将熊猫数据框中的初始日期与时间列合并为日期时间

Question

I'm reading text log files that have the year-month-day in the header, and a column with hours:minutes:seconds for each row. Sort of like this:

Yr=17  Mn= 3 Dy= 3

19:22:59.894       52
19:24:12.130      106
19:24:13.241      107
...

I have the date as a datetime.date object: eg datetime.date(2017, 3, 3) . I have the times as a Series:

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

How can I add together the scalar date with the array of times?

Also, some of these logs go past midnight. I'm thinking I need to use something like numpy.unwrap() to continue onto the next date, but I'm not sure how to do that with pandas.datetime .

Answer 1

Instead of using pd.to_datetime on the times , you could use pd.to_timedelta . Then, you could simply add the datetime for that file to the entire column and it would be converted to a column of datetimes.

Example

times = ['4:23:12.12', '11:25:43.23', '14:29:55.42']

df = pd.DataFrame(dict(times=times))



df
#             times
# 0 04:23:12.120000
# 1 11:25:43.230000
# 2 14:29:55.420000

file_date = datetime.date(2017, 3, 3)



df
#                     times
# 0 2017-03-03 04:23:12.120
# 1 2017-03-03 11:25:43.230
# 2 2017-03-03 14:29:55.420

---

Alternatively, you could try to read the dates/times upon input with parse_dates .