The data always has a record for state. The code is written so that it only gets the state if city='New York'.   However, when city isn't 'New York' you would end up reading that record set's state as the next set's name variable and every other set will input the wrong information for each field.   Thus, the record that contains state has to be input, regardless of whether it inputs state or not.   Art, CEO, AnalystFinder.com   ... View more

You're trying to do too much in a single DATA step.  Most every DOW loop will use the same groupings in both DO loops.  For example, have the top loop read all observations for a DIS PerformanceLevel combination, computing a summary statistic.  Then use the bottom DO loop to re-read the exact same observations, appending those summary statistics to every observation in the group.   To take an example, suppose you wanted the 5 sums appended to every observation for that DIS PerformanceLevel combination.  You could use:   data setloop3; do until (last.PerformanceLevel);    set setloop2;    by Dis PerformanceLevel;    sumPL1=sum(sumPL1,PL1);    sumPL2=sum(sumPL2,PL2);    sumPL3=sum(sumPL3,PL3);    sumPL4=sum(sumPL4,PL4);    sumPL5=sum(sumPL5,PL5); end; do until (last.PerformanceLevel);    set setloop2;    by Dis PerformanceLevel;    output; end; run;   Now the 5 sum variables contain subtotals for the DIS PerformanceLevel combination, but those 5 subtotals become part of every observation. ... View more