我有这样的DataFrame
事情:
data = [['2019-01-01', .1, .2],
['2019-01-02', .5, .3],
['2019-01-03', .2, .4]]
df = pd.DataFrame(data, columns=['date', 'hour01', 'hour02'])
date hour01 hour02
0 2019-01-01 0.1 0.2
1 2019-01-02 0.5 0.3
2 2019-01-03 0.2 0.4
如何融化它,使我有15分钟的正确间隔?像这样:
timestamp value
0 2019-01-01 00:00:00 0.1
1 2019-01-01 00:15:00 0.1
2 2019-01-01 00:30:00 0.1
3 2019-01-01 00:45:00 0.1
4 2019-01-01 01:00:00 0.2
5 2019-01-01 01:15:00 0.2
6 2019-01-01 01:30:00 0.2
7 2019-01-01 01:45:00 0.2
...
16 2019-01-03 00:00:00 0.2
17 2019-01-03 00:15:00 0.2
18 2019-01-03 00:30:00 0.2
19 2019-01-03 00:45:00 0.2
20 2019-01-03 01:00:00 0.4
21 2019-01-03 01:15:00 0.4
22 2019-01-03 01:30:00 0.4
23 2019-01-03 01:45:00 0.4
df.melt(id_vars=['timestamp'], value_vars=['hour_{}'.format(str(x).zfill(2)) for x in range(1, 24)])
给我这个:
27 2017-01-28 hour_01 34.90
28 2017-01-29 hour_01 36.04
29 2017-01-30 hour_01 36.51
... ... ...
16760 2018-12-02 hour_23 51.50
16761 2018-12-03 hour_23 54.00
16762 2018-12-04 hour_23 53.87
然后去哪儿?
解决方案基础上melt
,set_index
和ffill
:
df = df.melt(id_vars=['date'], var_name='hour')
df['timestamp'] = pd.to_datetime(df['date']) + pd.to_timedelta(df['hour'].str[4:].astype(int) - 1, unit='h')
df = df.set_index(pd.DatetimeIndex(df['timestamp']))
df = df.drop(columns=['timestamp', 'date', 'hour'])
df = df.resample('15T').ffill()
df = df.reset_index()
结果:
timestamp value
0 2019-01-01 00:00:00 0.1
1 2019-01-01 00:15:00 0.1
2 2019-01-01 00:30:00 0.1
3 2019-01-01 00:45:00 0.1
4 2019-01-01 01:00:00 0.2
.. ... ...
192 2019-01-03 00:00:00 0.2
193 2019-01-03 00:15:00 0.2
194 2019-01-03 00:30:00 0.2
195 2019-01-03 00:45:00 0.2
196 2019-01-03 01:00:00 0.4
本文收集自互联网,转载请注明来源。
如有侵权,请联系 [email protected] 删除。
我来说两句