pandas NaN after resample-CodePudding

I have this innocent looking data:

datetime,power
2022-02-14 15:09:58.163,114.07
2022-02-14 15:09:58.657,113.63
2022-02-14 15:10:32.237,114.28
2022-02-14 15:10:32.730,113.75
2022-02-14 15:10:33.195,113.76
2022-02-14 15:10:33.680,113.83
2022-02-14 15:10:34.195,114.44
2022-02-14 15:10:34.679,115.09

the following code produces NaNs

_df = pd.read_csv('measurements/nan_df.csv')
_df['datetime'] = pd.to_datetime(_df['datetime'])
_df.set_index('datetime', inplace=True)
_df.resample('s').mean()

datetime,power
2022-02-14 15:09:58,113.85
2022-02-14 15:09:59,
2022-02-14 15:10:00,
2022-02-14 15:10:01,
2022-02-14 15:10:02,
2022-02-14 15:10:03,
2022-02-14 15:10:04,
2022-02-14 15:10:05,
2022-02-14 15:10:06,
2022-02-14 15:10:07,
2022-02-14 15:10:08,
2022-02-14 15:10:09,

Any idea why?

CodePudding user response：

It is expected, because pandas by default create consecutive DatetimeIndex by DataFrame.resample, if not exist values is added NaNs.

If need remove values with misisng value use:

df = df.resample('s').mean().dropna()
print (df)
                       power
datetime                    
2022-02-14 15:09:58  113.850
2022-02-14 15:10:32  114.015
2022-02-14 15:10:33  113.795
2022-02-14 15:10:34  114.765