我正在尝试生成一个4-4-5日历。我有一个从29-03-2020到2022-04-09的日期的数据,由它们的等效周数组成。29-03-2020是财政年度的开始日期。我正试图生成一个列,其中包含它们各自所属的部分。
这是我要找的最后一个df,
a | count | quarter
2020-04-04 | 1 | Q1 2021
2020-04-11 | 2 | Q1 2021
.
.
.
2021-03-27 | 52 | Q4 2021
2021-04-03 | 53 | Q4 2021 #since 2020 is a leap year there are 53 weeks otherwise it will be 52 weeks
2021-04-10 | 1 | Q1 2022
2021-04-17 | 2 | Q1 2022
.
2022-03-02 | 52 | Q4 2022
2022-04-09 | 1 | Q1 2023我做了以下尝试,
import pandas as pd
import numpy as np
year_start = '2020-03-29'
year_end = '2022-04-09'
week_end_sat = pd.DataFrame(pd.date_range(year_start, year_end, freq=f'W-SAT'), columns=['a'])
first_day_of_year = week_end_sat.iloc[0, 0].replace(day=1, month=1)
baseline = pd.DataFrame(pd.date_range(first_day_of_year, periods=len(week_end_sat), freq=f'W-SAT'), columns=['a'])
week_end_sat['count'] = baseline['a'].dt.isocalendar().week
week_end_sat['quarter'] = 'Q' + baseline['a'].dt.quarter.astype(str) + ' ' + (baseline['a'].dt.year+1).astype(str)
week_end_sat['b'] = baseline['a']
all_days_df = pd.DataFrame(pd.date_range(year_start, year_end), columns=['a'])
merge_df = pd.merge(all_days_df,week_end_sat, on='a', how='left')
merge_df['count'] = merge_df['count'].bfill()
merge_df['quarter'] = merge_df['quarter'].bfill()
merge_df['week_day'] = merge_df['a'].dt.day_name()在第52周,也就是Q4 2021之前,我得到的硬币是正确的,但是在那之后,它就搞砸了。在第53周,我得到了Q1 2022等等。应该是Q4 2021年的第53周,因为这是一个闰年?有人能给我指点一下我可以纠正这个问题吗?
编辑
如果我也想为每个日期显示它所属的月份怎么办?
最终的df应该是这样的,
a | count | quarter | month
2020-04-04 | 1 | Q1 2021 | 2020-04
2020-04-11 | 2 | Q1 2021 | 2020-04
.
.
.
2021-03-27 | 52 | Q4 2021 | 2021-03
2021-04-03 | 53 | Q4 2021 | 2021-03
2021-04-10 | 1 | Q1 2022 | 2021-04
2021-04-17 | 2 | Q1 2022 | 2021-04
.
2022-03-02 | 52 | Q4 2022 | 2022-03
2022-04-09 | 1 | Q1 2023 | 2022-04发布于 2021-04-13 01:23:53
假设year_start总是一个财政年度的开始,那么您可以使用一个for循环来完成这个任务:
# Generate a week-to-accounting-month mapping
m = np.roll(np.arange(1, 13, dtype='int'), -3)
w = np.tile([4,4,5], 4)
acct_month = {
index + 1: month
for index, month in enumerate(np.repeat(m, w))
}
acct_month[53] = 3 # week 53, if exists, always belong to month 3
#
data = []
y, week = year_start.year, 1
month = year_start.replace(month=4, day=1)
# A fiscal year has 53 weeks if it starts on a leap year. 52 weeks otherwise
num_week = lambda: 53 if (y % 4 == 0 and y % 100 != 0) or (y % 400 == 0) else 52
for date in pd.date_range(year_start, year_end, freq='7D'):
data.append((date + pd.Timedelta(days=6), y+1, week, month))
week += 1
if week > num_week():
y += 1
week = 1
if acct_month[week] != month.month:
month += pd.offsets.MonthBegin(1)
df = pd.DataFrame(data, columns=['week_end', 'fy', 'week_no', 'acct_month'])
# Q1, Q2 and Q3 always have 13 weeks
# Q4 may have 13 or 14 weeks
q = np.ceil(df['week_no'].div(13).clip(upper=4)).astype('int')
df['quarter'] = 'Q' + q.astype(str) + ' ' + df['fy'].astype(str)结果:
week_end fy week_no acct_month quarter
0 2020-04-04 2021 1 2020-04-01 Q1 2021
1 2020-04-11 2021 2 2020-04-01 Q1 2021
2 2020-04-18 2021 3 2020-04-01 Q1 2021
3 2020-04-25 2021 4 2020-04-01 Q1 2021
...
48 2021-03-06 2021 49 2021-03-01 Q4 2021
49 2021-03-13 2021 50 2021-03-01 Q4 2021
50 2021-03-20 2021 51 2021-03-01 Q4 2021
51 2021-03-27 2021 52 2021-03-01 Q4 2021
52 2021-04-03 2021 53 2021-03-01 Q4 2021
53 2021-04-10 2022 1 2021-04-01 Q1 2022
54 2021-04-17 2022 2 2021-04-01 Q1 2022
55 2021-04-24 2022 3 2021-04-01 Q1 2022
...
100 2022-03-05 2022 48 2022-03-01 Q4 2022
101 2022-03-12 2022 49 2022-03-01 Q4 2022
102 2022-03-19 2022 50 2022-03-01 Q4 2022
103 2022-03-26 2022 51 2022-03-01 Q4 2022
104 2022-04-02 2022 52 2022-03-01 Q4 2022
105 2022-04-09 2023 1 2022-04-01 Q1 2023https://stackoverflow.com/questions/67062561
复制相似问题