I am seeking to display the following simulated -twoway- graph, in which subjects are identified by the variable ID (which in this simulated data are 3, 8 and 9). I am wondering whether it is possible to display these data but skipping the missing ID rows, so that from the bottom up, the first row shows ID==3, the second ID==8, and so on. How would I modify the -twoway- command so that the rows for missing subjects not in the data (i.e. ID==1, 2, 4 and so on) are skipped? (Note: this graph combines multiple observations per subject, but I don't believe that's creating my problem of the space for missing ID's being kept in the graph.)
Code:
clear
input id trtw start stop stw nn event
3 0 0 12 1 2 1
3 0 17 548 0 2 0
8 0 0 53 1 15 1
8 0 54 108 1 15 0
8 0 109 201 1 15 0
8 0 202 203 1 15 0
8 0 206 216 1 15 0
8 0 218 317 1 15 0
8 0 319 324 1 15 0
8 0 325 344 1 15 0
8 0 347 354 1 15 0
8 0 356 376 1 15 0
8 0 379 458 1 15 0
8 0 460 464 1 15 0
8 0 476 487 1 15 0
8 0 491 545 1 15 0
8 0 546 554 1 15 0
9 1 0 284 1 2 1
9 1 286 602 0 2 0
end
bysort id: gen end = start[_n+1]
twoway (rcap start stop id if id <=10, horiz) ///
(rcap stop end id if id<=10 & event ==1 & id[_n+1]==id, ///
horiz lwidth(thick) lcol(black)) ///
(rcap stop stop id if id <10 & event == 1 & end==., ///
horiz lwidth(thick) lcol(black)), ///
leg(off) xtitle(Calendar time (days)) ytitle(Child) ///
xlabel(0(100)700) xscale(range(0 700))