Pandas 格式化时间纳秒

Question

I am having a problem with formatting a time series in a pandas dataframe.

00:18:41.728.560.640 

is my time object. It is hour:min:sec:ms:us:ns.

My approach was:

df['Time'] = pd.to_datetime(df['Time'], format="%H:%M:%S.%f")

Results in:

ValueError: unconverted data remains: .560.640

But this is only working for milliseconds. Is there another way to format the time or what am I doing wrong?

Answer 1

Datetime struggles with the dots separating ms, us and ns. You can remove them with regex:

df['Time'] = pd.to_datetime(df['Time'].replace(r'(?<!:\d\d)\.', '', regex=True), format="%H:%M:%S.%f")

We use a negative lookbehind (?<::\d\d) to make sure the first dot isn't replaced by an empty string.

Answer 2

Another option would be:

pd.to_timedelta(df['Time'].str[::-1].str.replace('.', '', 2).str[::-1])

                          0
0 0 days 00:18:41.728560640