output out using proc means

[Barkamih]

Hi all 

I want to know what is wrong with this code, I'm quite sure this code is correct but it may be the issue because of the big sample size, I have 1.6 million observation. So, What is the best idea to solve this problem. 

when I submit this code, after few minutes the system showing proc means not responding.  

regards 

Ibrahim

proc sort data = have;
	by year month week;
	run; quit;

	proc means data = have;
	by year month week;
	output out = want sum(FATKG)= FAT_KG SUM(PROTKG)= PROT_KG SUM(MILKKG)= MILK_KG;
	RUN; QUIT;

[RW9]

1.6million observations will take a long time to run, there is no workaround unless you can reduce the number of observations.  Just leave it until its done.  

[Reeza]

You may want to suppress the output though - use NOPRINT. Generating that type of output can take a lot of time. If it sorted fine though, I'd expect PROC MEANS to run fine. And 1.6 million rows should complete in under 5 minutes in my experience (SAS UE is still less than a minute for me).

If you have a lot of variables, you can try dropping the variables you don't need in a data set option to make it faster. Not sure if that helps but worth a shot.

[Astounding]

It's not clear how many numeric variables exist in your data set.  Without a VAR statement, the default is to compute statistics for every numeric variable.  PROC MEANS might be "intelligent" enough to figure out that you only need statistics computed for 3 variables ... but it might not figure that out.  You might fix the problem just by adding to PROC MEANS:

var fatkg protkg milkkg;

[Reeza · Accepted solution]

I think @Astounding hit the nail on the head there. If you don't specify NOPRINT or VAR it will analyze all numeric variables and put that in the results window - and that will take time. So adding NOPRINT & a VAR statement should speed things up a lot. 

proc sort data = have;
by year month week;
run; 

proc means data = have NOPRINT;
by year month week;
var FATKG PROTKG MILKKG;
output out = want sum(FATKG)= FAT_KG SUM(PROTKG)= PROT_KG SUM(MILKKG)= MILK_KG;
RUN; 

---

@Astounding wrote:

It's not clear how many numeric variables exist in your data set.  Without a VAR statement, the default is to compute statistics for every numeric variable.  PROC MEANS might be "intelligent" enough to figure out that you only need statistics computed for 3 variables ... but it might not figure that out.  You might fix the problem just by adding to PROC MEANS:

var fatkg protkg milkkg;

---

[Barkamih]

 I appreciate your help my  SAS friends HAHAHAH,  now everything is fine and I got what I'm looking for. 

thank you very much again 

Ibrahim 

regards