@John04 wrote:   I am trying to find average amount by product_group, product, YM and then using that average for new YM. for e.g  Until now I tried partition by and indexing but I think it doesn't work in sas. Hoping that someone can point me in right direction on how to solve this problem.   data have; input YM Product_Group$ Product$ Amount; datalines; 200101 X P1 1.2 200201 X P1 8.9 200301 X P1 432.4 200101 Y P2 24.23 200201 Y P2 842.9 200301 Y P2 43.4 run; I want output something like:   data want; input YM Product_Group$ Product$ Average; datalines; 200401 X P1 147.5 200401 Y P1 303.51 run; This is the easiest way I know to get the mean amount by product_group and Product proc summary data=have nway; class product_group product; var amount; output out=want (drop=_:) mean=; run; But getting a value that does not exist in the source data, ie that 200401, as part of any summary will require further processing, such as data step to add the value.     Large economy size hint: Instead of using random numbers containing parts of a date, create actual date values with appropriate formats such as: data have; input YM :yymmn6. Product_Group$ Product$ Amount; format ym yymmn.; datalines; 200101 X P1 1.2 200201 X P1 8.9 200301 X P1 432.4 200101 Y P2 24.23 200201 Y P2 842.9 200301 Y P2 43.4 run; Then you have all the power of the SAS functions such as INTNX to increment years, such as Intnx('year',ym,1)  instead of weird arithmetic like 200101 +100 to simulate a next year. Plus graphs with dates will space tickmarks in nicer intervals. The above would assume the first day of a month for date. ... View more