Re: st: combining consecutive similar observations in time series datasets

[Jeph Herrin <stata@spandrel.net>]

You don't indicate if you need to worry about gaps. If not, then this will work:

 bys id coded (start) : egen newstart=start[1]
 bys id coded (end)   : egen newend = end[_N]
 bys id coded         : keep if _n==1

Or, if you really don't have any other variables

 collapse (min) start (max) end, by(id coded)

will work.

On 3/18/2013 2:51 AM, Adeel Yousaf wrote:
> Hi,
> how can I combine multiple rows of an activity to one row and generate
> its start and end time, in panel dataset, if the activity occurs
> adjacently. For example, in the first table 1 the activity code "11"
> has 3 rows from 4:00 am to 5:30am, I need to get its compact form
> shown in second table  :-
>
> actual dataset:-
> id      start   end     coded
> 100003  4:00    4:30    11
> 100003  4:30    5:00    11
> 100003  5:00    5:30    11
> 100003  5:30    5:45    13
> 100003  5:45    6:00    14
> 100003  6:00    6:30    14
> 100003  6:30    7:00    14
> 100003  7:00    7:15    15
> 100003  7:15    7:30    16
> 100003  7:30    8:00    16
> 100003  8:00    8:15    16
> 100003  8:15    8:30    17
> 100003  8:30    8:45    15
> 100003  8:45    9:00    15
> 100003  9:00    9:30    15
> 100003  9:30    10:00   15
>
> what i need is:-
> id      start   end     coded
> 100003  4:00    5:30    11
> 100003  5:30    5:45    13
> 100003  5:45    7:00    14
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/