st: From: Michael Stewart <michaelstewartresearch@gmail.com>

[owner-statalist@hsphsun2.harvard.edu]

Hello ,

I have a question regarding lists and have read Nicks paper on lists
and counting irregularly spaced intervals.

I posted a similar, but different question
http://www.stata.com/statalist/archive/2013-11/msg00882.html
previously.

My dataset is as follows


***********DATASET-BEGIN**************************
ID       VISITDATE         EVENT              LOCATION
1             Jan_1             May_1                  1
1             Jan_4             May_1 2
1 Jan_6 May_1 1
1 May_5 May_1 1
1 May_8 May_1 2
2 Feb_2 Feb_9 2
2 Feb_6 Feb_9 2
2 Feb_15 Feb_9 1
2 Feb_27 Feb_9 2
2 Mar_30 Feb_9 2

***********DATASET_END****************************


MY NEED:  I am trying to generate the following two variables and
looking to see if the programmers in the statalist may help me.

1) ***** for this variable, can you suggest an shorter/more elegant ********

VAR_1: this varaible shows VISITDATE when the LOCATION was the first
"2" [this is somewhat easier]
Ex:
For patient 1(ID=1), visit date corresponding to first 2 (in location)
would be Jan 4
For patient 2(ID=2), visit date corresponding to first 2 (in location)
would be Feb 2

*************code begin*******************
bysort ID  (VISITDATE): gen VAR_1=VISITDATE if LOCATION==2
bysort ID (VAR_1): gen LOCATION_FIRST2=VAR_1[1]
************code end************************

2)**********I do not know how to generate the following variable
******************

VAR_2: this variable shows VISITDATE  when the LOCATION is the second 2

Ex:
For patient 1(ID=1), visit date corresponding to second 2 (in
location) would be May 8
For patient 2(ID=2), visit date corresponding to second 2 (in
location) would be Feb 6

-- 
Thank you ,
Yours Sincerely,
Mike.
*
*   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/