It can be very useful to extract certain "fields" from the timestamp, such as the year, month, day, etc.

See eg https://pandas.pydata.org/docs/user_guide/timeseries.html#time-date-components for the ones available in pandas.

Using pandas as an example, there are the basic components of the datetime:

>>> ts = pd.Timestamp.now()
>>> ts
Timestamp('2021-02-24 10:47:54.294504')

>>> ts.year
2021
>>> ts.month
2
>>> ts.day
24
>>> ts.hour
10
>>> ts.minute
49
>>> ts.second
54
>>> ts.microsecond
607393
>>> ts.nanosecond
0

(only for the sub-second, this is not fully clear how to divide it in microseconds or milliseconds, etc)

But in addition also some more "advanced" like:

>>> ts.dayofyear
55
>>> ts.dayofweek
2
>>> ts.week
8
>>> ts.isocalendar()
(2021, 8, 3)

Reporter: Joris Van den Bossche / @jorisvandenbossche
Assignee: Rok Mihevc / @rok
Watchers: Rok Mihevc / @rok

Related issues:

• [C++] Implement week kernel that accepts WeekOptions (is related to)
• [R] bindings for lubridate's year, isoyear, quarter, month, day, wday, yday, isoweek, hour, minute, and second functions (is depended upon by)
• [R] Implement common date and time functions for dplyr (is depended upon by)

PRs and other links:

• GitHub Pull Request #10176

_{Note: This issue was originally created as ARROW-11759. Please see the migration documentation for further details.}