Multiples in proc transpose

[Walternate]

Hi everyone,

I have a pretty complex dataset  that looks like this:

ID      Date1        Event1        Date2      Event2    

1       4/5/13          a             7/4/13          r              

2       5/12/13        b             8/5/13          x                  

1       4/5/13          a            10/9/13         g           

3       7/7/12          a             5/5/13         m           

2      12/5/11          c            9/5/13          r             

Basically, each time a person experiences Event1 (corresponding to Date1), it generates a row. A person can experience multiple Event1s and thus generate multiple rows (eg ID 2 in the above example). The row also includes the date and type of event2 for that person. If a person has more than one Event2, it will generate another row for each Event1 (eg ID 1 in the example above).

What I want is to transpose this dataset so that it looks like this:

ID      Date1        Event1        DateR      DateX        DateG     DateM        

1       4/5/13          a             7/4/13          --            10/9/13       --                 

2       5/12/13        b             9/5/13        8/5/13          --             --                                 

3       7/7/12          a                --              --               --          5/5/13                 

2      12/5/11          c            9/5/13        8/5/13           --              --                       

The ultimate goal is to know how many people experienced events of each kind, and how long after the initial event. The problem is that if I transpose using the original Event2 variable as

the id variable, there are repeats and I'm not sure if/how that will work.

Any help is much appreciated!

[PGStats]

So, what do you want to happen with those repeats. Your expected data structure doesn't allow repeats. Do you want to keep only the first occurrence of a given type, or only the last, or something else?

PG