Dealing with dates

Dates are complicated. Just think of the Y2K bug, the pending Year 2038 problem, and time zones. It's a mess. We encounter dates naturally when dealing with the time-series data. pandas can create date ranges, resample time-series data, and perform date arithmetic operations.

Create a range of dates starting from January 1, 1900 with 42 days as follows:

print "Date range", pd.date_range('1/1/1900', periods=42, freq='D')

January has less than 42 days, so the end date falls in February as you can check for yourself:

Date range <class 'pandas.tseries.index.DatetimeIndex'>
[1900-01-01, ..., 1900-02-11]
Length: 42, Freq: D, Timezone: None

The following table from the pandas official documentation (refer to http://pandas.pydata.org/pandas-docs/stable/timeseries.html#offset-aliases) describes frequencies used in pandas:

Short code	Description
B	Business day frequency
C	Custom business day frequency (experimental)
D	Calendar day frequency
W	Weekly frequency
M	Month...