In a regression procedure, how to suppress some estimated coefficients?

[sasgenie]

I am using HPLOGISTIC and HPREG to analyze a model with thousands of fixed effects as control variables. They are not the independent variables that I am interested in. Now the output has thousands of rows in the estimated coefficient report table. Is there anyway to prevent these fixed effects from showing up? 

A sample code would go as follows, where "fixed" has thousands of levels. I am only interested in the coefficients of x1 and x2, but not fixed.

PROC HPREG data=test;

    CLASS fixed;

    MODEL y = x1 x2 fixed;

RUN;

[Rick_SAS · Accepted solution]

1. Use ODS EXCLUDE ParameterEstimates to suppress the printed table.
2. Use ODS OUTPUT ParameterEstimates=PE to store the table as a SAS data set
3. Use PROC SGRENDER to display the table by using the same template that PROC HPREG uses. You can use a WHERE statement to exclude any estimates that you don't want.

PROC HPREG data=test;
    CLASS fixed;
    MODEL y = x1 x2 fixed;
    ods exclude ParameterEstimates;
    ods output ParameterEstimates=PE;
RUN;

proc sgrender data=PE template=HPSTAT.HPREG.ParameterEstimates;
where Effect ^= "fixed";
run;

[Rick_SAS]

I wrote a short blog post that describes, in general, how to use a built-in ODS template to display a table of values:

"Use built-in ODS templates to display SAS tables"