Create a new column with label value

[Ronein]

Hello

I want to change the table from wide to long.

It means that for each ID there will be 3 rows (with information of X1 X2 X3 ).

I have 3 queations:

1-Since X3 is numeric and X1,X2 are char, should I need to do any convertion (Because in the long structure the values of X1 ,X2,X3 should be in same column)

2-Is there a better way to create the long structure table?

3-I want to add a new column to LongTbl that is called labelField that will get the value of label of X1/X2/X3.

What is the way to do it please in a way that it takes the label automatically ?(Here I type it but in real life I have many varaibles and I dont want to type the label for each varaible and I prefer to find a clever way to do it

Data WideTbl;
input ID date : date9. x1 $ x2  $ x3  ;
format date ddmmyy10.;
label x1='Team' x2='smoker' X3='status';
cards;
1 01jan2021 a y 1
2 01jan2021 b n 2
3 01jan2021 c n 2
4 01jan2021 a n 3
5 02jan2021 b y 1
6 02jan2021 a y 1
7 03jan2021 c y 2
8 03jan2021 a n 2
9 03jan2021 a y 3
;
run;

data LongTbl;
 set WideTbl;
     length category $3;
     category=x1; field='x1';labelField ='Team'; output;
     category=x2; field='x2'; labelField ='smoker';output;
     category=x3; field='x3';labelField ='status'; output;
     drop x1 - x3;
run;

[Reeza]

1. Yes, you'll need to convert your variables to the same type.

2. No, since you need to do type conversion, you may as well do it on in a data step

3. VLABEL() and VNAME() functions are what you're looking for here. 

PROC TRANSPOSE has the IDLABEL variable but not sure how that works when you go wide to long. 

Transposing data tutorials:
Wide to Long:
https://stats.idre.ucla.edu/sas/modules/how-to-reshape-data-wide-to-long-using-proc-transpose/

https://stats.idre.ucla.edu/sas/modules/reshaping-data-wide-to-long-using-a-data-step/

And sometimes a double transpose is needed for extra wide data sets:
https://gist.github.com/statgeek/2321b6f62ab78d5bf2b0a5a8626bd7cd

[Ronein]

Thanks, may you please show how to apply these solution here?

[Ronein]

Will it work well?

Data WideTbl;
input ID date : date9. x1 $ x2  $ x3  ;
format date ddmmyy10.;
label x1='Team' x2='smoker' X3='status';
cards;
1 01jan2021 a y 1
2 01jan2021 b n 2
3 01jan2021 c n 2
4 01jan2021 a n 3
5 02jan2021 b y 1
6 02jan2021 a y 1
7 03jan2021 c y 2
8 03jan2021 a n 2
9 03jan2021 a y 3
;
run;

/*To convert X3 to char*/
Data WideTbl2;
set WideTbl;
X3_char=compress(X3);
drop X3;
rename X3_char=X3;
Run;


data LongTbl;
set WideTbl;
length category $3 labelField $20. ;
 field='x1';category=x1;labelField = vlabel(x1);output;
 field='x2';category=x2;labelField = vlabel(x2);output;
 field='x3';category=x3;labelField = vlabel(x3);output;
 drop x1 - x3;
run;

[ballardw]

You really should avoid automatic numeric to character conversions that your code used. Explicitly PUT the numeric with an appropriate format AND add the -L modifier to left justify the value unless you want one or more leading spaces.

data LongTbl;
set WideTbl;
length category $3;
 field='x1';category=x1;labelField = vlabel(x1);output;
 field='x2';category=x2;labelField = vlabel(x2);output;
 field='x3';category=put(x3,best. -L);labelField = vlabel(x3);output;
 drop x1 - x3;
run;

[Ronein]

Thank you,May I ask some questions please:
What is the problem with compress(x3)?
What problem can be caused by using it?
Is PUT(X3,best.,-L) equivalent to LEFT(PUT(X3,best.)?

[Tom · Accepted solution]

Since you have unique BY variables just use PROC TRANSPOSE.

data WideTbl;
  input ID date : date9. x1 $ x2  $ x3  ;
  format date ddmmyy10.;
  label x1='Team' x2='smoker' X3='status';
cards;
1 01jan2021 a y 1
2 01jan2021 b n 2
3 01jan2021 c n 2
4 01jan2021 a n 3
5 02jan2021 b y 1
6 02jan2021 a y 1
7 03jan2021 c y 2
8 03jan2021 a n 2
9 03jan2021 a y 3
;

proc transpose data=widetbl 
  out=LongTbl(rename=(_name_=field _label_=labelField col1=category))
;
  by id date;
  var x1-x3;
run;

If you have mixed numeric and character then CATEGORY will automatically become character. 

                                    label
Obs    ID          date    field    Field       category

  1     1    01/01/2021     x1      Team      a
  2     1    01/01/2021     x2      smoker    y
  3     1    01/01/2021     x3      status               1
  4     2    01/01/2021     x1      Team      b
  5     2    01/01/2021     x2      smoker    n
  6     2    01/01/2021     x3      status               2
  7     3    01/01/2021     x1      Team      c
  8     3    01/01/2021     x2      smoker    n
  9     3    01/01/2021     x3      status               2

But you might want to add a step to left justify the strings generated from the numeric variables.

data LongTbl; set LongTbl;
  category=left(category);
run;

                                    label
Obs    ID          date    field    Field     category

  1     1    01/01/2021     x1      Team         a
  2     1    01/01/2021     x2      smoker       y
  3     1    01/01/2021     x3      status       1
  4     2    01/01/2021     x1      Team         b
  5     2    01/01/2021     x2      smoker       n
  6     2    01/01/2021     x3      status       2
  7     3    01/01/2021     x1      Team         c
  8     3    01/01/2021     x2      smoker       n
  9     3    01/01/2021     x3      status       2

[Ronein]

Thank you,
As I see you didnt use Idlabel in proc transpose. why?
How did you tell SAS to create a varaible with label value?
Thank you

[Tom]

The IDLABEL statement is for transposing the other way, when you are using the it to make new variables.

The fact that at least one of the variables being transposed had a LABEL attached to it is what caused PROC TRANSPOSE to create the _LABEL_ variable.