topic Group by sum in proc summary in Statistical Procedures

Group by sum in proc summary

BrahmanandaRao — Wed, 24 Mar 2021 06:09:41 GMT

proc summary data=sashelp.prdsal2 print ;
class county month product actual_sales;
var  year;
output out= prdsales actual_sales sum=;
run;

Re: Group by sum in proc summary

qoit — Wed, 24 Mar 2021 06:18:52 GMT

I am quite unclear regarding what you are trying to achieve here. OUT= option of the OUTPUT statement is meant to output the created data set by the SUMMARY procedure. Adding to it, the "ACTUAL_SALES" variable does not exist in the data set sashelp.prdsal2, and therefore the error message. Use the below, or better yet, the SQL procedure.


proc means data=sashelp.prdsal2 nway sum;
class county month product;
var  actual;
output out= prdsales (drop=_type_ _freq_) sum= / autoname;
run;

Re: Group by sum in proc summary

BrahmanandaRao — Wed, 24 Mar 2021 06:39:38 GMT

Actual sales variable in prdsal2 dataset why it shows error variable not found

Re: Group by sum in proc summary

PeterClemmensen — Wed, 24 Mar 2021 06:42:01 GMT

What is the question here?

Re: Group by sum in proc summary

Kurt_Bremser — Wed, 24 Mar 2021 09:21:37 GMT

@BrahmanandaRao wrote:
Actual sales variable in prdsal2 dataset why it shows error variable not found

The variable is positively NOT there:

proc sql;
select name from dictionary.columns
where libname ="SASHELP" and memname = "PRDSAL2";
quit;

Result:

Column Name
COUNTRY
STATE
COUNTY
ACTUAL
PREDICT
PRODTYPE
PRODUCT
YEAR
QUARTER
MONTH
MONYR

You may have meant to use ACTUAL.

When SAS tells you something ain't there, while you think otherwise, then it's always YOU who's in error.

Re: Group by sum in proc summary

ballardw — Wed, 24 Mar 2021 14:08:10 GMT

A random variable name on the OUTPUT statement is going to throw an error if not used properly.

In your picture the variable name underlined has no context.

All those "expected" are Statistics or the option specifier / . Statistics, associated variables and variable names are expected. But at name without context in one of those places is invalid.

If you are creating a statistic for a specific variable you use statistic(variablename(s))= <name of new variable(s) if desired