topic Re: Transforming column variable to rows with one ID per row in SAS Programming

Transforming column variable to rows with one ID per row

love_epi — Mon, 16 May 2022 17:50:46 GMT

Hello,

I am trying to transform a column variable into several dummy variables with one row per ID.

Here is the data I have:

ID, Referral Type

1, Housing

1, Transportation

2, Food

3, Housing

3, Food

4, Transportation

4, Mental Health

Here is the data I want:

ID, Housing, Transportation, Food, Mental Health

1, 1, 1, 0, 0

2, 0, 0, 1, 0

3, 1, 0, 1, 0

4, 0, 1, 0, 1

Any advice would be greatly appreciated! Thanks 🙂

Re: Transforming column variable to rows with one ID per row

ballardw — Mon, 16 May 2022 18:11:45 GMT

How do you expect to use that resulting data set?

Your use of "dummy variables" makes me think you may want to use this in some sort of regression. In which case why cause yourself work. Most of the regression procedures support CLASS variables which SAS will use the levels to create dummy variables for you. For most purposes the form of data you have is preferred.

Or if you are looking for a report for people to read some thing like this perhaps:
Note use of DATA Step code to provide a usable data set.

data have;
  infile datalines dsd dlm=' ';
  input ID Referral :$15.;
datalines;
1 Housing
1 Transportation
2 Food
3 Housing
3 Food
4 Transportation
4 "Mental Health"
;

proc tabulate data=have;
   class id referral;
   table id,
         referral*n=' '
         /misstext='0'
   ;
run;

Re: Transforming column variable to rows with one ID per row

Ksharp — Tue, 17 May 2022 06:42:02 GMT

data have;
  infile datalines dsd dlm=' ';
  input ID Referral $15.;
  var=1;
datalines;
1 Housing
1 Transportation
2 Food
3 Housing
3 Food
4 Transportation
4 Mental Health
;

proc transpose data=have out=temp(drop=_name_);
by id;
id Referral;
var var;
run;
proc stdize data=temp out=want missing=0 reponly;
run;