Concatenation variables

[KrishnaChandra]

Hai

  I have a doubt in data set programming

Here i can show my dataset.

ID	Test	1_Test	2_Test	3_test
U1	T	a
U1	T	b
U1	T	c
U2	B		i
U2	B		j
U3	C			k
U3	C			l
U3	C			m
U3	C			n

I would like to become the Varible 4_Test as shown in below

ID	4_test
U1	T
U1	a
U1	b
U1	c
U2	B
U2	i
U3	j
U3	C
U3	k
U3	l
U3	m
U3	n

Please Help me :smileyplain:

Regards

Krishna

[Reeza]

Have you tried a transpose?

Otherwise a data step is a good way, using an output statement to control the output.

data want;

set have;

array t_3) test1-test3;

test4=test;

output;

do i=1 to 3;

test4=t_(i);

output;

end;

keep ID test4;

run;

[PGStats]

You will have to use variable names starting with alphabetic characters (1_test is not a valid name, change it to test_1), then expand your dataset with:

data myDataset;

infile datalines missover;

length ID Test Test_1-Test_3 $8;

input ID Test Test_1 Test_2 Test_3;

datalines;

U1 T a

U1 T b

U1 T c

U2 B . i

U2 B . j

U3 C . . k

U3 C . . l

U3 C . . m

U3 C . . n

;

data want;

set myDataset; by test notsorted;

array t Test_:;

if first.test then output;

do i = 1 to dim(t);

    test = t{i};

    if not missing(test) then output;

    end;

keep ID test;

run;

proc print data=want noobs; run;

PG

[KrishnaChandra]

Hai PG,

   Thanx.. But the test value T,B,C are  repeats ,then how it possible to out the observation(first.Test)???

[PGStats]

Try replacing

set myDataset; by test notsorted;

by

set myDataset; by ID test notsorted;

That will output one observation for each consecutive ID with the same test.

PG

[PoornimaRavishankar]

Also not missing(test) has to change since they are character values. try if trim(left(compress(test))) ne ''

[PGStats]

The MISSING function works both for numeric and character arguments. The function returns true for empty strings. - PG

[KrishnaChandra]

but first.test then output;  The code doesn't work properly.The test value not shown .i meant(T B C)

[Reeza · Accepted solution]

This is tested based on the data you provided. Mistakes in original code include a bug in array declaration (character and missing brace) and extra output statement.

data have;

infile datalines missover;

length ID Test Test1-Test3 $8;

input ID Test Test1-test3;

datalines;

U1 T a

U1 T b

U1 T c

U2 B . i

U2 B . j

U3 C . . k

U3 C . . l

U3 C . . m

U3 C . . n

;

data want;

set have;

by ID ;

array t_(3) $ test1-test3;

test4=test;

if first.id then output;

do i=1 to 3;

test4=t_(i);

if not missing(test4) then output;

end;

keep ID test4;

run;

[KrishnaChandra]

Thanx Reeza

[Reeza]

data want;

set have;

by ID ;

array t_3) test1-test3;

test4=test;

if first.id then output;

do i=1 to 3;

test4=t_(i);

if not missing(test4) then output;

output;

end;

keep ID test4;

run;

[KrishnaChandra]

Thanx PG