Python 按过去7周内的日期选择数据集记录

根据下面的数据集，我想筛选所有日期属于最后7周的记录

record_id,date,site,sick,funny,happy
CDEC1947-6,9/2/2018,2,1,1,1
IJKC1953-4,9/29/2018,2,1,1,1
FGHC1724-9,10/25/2018,2,3,1,1
FGHC2929-1,10/31/2018,4,1,1,1
CDEC1912-0,11/1/2018,1,1,1,1
IJKC1726-4,11/2/2018,1,3,1,1
IJKC1728-0,10/26/2018,2,3,1,1
ABCC1730-6,11/2/2018,2,3,1,1
ABCC1731-4,11/2/2018,2,3,1,1
CDEC1733-0,10/22/2018,1,3,1,1
CDEC1735-5,11/2/2018,2,3,1,1
IJKC1914-6,10/27/2018,2,6,1,1
ABCC1916-1,10/23/2018,2,6,1,1
IJKC1918-7,11/2/2018,2,1,1,1
CDEC1920-3,10/24/2018,1,6,1,1
IJKC1943-5,11/2/2018,2,4,1,1
ABCC1945-0,11/2/2018,1,4,1,1
ABCC1949-2,10/25/2018,2,4,1,1
CDEC1951-8,11/2/2018,2,5,1,1
CDEC2924-2,11/3/2018,4,1,1,1
CDEC2927-5,11/3/2018,1,1,1,1
ABCC2925-9,11/4/2018,4,1,1,1
IJKC1941-9,11/4/2018,2,4,1,1
ABCC2922-6,11/5/2018,1,1,1,1

我试过很多把戏，但都没有成功。 其中之一是：

df['data_recrutamento'] = pd.to_datetime(df['data_recrutamento'])

m1 = (df['sick'] == 1) | (df['funny'] == 1) | (df['happy'] == 1)  
m2 = df['date'] >= pd.Timestamp('today') - pd.DateOffset(days=7)  
m3 = ~df['date'].dt.weekday.isin([5, 6])                         

dates_last7_weekdays = df.loc[m1 & m2 & m3, 'site'].value_counts()
dates_last7_weekdays

dates_last7_weekdays = df.loc[m1 & m2 & m3, 'site'].value_counts()
dates_last7_weekdays

其他尝试示例：

import pandas as pd
import numpy as np
from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import *
import plotly.graph_objs as go
import datetime
from datetime import date
from datetime import timedelta
today = date.today()
from IPython.core.interactiveshell import InteractiveShell
%matplotlib inline

df=pd.read_csv("dataset.csv", encoding="utf-8",low_memory=False)

df["date"]=pd.to_datetime(df["date"])
df["site"]=df["site"].astype("category") # Convert to category
df['sick']=df['sick'].astype('category')
df["funny"]=df["funny"].astype("category")
df["happy"]=df["happy"].astype("category")

df = df.sort_values(by='date', ascending='True')
df.head()

    record_id   date    site    sick    funny   happy
0   CDEC1947-6  2018-09-02  2   1   1   1
1   IJKC1953-4  2018-09-29  2   1   1   1
9   CDEC1733-0  2018-10-22  1   3   1   1
12  ABCC1916-1  2018-10-23  2   6   1   1
14  CDEC1920-3  2018-10-24  1   6   1   1
2   FGHC1724-9  2018-10-25  2   3   1   1
17  ABCC1949-2  2018-10-25  2   4   1   1
6   IJKC1728-0  2018-10-26  2   3   1   1
11  IJKC1914-6  2018-10-27  2   6   1   1
3   FGHC2929-1  2018-10-31  4   1   1   1
4   CDEC1912-0  2018-11-01  1   1   1   1
7   ABCC1730-6  2018-11-02  2   3   1   1
10  CDEC1735-5  2018-11-02  2   3   1   1
5   IJKC1726-4  2018-11-02  1   3   1   1
13  IJKC1918-7  2018-11-02  2   1   1   1
15  IJKC1943-5  2018-11-02  2   4   1   1
16  ABCC1945-0  2018-11-02  1   4   1   1
18  CDEC1951-8  2018-11-02  2   5   1   1
8   ABCC1731-4  2018-11-02  2   3   1   1
19  CDEC2924-2  2018-11-03  4   1   1   1
20  CDEC2927-5  2018-11-03  1   1   1   1
22  IJKC1941-9  2018-11-04  2   4   1   1
21  ABCC2925-9  2018-11-04  4   1   1   1
23  ABCC2922-6  2018-11-05  1   1   1   1

