SAS - calculate average values of previous n quarters

Question

I have a SAS dataset that looks like:

             Obs    qy         Time       beginning   end                   

               1    2008 Q1    2008Q1      1           1                    
               2    2008 Q1    2008Q1      1           2                    
               3    2008 Q2    2008Q2      1           1                    
               4    2008 Q2    2008Q2      1           2                    
               5    2008 Q2    2008Q2      1           1                    
               6    2008 Q3    2008Q3      1           1                    
               7    2008 Q4    2008Q4      1           1                    
               8    2008 Q4    2008Q4      1           2                    
               9    2008 Q4    2008Q4      1           1                    
              10    2009 Q1    2009Q1      1           1   

and I would like to have some new variables containing the accounts that had beginning=1 and end=2 per quarter, for the n previous quarters (n could be 1,2,3,4 etc). How can I do that?

Desired output:

             Obs    qy         Time       beginning   end      one_quarter_back             

               1    2008 Q1    2008Q1      1           1   0.25                  
               2    2008 Q1    2008Q1      1           2   0.25                 
               3    2008 Q2    2008Q2      1           1   0.5                 
               4    2008 Q2    2008Q2      1           2   0.5              
               5    2008 Q2    2008Q2      1           1   0.5              
               6    2008 Q3    2008Q3      1           1   0.33              
               7    2008 Q4    2008Q4      1           1   0.4             
               8    2008 Q4    2008Q4      1           2   0.4              
               9    2008 Q4    2008Q4      1           1   0.4                 
              10    2009 Q1    2009Q1      1           1   0.15

(The above assuming that the respective calculations led to these numbers, and we would have similar for 2,3,4 etc quarters back)

Answer 1

You can use the LAG function. LAG enables you to retrieve previous values.

If you use LAG or LAG1, you are looking for that column's value from the previous row. LAG2 gives you the previous value two rows back. LAG3 gives you the previous value three rows back, and so on.

one_quarter_back = lag1(qy)

Your first observation will be null, but the seccond will have the value from the previous observation. The third will have the value from the seccond and so on.