08-15-2016 05:08 PM
Hello, I created the following grouping of variables:
%let practice_size= doctors, nurses, patients;
%let market=income, population, race;
%let organization=med_school, multi_market, affiliation;
For each of the variables listed above, I then created a new variable called check_variablename.
So I have check_doctors, check_nurses... check_affiliation. I am now trying to sum these check variables by the categories I created above. So I want FV_practice_size to be the sum of check_doctors, check_nurses, and check_patients. Is there a way to combine this "check" prefix with the %let statements. The following code is not functioning how I want it to.
The data looks like this:
doctors comp_doctors check_doctors nurses comp_nurses check_nurses
100 90 1 500 425 0
08-15-2016 06:14 PM
Since you don't want to process the variable names by CHECK then I suggest that the process that created the variables should use the appropriate prefix Practice_doctor, Practice_Nurse, Practice_Patient, Market_Income, Market_population, Market_race.
Then you could get the sums using the correct prefix.
Or possibly have a completely different structure such as
Roll with values of practice, market or organization
Group with the values like doctor, nurse, income, population, med_school media market
and the last variable holding a (check?) value.
Then Proc summary, means, report or tabulate would allow getting totals by the Roll variable and crossed with Group as needed.
08-15-2016 06:39 PM
For the small number of variables you have, you might as well type out the list. But if you are interested in extending your macro language skills (or if the lists are in reality much longer), you could take this approach:
08-15-2016 10:43 PM
You could change it before apply it . %let practice_size= doctors, nurses, patients; %let check=%sysfunc(prxchange(s/(\w+)/check_$1/,-1,%bquote(&practice_size))); %put &check ;