st: RE: RE: RE: combine data by groups

["Nick Cox" <n.j.cox@durham.ac.uk>]

Yes, that's right. 

A more cautious approach would test for consistency of non-missing
values within -personid-. 

Nick 
n.j.cox@durham.ac.uk 

Martin Weiss

"-collapse- to maxima."

In this _particular_ constellation, the minimum will serve Guang just as
well:


*******
clear*

inp str10 personid  byte(parentsid  momlive dadlive)
001     1      0  .
001     2      .   1
002     1      1  .
002     2        .   1
end

compress
list, noo

collapse (min) momlive dadlive, by(personid)
list, noo
*******

Nick Cox

-collapse- to maxima. 

G. Dai

hi all, this is an easy question,but somehow I can't figure it.
The data is like:
personid     parentsid              momlive                   dadlive
001                   1                      0
 .
001                   2                      .
  1
002                   1                      1
 .
002                   2                        .
   1

where personid and parentsid  are person and her parents' id.
when parentid==1, we only record whether mom still live or not (momliv)
similarly
for parentid==2 and dadlive.

I want to change the data set to the following form

personid         momlive                   dadlive
001                  0                              1
002                   1                              1

what is the best way to do this?

*
*   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/