Shorter way to do variable list 1 = variable list 2

[benjaminchua002]

I have two arrays, year and yearsum. Yearsum is the sum of values for each year. 

I am trying to create a new row 'Total' and use the yearsum values as the total values. My code works but it's incredibly inefficient. I am not sure if I can use a DO loop because it will generate 10 rows of 'Total'. 

if last then do;
Variable = 'Total';
year1 = yearsum1;
year2 = yearsum2;
year3 = yearsum3;
year4 = yearsum4;
year5 = yearsum5;
year6 = yearsum6;
year7 = yearsum7;
year8 = yearsum8;
year9 = yearsum9;
year10 = yearsum10;
output;
end;

Is there a way to write year{i} = yearsum {i} more efficiently?

I did a DO loop for the top portion but I can't apply it to the above code.

do i = 1 to 10;
yearsum{i} = sum(yearsum{i}, year{i});
end;

End result:

[Shmuel]

You just need to define ARRAY statements:

array year year1-year10;
array yearsum yearsum1-yearsum10;

do i = 1 to 10;
    yearsum{i} = sum(yearsum{i}, year{i});
end;

[Kurt_Bremser]

A similar result can easily be achieved with PROC REPORT from your untransposed vertical dataset.

[PaigeMiller]

Yes, you could use arrays, but any time you have a wide data set, with a specific year in the variable names (or any calendar information in the variable names), you have chosen a data arrangement that is more difficult to work with, and probably requires a lot more programming than if you kept the data in a long format, with year as a variable containing values 1 through 10.

You see, SAS already provides several built-in methods of doing what you are trying to do, they have done the work so you don't have to, if you use a long data set, then you have much less programming. PROC REPORT get this output table easily, without you having to do all the coding of individual years.

See Maxim 19. Embrace long data sets. Make your coding simpler.

[Kurt_Bremser]

A simplified example for using the longitudinal dataset in PROC REPORT:

data have;
input variable $ year value;
datalines;
A 1 111
A 2 222
A 3 333
B 1 444
B 2 555
B 3 666
;

proc report data=have;
column variable year,value;
define variable / group;
define year / across;
define value / "" sum;
run;