Date Variables: Multiple Date Observations put into one observation with multiple date variables

[mbrlvlctrl]

Hello Community,

I just need help getting ideas on how to approach this and it seems simple, but I can only think of inefficient ways to get this done.

I have information dealing with start and end dates but it is currently set up so that if a person comes in multiple times, it's a separate observation in itself

id   start            end

1    1/1/10          1/10/10

1    2/25/10         3/1/10

kind of like that.

I want to make it so it looks like this

id   start1          end1        start2          end2

1    1/1/10         1/10/10     2/25/10       3/1/10

Any guidance would be helpful.

[Astounding]

Does efficiency really matter?  You only need to transform the data once.  Anyway here's one way (assuming your data is sorted by ID):

proc transpose data=have prefix=start out=start_dates;

   by id;

   var start;

run;

proc transpose data=have prefix=end out=end_dates;

   by id;

   var end;

run;

data want;

   merge start_dates end_dates;

   by id;

run;

You will find that some programming tasks are easier, but they now depend on you knowing how many start and end dates you have.

Good luck.