Subjects with no failure event not in risk pool, Stset

[Mary Carmack]

Hello-

I am using the following commands for my set of patients with multiple records and multiple events of "failure":

replace day = date(t, "DMY")
stset day, id(patient) origin(code==1) failure(code==2) exit(time mdy(3,15,2019))

With the following dataset:

* Example generated by -dataex-. To install: ssc install dataex
clear
input str2 patient str11 t float(code day)
"p1" "1jan2019"  1 21550
"p1" "1mar2019"  2 21609
"p2" "5jan2019"  1 21554
"p3" "10jan2019" 1 21559
"p4" "2jan2019"  1 21551
"p4" "1feb2019"  2 21581
"p4" "3mar2019"  2 21611
end

I choose exit to be a time because my followup period ends March 15, 2019. However, if I look at the values stset generates, it never places patients 2 and 3 into the risk pool (instead, _st is 0). From my understanding, they should enter the risk pool at 5jan2019 (for p2) and 10jan2019 (for p3), and then be censored on March 15, 2019. Am I misunderstanding the risk pool criteria?

clear
input byte(_t _t0 _st _d)
 .  . 0 .
59  0 1 1
 .  . 0 .
 .  . 0 .
 .  . 0 .
30  0 1 1
60 30 1 1
end

Please advise, help much appreciated.