data set up variables

[vraj1]

i have a sample dataset as below

subj	var2	var3	var4
1	time	12043	mark
1	jack	43535	steve
2	bob	52352	vsvgv
2	steve	24565	fbhdhn
3	kdire	2355	rtgeff
3	dfsde	63634	rwege
3	kjyytt	34235	fgsggsd

and i need to set uplike below

subj	value	desc
1	time	var2
1	12043	var3
1	mark	var4
1	jack	var2
1	43535	var3
1	steve	var4
2	bob	var2
2	52352	var3
2	vsvgv	var4
2	steve	var2
2	24565	var3
2	fbhdhn	var4

can anyone help me how to do this

[gamotte · Accepted solution]

Hello,

data have;
    infile cards dlm=",";
input subj var2 $ var3 var4 $;
cards;
1,time,12043,mark
1,jack,43535,steve
2,bob,52352,vsvgv
2,steve,24565,fbhdhn
3,kdire,2355,rtgeff
3,dfsde,63634,rwege
3,kjyytt,34235,fgsggsd
;
run;

data want;
    set have;
    array nums(*) _NUMERIC_;
    array chars(*) _CHARACTER_;

	format value $10. desc $4.;
    keep subj value desc;

    /* We exclude subj fom the loop */
    do i=2 to dim(nums);
        value=strip(put(nums(i),10.));
        desc=vname(nums(i));
		output;
    end;

    do i=1 to dim(chars);
        value=chars(i);
        desc=vname(chars(i));
		output;
    end;
run;

[PeterClemmensen]

Like this

data have;
input subj var2$ var3$ var4$;
datalines;
1 time 12043 mark 
1 jack 43535 steve 
2 bob 52352 vsvgv 
2 steve 24565 fbhdhn 
3 kdire 2355 rtgeff 
3 dfsde 63634 rwege 
3 kjyytt 34235 fgsggsd
;

data want;
	set have;
	array val{2:4} var2-var4;
	do i=2 to 4;
		value = val[i];
		desc = vname(val[i]);
		output;
	end;
	drop var2-var4 i;
run;

[RW9]

Out of interest, why?  Neither of those two datasets you provide as example are of much use to anyone.  Could just be the test data you provide, however there is no way from either of those to tell if 43535 should belong to time or jack for instance.  I would suggest sorting out the data before getting to this stage, and assign proper ids and such like so that you can correctly identify a unique datapoint.