The datetime module is part of the Python standard library, and is very popular and widely used. For this reason, it is a good idea to have some familiarity with it, as you will likely cross paths with it. The datetime module is actually fairly simple with a total of only six types of objects: date, time, datetime, timedelta along with two others on timezones. The pandas Timestamp and Timedelta objects have all the functionality of their datetime module counterparts and more. It will be possible to remain completely in pandas when working with time series.
Step 1 shows how to create datetimes, dates, times, and timedeltas with the datetime module. Only integers may be used as each component of the date or time, and are passed as separate arguments. Compare this to step 5 where the pandas Timestamp constructor can accept the same components as arguments, as well as a wide variety of date strings. In addition to integer components and strings, step 6 shows how a single numeric scalar can be used as a date. The units of this scalar are defaulted to nanoseconds (ns) but are changed to days (D) in the second statement with the other options being hours (h), minutes (m), seconds (s), milliseconds (ms), and microseconds (µs).
Step 2 details the construction of the datetime module's timedelta object with all of its parameters. Again, compare this to the pandas Timedelta constructor shown in step 9, which accepts these same parameters along with strings and scalar numerics.
In addition to the Timestamp and Timedelta constructors, which are only capable of creating a single object, the to_datetime and to_timedelta functions can convert entire sequences of integers or strings to the desired type. These functions also provide several more parameters not available with the constructors. One of these parameters is errors, which is defaulted to the string value raise but can also be set to ignore or coerce. Whenever a string date is unable to be converted, the errors parameter determines what action to take. When set to raise, an exception is raised and program execution stops. When set to ignore, the original sequence gets returned as it was prior to entering the function. When set to coerce, the NaT (not a time) object is used to represent the new value. The second statement of step 8 converts all values to a Timestamp correctly, except for the last one, which is forced to become NaT.
Another one of these parameters available only to to_datetime is format, which is particularly useful whenever a string contains a particular date pattern that is not automatically recognized by pandas. In the third statement of step 7, we have a datetime enmeshed inside some other characters. We substitute the date and time pieces of the string with their respective formatting directives.