python合并每五分钟数据_Python时间序列：将字典中的每日数据合并为每周d

如果我很理解您的问题，我认为您可以使用datetime和{}从{}模块中解决它，如下例：from datetime import datetime, timedelta

def get_days_of_week(year, week=1):

# number of the days

days = {'Monday': 1, 'Tuesday': 2, 'Wednesday': 3,

'Thursday': 4, 'Friday': 5, 'Saturday': 6, 'Sunday': 7}

# construct the datetime object with the year and the desired week

a = datetime.strptime('{0}'.format(year), '%Y') + timedelta(days=7*(week-1))

# Every week should start by Sunday .. So escaping days untill the first Sunday

a += timedelta(days=7-days.get(a.strftime('%A'), 0))

for k in range(0, 7):

yield (a + timedelta(days=k)).strftime('%Y%m%d')

days = list(get_days_of_week(2016, week=1))

print('2016 / week = 1:', days)

days = list(get_days_of_week(2016, week=22))

print('2016 / week = 22:', days)

输出：

^{pr2}$

编辑：

根据您上次的编辑，此代码可以满足您的需要：from datetime import datetime, timedelta

def get_days_of_week(data):

# number of the days

days = {'Monday': 1, 'Tuesday': 2, 'Wednesday': 3,

'Thursday': 4, 'Friday': 5, 'Saturday': 6, 'Sunday': 7}

date = datetime.strptime('{}'.format(data), '%Y%m%d')

# get week number

week = int(date.strftime('%U'))

# get year

year = date.strftime('%Y')

# construct the datetime object with the year and the desired week

a = datetime.strptime(year, '%Y') + timedelta(days=7*week)

# Every week should start by Synday .. So escaping days untill the first Sunday

a += timedelta(days=7-days.get(a.strftime('%A'), 0))

return {int(str(data)[:-2]): [int((a + timedelta(days=k)).strftime('%Y%m%d')) for k in range(0, 7)]}

week_dict = {}

week_dict.update(get_days_of_week(20160101))

week_dict.update(get_days_of_week(20160623))

print(week_dict[201601])

print(week_dict[201606])

print(week_dict)

输出：[20160103, 20160104, 20160105, 20160106, 20160107, 20160108, 20160109]

[20160626, 20160627, 20160628, 20160629, 20160630, 20160701, 20160702]

{ 201601: [ 20160103,

20160104,

20160105,

20160106,

20160107,

20160108,

20160109],

201606: [ 20160626,

20160627,

20160628,

20160629,

20160630,

20160701,

20160702]}