more value in same cell based on one column

[Deepankar]

I have an input data like this

Column1	Column2	Column3
1	AB	14/03/2018
2	DC	23/07/2002
3	NP	13/11/1910
1	LP	25/09/1901
2	JP	06/12/1925
3	QR	03/04/1974

i am looking for an output like

Column1	Column2	Column3
1	AB-LP	14/03/2018-25/09/1901
2	DC-JP	23/07/2002 -06/12/1925
3	NP-QR	13/11/1910-03/04/1974

[Reeza]

Is it always only 2 per record?

1. Sort data (PROC SORT)

2. Use BY/LAST to identify LAST records

3. Use LAG() to capture previous value

4. Use CATX()

Or Transpose and then use CATX on the output to combine the variables.

[novinosrin]

data have;
input (Column1	Column2	Column3) (:$10.);
datalines;
1	AB	14/03/2018
2	DC	23/07/2002
3	NP	13/11/1910
1	LP	25/09/1901
2	JP	06/12/1925
3	QR	03/04/1974
;


data _null_;
length _Column2 _Column3 $100;
if _n_=1 then do;
   dcl hash H (ordered: "A") ;
   h.definekey  ("Column1") ;
   h.definedata ("Column1","_Column2", "_Column3") ;
   h.definedone () ;
   call missing(of _all_);
   end;
set have end=l;
if h.find() ne 0 then h.replace(key:column1,data:Column1,data:Column2,data:Column3);
else do;_Column2=catx('-',_Column2,Column2);
_Column3=catx('-',_Column3,Column3);
h.replace();
end;
if l then h.output(dataset:'want');
run;

[Ksharp]

data have;
input (Column1	Column2	Column3) (:$10.);
datalines;
1	AB	14/03/2018
2	DC	23/07/2002
3	NP	13/11/1910
1	LP	25/09/1901
2	JP	06/12/1925
3	QR	03/04/1974
;
proc sort data=have;
by column1;
run;
data want;
length new2 new3 $ 200;
do until(last.column1);
 set have;
 by column1;
 new2=catx('-',new2,column2);
 new3=catx('-',new3,column3);
end;
drop column2 column3;
run;