I have a column of event\'s timestamp in my dataframe, looks like \"2016-06-10 18:58:25:675\", the last 3 digits are milliseconds, is there a efficient way to transform this
Normally there is a period / full-stop between the seconds and microseconds.
Since your data does not have that you can specify how your data is formatted:
pd.to_datetime('2016-06-10 18:57:35:317', format='%Y-%m-%d %H:%M:%S:%f')