This might not be practical if you have many different series of data with different years, but, you could include a where statement that says to do the expand where year = 2000 or year = 2007. Depending on which moving average method you're using, if it's in a data step, then have an if that says if year=2007 and month=1 then start the moving average over, or something to that effect. But if different series have different years, you would either have to start running them separately ( a pain) or perhaps you could build a data set that you merge in that provides that info (one column would indicate the item, another would have the year(s) that are relevant with an additional record for each year). Then do proc expand/Moving Avg on each, then append those data sets. But again, that would be no fun if you have different series with different years/ranges of years.
... View more
