Reputation: 28705
In this first data-set each employee has one team lead and one supervisor. I can transpose that no problem.
data a;
input employee_id ReportsTo $ ReportsToType $12.;
cards;
100 Jane Supervisor
100 Mark Team_lead
101 Max Supervisor
101 Marie Team_lead
102 Sarah Supervisor
102 Sam Team_lead
;
run;
proc transpose data = a
out = aTP(drop = _:);
by employee_id;
id ReportsToType;
var ReportsTo;
run;
/* Output */
/*employee_id Supervisor Team_lead */
/*100 Jane Mark */
/*101 Max Marie */
/*102 Sarah Sam */
Now, what if an employee can have anywhere from 1 to 3 team leads?
data b;
input employee_id ReportsTo $ ReportsToType $12.;
cards;
100 Jane Supervisor
100 Mark Team_lead
100 Jamie Team_lead
101 Max Supervisor
101 Marie Team_lead
101 Satyendra Team_lead
101 Usha Team_lead
102 Sarah Supervisor
102 Sam Team_lead
;
run;
/* Desired Output */
/*employee_id Supervisor Team_lead1 Team_lead2 Team_lead3 */
/*100 Jane Mark Jamie */
/*101 Max Marie Satyendra Usha */
/*102 Sarah Sam */
Using proc transpose gives an error telling me I can't have more than one identical ID variable in each group. Is there a procedure for transposing which does allow this?
ERROR: The ID value "Team_lead" occurs twice in the same BY group
Upvotes: 1
Views: 3410
Reputation: 8513
You need to change your input data so that rather than the word Team_lead
repeating, it shows it incrementing... i.e. Team_lead1
, Team_lead2
, etc...
You can use by-group processing and the retain statement to achieve this:
proc sort data=b;
by employee_id reportstotype;
run;
data want;
set b;
by employee_id reportstotype;
retain cnt .;
if first.reportstotype then do;
cnt = 1;
end;
if upcase(reportsToType) eq 'TEAM_LEAD' then do;
reportsToType = cats(reportsToType,cnt);
end;
cnt = cnt + 1;
run;
Then simply call proc transpose
like you did beforehand:
proc transpose data=want out=trans;
by employee_id;
id reportsToType;
var reportsTo;
run;
Upvotes: 1