Selecting first non-missing value for each subject

[gp123]

I have a dataset in which i need to identify the first non missing subject within a dose. 

My data looks like below

subject date         dose
10002 1/27/2015 10MG
10002 2/24/2015 10MG
10002 6/1/2015   20MG
10002 9/29/2015 10MG
10002 1/24/2016 10MG

I want my final dataset to look like below,

subject     date           dose
10002    1/27/2015   10MG
10002    6/1/2015     20MG
10002    9/29/2015   10MG

Is there any particular way to do this? If i do first.dose, its giving me only first 2 records of my final dataset. 

[art297 · Accepted solution]

data have;
  input subject date mmddyy10. dose $;
  format date mmddyy10.;
  cards;
10002 1/27/2015 10MG
10002 2/24/2015 10MG
10002 6/1/2015   20MG
10002 9/29/2015 10MG
10002 1/24/2016 10MG
;

data want;
  set have;
  by subject dose notsorted;
  if first.dose;
run;

Art, CEO, AnalystFinder.com

[gp123]

Thanks Art. But this gives me only 2 records.

subject     date           dose
10002    1/27/2015   10MG
10002    6/1/2015     20MG

I need the first non missing record every time the dose changes.

[art297]

You're doing something other than what I did. I get all 3 records.

Is it possible that you sorted your file before running the code I suggested?

Art, CEO, AnalystFinder.com

[gp123]

Yes, you are right. I had sorted my dataset before. I corrected it and it gives me 3 records.

Thank you very much.