Get first unique values

[catkat96]

Hello, this is the dataset I have available:

data Have;
input ID arrange_date :ddmmyy. datestamp :ddmmyy.;
format arrange_date datestamp ddmmyy10.;
datalines;
001 05/06/2020 01/06/2020
001 05/07/2020 01/06/2020
001 05/08/2020 01/06/2020
001 05/06/2020 02/06/2020
001 05/07/2020 02/06/2020
001 05/08/2020 02/06/2020
001 05/06/2020 03/06/2020
001 05/07/2020 03/06/2020
001 05/08/2020 03/06/2020

001 10/10/2020 03/10/2020
001 10/11/2020 03/10/2020
001 10/12/2020 03/10/2020
001 10/10/2020 04/10/2020
001 10/11/2020 04/10/2020
001 10/12/2020 04/10/2020

002 20/03/2020 10/03/2020
002 20/04/2020 10/03/2020
002 20/05/2020 10/03/2020
002 20/06/2020 10/03/2020
002 20/03/2020 11/03/2020
002 20/04/2020 11/03/2020
002 20/05/2020 11/03/2020
002 20/06/2020 11/03/2020;
run;

 To clarify: 

-There are many client ID's.

-The second column refers to certain arranged dates when the client needs to perform an action. 

-The third column is the datestamp. 

-There is one row per client's arrangement date, per datestamp. As you can see, client 001 has two different "periods" in this table, one that goes from 05/06 and another that starts in 10/10.

I'm interested in keeping only the first unique occurrences of arrange_date. In this case it would look like this:

data Have;
input ID arrange_date :ddmmyy. datestamp :ddmmyy.;
format arrange_date datestamp ddmmyy10.;
datalines;
001 05/06/2020 01/06/2020
001 05/07/2020 01/06/2020
001 05/08/2020 01/06/2020

001 10/10/2020 03/10/2020
001 10/11/2020 03/10/2020
001 10/12/2020 03/10/2020

002 20/03/2020 10/03/2020
002 20/04/2020 10/03/2020
002 20/05/2020 10/03/2020
002 20/06/2020 10/03/2020

;
run;

As you can see, I only kept the first occurrence of each "arrange_date". 

I'm okay if it's done either in proc sql or a data step.

[andreas_lds · Accepted solution]

Try

proc sort data=have out=want nodupkey;
   by id arrange_date;
run;

[catkat96]

yes, perfect! I was juggling more complex things in my mind but sometimes simpler is better!!