How do I identify the first instance an observation appears based on a variable? (repeated measures)

[asgee]

Hi all, 

Repeated measures data is frying my brain 😣... I'm working on a dataset that has multiple visits overtime. I want to reduce the visits per ID down to just one observation per ID. A sample of the data is below:

ID	Visit	Label
AAA	1	.
AAA	2	No
AAA	3	Yes
BBB	1	.
BBB	2	Yes
CCC	1	Yes
DDD	1	.
DDD	2	No
DDD	3	No
EEE	1	.
EEE	2	.
FFF	1	Yes
GGG	1	No
HHH	1	.
HHH	2	.
HHH	3	.

Everything is sorted in chronological order based on their visit (1->2->3->etc.). Basically I have three criteria to follow per ID:

1) If the ID has at least one label with a "Yes", pick the row where a "Yes" label first appears. 

2) If the ID ONLY has "No" / "missing" labels across all visits, pick the row where a "No" label first appears.

3) If the ID ONLY has missing observations across all visits, pick the row where a missing label first appears.

Initially, the code I have only reduces the observations down to their very first visit (regardless of their label). Now I'm having trouble with coding the part of identifying "the first time a "___" appears". 

I'd want to hopefully produce a table that looks something like this:

ID	Visit	Label
AAA	3	Yes
BBB	2	Yes
CCC	1	Yes
DDD	2	No
EEE	1	.
FFF	1	Yes
GGG	1	No
HHH	1	.

Any help or advice would be really appreciated! 

[novinosrin · Accepted solution]

Hi @asgee  Solution plagiarized from @Reeza . She stumped with me with this elegant approach. I don't like intelligent people

data have;
input ID $	Visit	Label $;
cards;
AAA	1	.
AAA	2	No
AAA	3	Yes
BBB	1	.
BBB	2	Yes
CCC	1	Yes
DDD	1	.
DDD	2	No
DDD	3	No
EEE	1	.
EEE	2	.
FFF	1	Yes
GGG	1	No
HHH	1	.
HHH	2	.
HHH	3	.
;

proc sort data=have out=_have;
 by id descending label;
run;

data want;
 set _have;
 by id;
 if first.id;
run;

[asgee]

Hi @novinosrin (& @Reeza), thanks for your help, the code works perfectly! Ahahah both you and Reeza are amazing intelligent people, can't thank you enough both for all your help.