我想要一個簡短的數據框：每分鐘安排一次添加缺失分鐘數的行（從 09:05 到 09:20）然后每 5 分鐘重新采樣一次time cars flow 0 9:07 737 In 1 9:06 22 Out 2 9:18 42 In 3 9:19 36 Unknown我嘗試過的：data = {'time': ["9:07", "9:06", "9:18", "9:19"], 'cars' : [737, 22, 42,36],'flow': ["In","Out","In","Unknown"]}df = pd.DataFrame(data)idx = pd.date_range("9:05", "09:20", freq="1min")idx = idx.rename('time')df = df.set_index('time')df.index = pd.to_datetime(df.index)df = df.reindex(idx, fill_value=0)df = df.groupby('flow').resample('5T')['cars'].sum() # how_many_volumeprint(df)它返回：flow time 0 2020-10-21 09:05:00 0 2020-10-21 09:10:00 0 2020-10-21 09:15:00 0 2020-10-21 09:20:00 0In 2020-10-21 09:05:00 737 2020-10-21 09:10:00 0 2020-10-21 09:15:00 42Out 2020-10-21 09:05:00 22Unknown 2020-10-21 09:15:00 36但想要的是：In 2020-10-21 09:05:00 737 2020-10-21 09:10:00 0 2020-10-21 09:15:00 42 2020-10-21 09:20:00 0Out 2020-10-21 09:05:00 22 2020-10-21 09:10:00 0 2020-10-21 09:15:00 0 2020-10-21 09:20:00 0Unknown 2020-10-21 09:05:00 0 2020-10-21 09:10:00 0 2020-10-21 09:15:00 36 2020-10-21 09:20:00 0有什么方法可以實現呢？
1 回答
撒科打諢
TA貢獻1934條經驗 獲得超2個贊
df2 = df.pivot(columns='flow').resample('5T').sum().stack()
df2.index = df2.index.swaplevel(0, 1)
>>> df2.sort_index()
cars
flow time
In 2020-10-20 09:05:00 737.0
2020-10-20 09:10:00 0.0
2020-10-20 09:15:00 42.0
Out 2020-10-20 09:05:00 22.0
2020-10-20 09:10:00 0.0
2020-10-20 09:15:00 0.0
Unknown 2020-10-20 09:05:00 0.0
2020-10-20 09:10:00 0.0
2020-10-20 09:15:00 36.0
我個人會將數據保留為表格形式：
>>> df.pivot(columns='flow').resample('5T').sum()
cars
flow In Out Unknown
time
2020-10-20 09:05:00 737.0 22.0 0.0
2020-10-20 09:10:00 0.0 0.0 0.0
2020-10-20 09:15:00 42.0 0.0 36.0
添加回答
舉報
0/150
提交
取消