Macro variables are just strings.  You have made two macro variables with the same 9 characters.   If you have a macro variable with a value that the DATE informat understands then you can use as a date literal in your code by adding quotes and the letter d.  So here is another convoluted way to make a copy of your DCS3YR macro variable.  %let copy2 = %sysfunc(putn("&dcs3yr"d,date9.));   ... View more

First thing to do: create a version without any macro-statement. If you already have that code, post it, so we can help. It seems as if the target-varialbes can be calculated using a normal do-loop and arrays. But to suggest something usefull i need to see data you have and what you expect as result. ... View more

It sounds like you want to solve an optimization problem, but it is not clear to me which of your quantities are decision variables to be determined by the solver and which are input parameters that have fixed values.  In PROC OPTMODEL, you can use a VAR statement to declare decision variables, a NUM statement to declare numeric parameters, and a MIN statement to declare a minimization objective. ... View more

Kiteulf, The macro example supplied by Astounding is a good starting point. Except for a small glitch: rather than using a KEEP statement to keep the output variables, I would use a DROP statement to drop the input variables. That way, you can get away with using multiple macro calls in the same data step.   I would write the macro something like this (not tested):     %macro calc (varname); do; PSI _&varname = ABS((DEV_&varname - &varname)*log(DEV_&varname / &varname))/100; length ALERT_&varname $25;   select; when(missing(PSI_&varname)) ALERT_&varname ='OBS - ingen observasjoner'; when(PSI_&varname <= 0.05) ALERT_&varname = 'Normal PSI'; when(PSI_&varname <= 0.1) ALERT_&varname = 'OBS - Avvik fra PSI oppdaget'; when(PSI_&varname <= 0.2) ALERT_&varname ='OBS! - Høy PSI oppdaget'; when(PSI_&varname <= 0.3) ALERT_&varname ='OBS!! - Svært høy PSI oppdaget'; otherwise ALERT_&varname ='OBS!!! - Ekstermt høy PSI oppdaget'; end; drop DEV_&varname &varname; end %mend calc; I changed the IF statements to a SELECT, in order to make the code run faster. The way your code was written, a high value would first be assigned the default, then it would be rewritten each time the PSI value was shown to be higher. And a normal value would only be written once, but it would be tested in every IF statement. Instead, with the SELECT statement, the output variable only gets written once, and the PSI is tested as few times as possible. To use the SELECT statement like this, it is important to remember that the order of the statements is crucially important - the smallest value (missing) first, and then compare  to increasing limits. You could also write it with descending values (using ">" instead of "<=", and the missing case in the OTHERWISE), but I assumed that the normal (small) values would be most frequent, meaning that the code would run faster that way.   Instead of using KEEP (when you use KEEP several times, with different variables, what is actually going to be kept?), I used DROP to drop the input variables, which will work even if you call the macro many times in the same datastep.     Note that I wrapped the code in the macro in DO; - END, and did not put a semicolon after the final END. This style of coding means that the macro can (and should) be used just like a normal SAS statement, and you can write code like data X; set Y; if p>4 then %calc(a); else %calc(b); run; Regards, Søren   ... View more