topic Re: Dynamic Catx Macro for % and CIs in SAS Programming

Dynamic Catx Macro for % and CIs

delgoulding — Wed, 31 Oct 2018 17:31:39 GMT

I want to create a macro to dynamically catx a percentage with a lower and upper CI.

The variables always look like this:

a_percent a_low a_high

b_percent b_low b_high

c_percent c_low c_high

Originally I was trying to put the suffix as a prefix so I can create arrays for percents, lows, and highs. Then I would be able to reference the arrays in a catx macro.

data _null_;
if 0 then set have;

array percent {*} percent:;
array low {*} low:;
array high {*} high:;

call symputx("dim_percent", dim(percent));
call symputx("dim_low",dim(low));
call symputx("dim_high", dim(high));

do i = 1 to dim(percent);
call symputx(cats("percent_",i),vname(percent{i}));
end;
do i = 1 to dim(low);
call symputx(cats("low_",i),vname(low{i}));
end;
do i = 1 to dim(high);
call symputx(cats("high_",i),vname(high{i}));

end;

run;

And the ultimate goal is to be able to loop through something like this:

var1=catx("^n(", put(percent(i), percent20.3), put(low(i), percent20.3));
varwant=catx(", ", var1, put(high(i),percent20.3))||")";

Re: Dynamic Catx Macro for % and CIs

Reeza — Wed, 31 Oct 2018 17:48:52 GMT

Ok, so what is your question for us?

My recommendation for something like this, is not to roll your own. Instead, reformat your data so you have a data set with:

Value, UCLM, LCML

Then you can use a single formula to create your values. Then if you need a different structure, reformat your data again.

Re: Dynamic Catx Macro for % and CIs

Astounding — Wed, 31 Oct 2018 20:36:31 GMT

Clearly, you have started off on the wrong foot. Consider this statement:

array percent {*} percent: ;

It appears that the intent is to take all variable names that contain "percent" somewhere in the name. But that's not what it does. It actually takes all variable names that begin with "percent" and it appears that you don't have any.

I agree with Reeza that you would make the problem a lot simpler by using a different structure for your data.

Re: Dynamic Catx Macro for % and CIs

delgoulding — Wed, 31 Oct 2018 21:45:05 GMT

Hi! Thank you for your responses! Let me clarify my question. I am not very familiar with arrays and have very large dataset that I am trying to put together the percentage with CIs.

So my first step would be renaming my variables to having matching prefixes. However, I am unsure how to more forward after that. I have the code below, but I can only get var1 to output once and the variable I want never outputs.

data want; set have;

rename a_percent= percent_a;
rename a_low=low_a;
rename a_high=high_a;
rename b_percent= percent_b;
rename b_low=low_b;
rename b_high= high_b;
run;

%let outcomes= a b;

data want2; set want;

array percent {*} percent:;
array low {*} low:;
array high {*} high:;
array stat {*} &outcomes;

do i= 1 to dim(percent);
do j= 1 to dim(low);

var1=catx("^n(", put(percent(i), percent20.3), put(low(j), comma20.3));

end;
end;

do k= 1 to dim(high);
do s= 1 to dim(stat);
stat(s)=catx(", ", var1, put(high(k), comma20.3))||")";
end;
end;

run;

Thank you in advance! I really appreciate any feedback/help!

Re: Dynamic Catx Macro for % and CIs

Reeza — Wed, 31 Oct 2018 22:14:13 GMT

Go back 5 steps. How did you get your data? What does the input data look like? What do you want as output?