time constraints - use the first

[monday89]

I have the following dataset. What I want to do is create a new time_3 column that uses the FIRST date. if one of them is missing, use the one that is available. 

subjectid	time_1	time_2
1	1/2/2012 1:50
2	1/4/2012 13:53	1/3/2012 1:50
3	1/5/2012 1:50	2/6/2012 9:32
4		1/9/2019 12:12

subjectid	time_1	time_2	time_use
1	1/2/2012 1:50		1/2/2012 1:50
2	1/4/2012 13:53	1/3/2012 1:50	1/3/2012 1:50
3	1/5/2012 1:50	2/6/2012 9:32	1/5/2012 1:50
4		1/9/2019 12:12	1/9/2019 12:12

I had to do in two data sets and is there any one step that solves all? 

data want; set have;

if time_1 < time_2 then time_use= time_1;
else if time_2 < time_1 then time_use= time_2;
format time_use datetime16.;run;
run;

data have; set have;

if time_use= . and ventstopdtm = . then time_use= time_2;
else if time_use= . and time_2 = . then time_use= time_1 ;
else time_use= time_use;
format time_use datetime16.;
run;

[kelxxx]

data want;
set have;
if nmiss(time_1,time_2) then time_use=max(time_1,time_2);
else time_use=min(time_1,time_2);
run;

hi,

im not sure that is optimal but i think that it s worked

can you test it?

have a good day

A

[novinosrin]

data want;
 set have;
 if nmiss(time1,time2) then time3=coalesce(time1,time2);
 else time3=min(time1,time2);
run;

[mkeintz]

Since you are examining only 2 time variables, min(time1,time2) is the same as coalesce(time1,time2)  for your purposes (i.e. when there is one or more missing value):

data want;
   set have;
   newtime=min(time1,time2);
run;