Creating a time index

[Simon Sampson]

My data currently has a period variable which is a number with the first 4 digits being the year, the second two being the month - e.g. Jan 2012 is 201201. These run from Jan 2012 - Dec 2015. Is there a way to create a new variable where these periods are numbered from 1 to 48?

[Nick Cox]

See

help dates

for guidance here. The easiest method I know using official commands is this:

clear

input mydate
201201
201512
end

gen mdate = ym(floor(mydate/100), mod(mydate, 100))
format mdate %tm

list

Note that the perhaps more obvious monthly(strofreal(mydate), "YM") does not work (as of Stata 14) with run-together integers. For the same reason, numdate (SSC) will not work either.

The rather ancient todate (SSC) (first made public 2001) does work here. You must install it before you can use it, but need only do that once on the same system.

capture ssc install todate
todate mydate, pattern(YYYYMM) gen(mdate2)

[Simon Sampson]

I see the code works to change each period from 201201 to 2012m1 as appropriate. However I was hoping to simply number each period from 1 to 48. I.e. 201201 =1, 201202 = 2 etc. That said, I'm not really sure if this is even necessary - what would be best to use to denote time in panel data?

[William Lisowski]

You will want to follow the guidance in help datetime, which is without a doubt the most visited documentation on my system, with the second-most-visited being Chapter 24 (Working with dates and times) of the Stata User's Guide PDF available from the PDF Documentation item on Stata's Help menu. Before working with dates and times, any Stata user should read the very detailed Chapter 24 thoroughly. After that, the help datetime documentation will usually be enough to point the way.

With regard to working with panel data, you should begin by familiarizing yourself with the introductory material in the introductory section ("xt") of the Stata Longitudinal-Data/Panel-Data Reference Manual PDF. In general to the extent that you have meaningful times, it is not a good idea to discard that information in favor of a sequence number.

[Nick Cox]

I agree with all of William's advice. But it follows from studying the help for dates as both he and I suggest that time indexes 1 up could be obtained from the monthly dates generated by just subtracting the constant ym(2011, 12). That shifts the origin in your case.

For many of these little date problems I check my logic by using display for one-off calculations. di is allowed as an abbreviation.

. di ym(2012, 1)
624

. di ym(2011, 12)
623

This is just how you would do it: it is neither necessary nor helpful.