我有一个数据框,日期索引类型是时间戳,时间列是 datetime.Time :
Time Value
Date
2004-05-01 0:15 3.58507
2004-05-02 0:30 3.84625
...
如何将其转换为:
Value
Date
2004-05-01 0:15 3.74618
2004-05-01 0:30 3.58507
2004-05-01 0:45 3.30998
我编写了一个可行的代码,但它不是非常pythonic:
ind = frame.index.get_level_values(0).tolist()
tms = frame['Time']
new_ind = []
for i in range(0, len(ind)):
tm = tms[i]
val = ind[i] + timedelta(hours=tm.hour, minutes=tm.minute, seconds=tm.second)
new_ind.append(val)
frame.index = new_ind
del frame['Time']
答案 0 :(得分:4)
您可以先转换Time
to_timedelta
列,然后添加到index
,drop
列Time
,如有必要,请设置索引name
:
df.Time = pd.to_timedelta(df.Time + ':00', unit='h')
df.index = df.index + df.Time
df = df.drop('Time', axis=1)
df.index.name = 'Date'
print (df)
Value
Date
2004-05-01 00:15:00 3.58507
2004-05-02 00:30:00 3.84625
如果列Time
为datetime.time
,则我首先将作品投射到string
(如有必要,添加:00
):
df.Time = pd.to_timedelta(df.Time.astype(str), unit='h')
df.index = df.index + df.Time
df = df.drop('Time', axis=1)
df.index.name = 'Date'
print (df)
Value
Date
2004-05-01 00:15:00 3.58507
2004-05-02 00:30:00 3.84625