days_diff = []
for i in df.loc[:, 'date']:
    days_diff.append(((datetime.datetime.today() - i).days))

final=df[(pd.Series(days_diff) <= 7) & ((df.loc[:, 'sick'] == 1)|(df.loc[:, 'funny'] == 1)|(df.loc[:, 'happy'] == 1) )] 

C:\Users\H\Miniconda3\lib\site-packages\ipykernel_launcher.py:10: UserWarning:

Boolean Series key will be reindexed to match DataFrame index.

len(final)
21

final

    record_id   date    site    sick    funny   happy
9   CDEC1733-0  2018-10-22  1   3   1   1
12  ABCC1916-1  2018-10-23  2   6   1   1
14  CDEC1920-3  2018-10-24  1   6   1   1
17  ABCC1949-2  2018-10-25  2   4   1   1
11  IJKC1914-6  2018-10-27  2   6   1   1
10  CDEC1735-5  2018-11-02  2   3   1   1
13  IJKC1918-7  2018-11-02  2   1   1   1
15  IJKC1943-5  2018-11-02  2   4   1   1
16  ABCC1945-0  2018-11-02  1   4   1   1
18  CDEC1951-8  2018-11-02  2   5   1   1
19  CDEC2924-2  2018-11-03  4   1   1   1
20  CDEC2927-5  2018-11-03  1   1   1   1
22  IJKC1941-9  2018-11-04  2   4   1   1
21  ABCC2925-9  2018-11-04  4   1   1   1
23  ABCC2922-6  2018-11-05  1   1   1   1

---

因为这些日期属于与2018-11-06至2018-11-29的最后7个工作日相对应的日期间隔，所以我在撰写2018-11-06时参考的是今天，但明天应该是2018-11-07。

一个简单的方法是减去并找出日差，并将其用于子集。我们使用datetime.datetime.today来获取今天的日期时间。然后，我们使用该日期时间从df.loc[：，'dates']列中减去每个条目。为了确保我们不会因为天数的不同而有时间，我们在最后使用了…天。然后，我们使用比较小于或等于操作创建一个布尔序列，指示哪些条目小于或等于7天。最后，我们使用这个布尔级数来过滤数据帧

import datetime

days_diff = []
for i in df.loc[:, 'date']:
    days_diff.append(((datetime.datetime.today() - i).days))

#subset your data frame
df[pd.Series(days_diff) <= 7]
#or to include the other conditions as well,
df[(pd.Series(days_diff) <= 7) & ((df.loc[:, 'sick'] == 1)|(df.loc[:, 'funny'] == 1)|(df.loc[:, 'happy'] == 1) )]

---

注意：首先将日期列转换为适当的日期时间

它可以工作，但在我的真实数据集上输出20个不同的日期。也给出了这个警告：C:\Users\H\Miniconda3\lib\site packages\ipykernel\u launcher.py:8:UserWarning:Boolean系列键将被重新索引以匹配数据帧索引。请在代码中添加一些注释，我非常喜欢它，它非常干净，代码行数很少。Thanks@MGB.py，我希望现在清楚了。如果您有任何其他问题，请告诉我谢谢您的评论，但是这个条件对过滤器有什么影响？df['sick']==1 | df['fully']==1 | df['happy']==1抱歉编辑以包括其他条件

import datetime

days_diff = []
for i in df.loc[:, 'date']:
    days_diff.append(((datetime.datetime.today() - i).days))

#subset your data frame
df[pd.Series(days_diff) <= 7]
#or to include the other conditions as well,
df[(pd.Series(days_diff) <= 7) & ((df.loc[:, 'sick'] == 1)|(df.loc[:, 'funny'] == 1)|(df.loc[:, 'happy'] == 1) )]