Your graph would seem obtainable by
linkplot y day, link(id) c(L)
Nick
[email protected]
Allan Reese
> Thanks to Nick Cox for drawing my attention to "help
> stylelists". That will save a lot of typing in future.
>
> Michael Blasnik is right that plotting these graphs from the
> long format data is a better solution. I've tried that in
> the past, and failed because data were not sorted correctly
> or there might be single points. In these data each
> individual was necessarily observed over a series of days, so
> the data can be reverse-sorted by the start-date for each
> individual, then forward-sorted on the individual's dates.
> c(L) then works fine.
>
> Long data has vars id, day, y
>
> egen minday= min(day), by(id)
> gsort -minday day
> scatter y day, c(L) ms(i)
>
> Wow, that's saved some typing! Thanks.
>
> I'd got in the habit of structuring data wide (multiple
> variables) so that different variables (subsets) could be
> distinguished. It's also easier to stick with scatter, rather
> than use other specialist commands and look up which options
> are allowed with each. There are other ways of doing the following:
>
> gen malewt = weight if sex=="M"
> gen femwt = weight if sex=="F"
> gen vbig = weight>100
> scatter malewt femwt vbig day, c(l l .) ms(i i t) mlab(. . id)
>
> Thanks to Vince Wiggins for technical background. In the
> immediate application, finding the limit of 100 lines just
> helped me decide the data fell naturally into subsets, with
> fewer than 100 individuals in each. So no problem in waiting
> for future updates ;-) As Vince points out, if there's a
> limit, some fool will try to bend it.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/