的值进行排序。下面我将向您展示给定数据的格式示例：

发起人	价格	日期
XXX	560	13/05/202011:05:35
Glovo 应用程序	250	12/05/2020 13:07:15
Glovo 应用程序	250	13/04/2020 12:09:25

expected output:
if the user selects a date from the 10/04/2020 | 00:00:00 to 15/05/2020 |00:00:00 :
Glovoapp: 500
XXX: 560

if the user selects a date from the 10/04/2020 00:00:00 to 01/05/2020 00:00:00: 
Glovoapp: 250

到目前为止，我能够在没有日期过滤的情况下根据发起者计算价格总和。对我应该做什么有什么建议吗？

 def sum_method(self):
        montant_init = self.data.groupby("Initiateur")["Montant (centimes)"].sum()
        print(montant_init)
        return montant_init

^我就是用这个方法计算的。我希望我已经足够清楚了，谢谢。

试过答案；请指正：

class evaluation():

    def __init__(self, df):
        self.df = df

    # Will receive 'actual' datetime from df, and user defined 'start' and 'stop' datetimes.
    def in_range(actual, start, stop):
        return start <= actual <= stop

    def evaluate(self):
        user_start = input("Enter your start date (dd.mm.yyyy hour:min:second): ")
        user_stop = input("Enter your end date (dd.mm.yyyy hour:min:second): ")

        # creates series of True or False selecting proper rows.
        mask = self.df['Date'].apply(self.in_range, args=(user_start, user_stop))

        # Do the groupby and sum on only those rows.
        montant_init = self.df.loc[mask].groupby("Initiateur")["Montant (centimes)"].sum()
        print(montant_init)

output when printing: self.df.loc[mask]
Empty DataFrame
Columns: [Opération, Initiateur, Montant (centimes), Monnaie, Date, Résultat, Compte marchand, Adresse IP Acheteur, Marque de carte]
Index: []

Answer 1

下面的作品。有两个步骤：

为 select 右边的行制作遮罩
然后只对这些行进行分组和求和

遮罩功能：

# Will receive 'actual' datetime from df, and user defined 'start' and 'stop' datetimes.
def in_range(actual, start, stop):
    return start <= actual <= stop

然后应用蒙版并执行groupby:

# creates series of True or False selecting proper rows.
mask = df['date'].apply(in_range, args=(user_start, user_stop))

# Do the groupby and sum on only those rows.
df2 = df.loc[mask].groupby('Initiator').sum()

请注意，user_start 和 user_stop 应该是用户定义的开始和停止日期时间。

大功告成！

更新：将方法作为 class 的一部分包含在内：

class evaluation():

    def __init__(self, df):
        self.df = df

    # Will receive 'actual' datetime from df, and user defined 'start' and 'stop' datetimes. Add 'self' as arg in method.
    def in_range(self, actual, start, stop):
        return start <= actual <= stop

    def evaluate(self):
        user_start = pd.to_datetime(input("Enter your start date (yyyy.mm.dd hour:min:second): "))
        user_stop = pd.to_datetime(input("Enter your end date (yyyy.mm.dd hour:min:second): "))

        # creates series of True or False selecting proper rows.
        mask = self.df['Date'].apply(self.in_range, args=(user_start, user_stop)) 

        # Do the groupby and sum on only those rows.
        amount_init = self.df.loc[mask].groupby("Initiator")["Price"].sum()
        print(amount_init)

然后实例化一个新的对象class:

import pandas as pd
import dateutil.parser as dtp
import evaluation as eval # this is the class we just made

data = {
    'Initiator': ['XXX', 'Glovoapp', 'Glovoapp'],
    'Price': [560, 250, 250],
    'Date': [dtp.parse('13/05/2020 11:05:35'), dtp.parse('12/05/2020 13:07:15'), dtp.parse('13/04/2020 12:09:25')]
}

df = pd.DataFrame(data)
eval_obj = eval.evaluation(df)
eval_obj.evaluate()

按日期对我的数据框进行排序（d/m/y + 小时：分钟：秒）

Sorting my data frame by date (d/m/y + hour: min: sec)

python

oop

date

operation

pandas