I would appreciate if I could get help with creating new variables from rows of an existing variable. I am doing a survival analysis (recurrent events). I have nearly 45000 subjects who have intervals varying from 1 to 17. For each interval, I need to get the cumulative dose of the medication. The number of individual medication varies from 1 to 7. As I need one row for each interval per subject, I need to long format for each interval to wide format, meaning converting 'n' rows of variable drug code to 'n' columns of drug code (code1, code2, ... coden). I am unsuccessful in creating loop for each interval per subject id.
I am giving an example of dataset.
Code:
* Example generated by -dataex-. For more info, type help dataex clear input float(id interval code cdose) 1 1 10 100 1 1 15 150 1 1 50 200 1 2 10 1000 1 2 15 1000 2 1 10 500 2 1 30 250 2 2 25 245 2 2 30 250 2 3 25 250 2 3 30 500 3 1 10 150 3 1 20 200 3 2 10 150 4 1 15 150 4 2 25 300 4 2 30 300 end
As the intervals subject varies from subject to subject and also the number of medications used varies from interval to interval, I have been struggling to get single row for each interval.
I very much appreciate help.
With warm regards,
Sateesh
0 Response to convert 'n' rows of a variable to 'n' columns
Post a Comment