Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# RE: st: complex panel data variable generation question

 From "Hoffman, George" To "statalist@hsphsun2.harvard.edu" Subject RE: st: complex panel data variable generation question Date Thu, 1 Sep 2011 22:45:19 -0500

```Nick -
Works perfectly for what I asked for (which wasn't exactly what I wanted , but close enough).
Thank you for helping me over another hurdle
George Hoffman

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Nick Cox
Sent: Thursday, September 01, 2011 12:33 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: complex panel data variable generation question

It sounds as if you need to get your observations into the right order.

bysort i epoch (t) : gen last = _n == _N
bysort i (last t) : gen diff = x - x[_n-1] if last

On Thu, Sep 1, 2011 at 5:28 PM, Hoffman, George <ghoffman@mcw.edu> wrote:
> I have a panel dataset of observations in individuals (i=1-15) over time (t=1-4000)
> The observations are sequential by minute, but the duration of observation for each individual is not identical.
> I have xtdes, i(i) t(t)
>
> Within the observation period are epochs (1-8) defined by events that do not occur at identical intervals but do occur sequentially.
> In additional to the observations that occur at minute intervals are measurements made only at the end of each epoch.
>
> What I would like to do is generate a difference between the measures made once per epoch, but since these observations do not occur at regular intervals I can't get his to work (using sorting, by i: gen...)
>
*
*   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/

*
*   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/
```