st: Re: question about working with dates and times

["Joseph Coveney" <jcoveney@bigplanet.com>]

Paul Visintainer wrote:

Thanks, Joe.  I think I get it.  Rather than just specifying the interval, one
specifies the desired interval from the zero-point.

--------------------------------------------------------------------------------

In your original post, you mentioned wanting 36 hours, 14 days, etc.  You can 
take advantage of the -tc()- pseudofunction by treating 
-tc("02jan1960 00:00:00")- as a constant, representing how many milliseconds 
there are in one day.  See below for the 36-h, 14-d examples.  

The limitation of this tactic is illustrated as -day2- in my first reply, 
namely, the increment for "3 months" depends upon which three months of the 
year you're trying to span.

Joseph Coveney

. generate double day0 = tc("06oct2007 00:00:00")

. 
. * 36 h
. generate double day1 = day0 + ///
>     36 / 24 * tc("02jan1960 00:00:00")

. 
. * 14 days
. generate double day2 = day0 + ///
>     14 * tc("02jan1960 00:00:00")

. 
. format day? %tc

. 
. list day?, noobs

  +--------------------------------------------------------------+
  |               day0                 day1                 day2 |
  |--------------------------------------------------------------|
  | 06oct2007 00:00:00   07oct2007 12:00:00   20oct2007 00:00:00 |
  +--------------------------------------------------------------+


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/