How to sort by timestamps in pandas?

Question

So, I have timestamps that look like the following:

20140804:10:00:13.281486

20140804:10:00:13.400113

20140804:10:00:13.555512

20140804:10:00:13.435677

I have them in a DataFrame and I am trying to sort them in an ascending order. I have tried the following. But, it doesn't seem to work

df['yyyymmdd'] = pd.to_numeric(df['yyyymmdd'], errors='coerce')

df['hh'] = pd.to_numeric(df['hh'], errors='coerce')

df['mm'] = pd.to_numeric(df['mm'], errors='coerce')

df['ss'] = pd.to_numeric(df['ss'], errors='coerce')

df=df.sort(['yyyymmdd', 'hh','mm','ss'], ascending=[True, True,True,True])

Any help is appreciated.

Answer 1

You just have to ensure you denote the format specification properly, and you can use pd.to_datetime to convert them to actual datetimes before sort_values.

pd.to_datetime(stamps, format="%Y%m%d:%H:%M:%S.%f").sort_values()

This is much more direct than decomposing the timestamps in components and performing a multiple-criteria sort as you were attempting.

Demo

>>> stamps
0    20140804:10:00:13.281486
1    20140804:10:00:13.400113
2    20140804:10:00:13.555512
3    20140804:10:00:13.435677
dtype: object

>>> pd.to_datetime(stamps, format="%Y%m%d:%H:%M:%S.%f").sort_values()
0   2014-08-04 10:00:13.281486
1   2014-08-04 10:00:13.400113
3   2014-08-04 10:00:13.435677
2   2014-08-04 10:00:13.555512
dtype: datetime64[ns]