Outputting by sum with proc freq

[js5]

Hello,

I have the following code:

proc freq data=inputds noprint;
	by notsorted a b c;
	tables x*y*z/out=freq;
	weight count/zeros;
run;

proc means data=inputds completetypes noprint nway;
	by notsorted a b c;
	var count;
	output out=means(drop=_:) sum=count_sum_c;
run;

data final;
	merge freq means;
	by notsorted a b c;
run;

Is it possible to produce the dataset in one single step? having a notsorted means that the merge is not possible, and due to the fact that the input and freq datasets contain over 90 million rows, I would prefer not to have to sort anything unless absolutely unavoidable. Thank you for your tips in advance.

[Tom]

Without data or even a description of what you are trying to do it is going to be hard.

Why not just let PROC MEANS add up the WEIGHT variable for you (especially since you want to include the zero weight observations)?

proc means data=inputds completetypes noprint nway;
  class a b c;
  output out=means(drop=_:) n=nobs sum(count)=count_sum_c;
run;

[ballardw]

You might describe what you are wanting as output and do you actually need a data set? If the data set is not used for anything other than reporting it may be that a report procedure is the way to go.

And providing data in the form of a working data step usually helps.

Provide a subset and what you expect as a final result for that subset of data.

I am a bit concerned that you may not be getting what you want as output to begin with by using not sorted.

IF two data sets have the same number of observations you might try as without a BY then SAS does a line by line merge regardless of values. The output of the two data sets may be in the same order because of the way SAS process notsorted data.

data final;
	merge freq means;
run;