Reply made to same question as cross posted to stackoverflow...
Use a `BY` statement to independently process like grouped subsets of a data set. Use a `WHERE` statement to select the subset(s) to process.
Example:
ods html file='output.html' style=plateau;
proc sort data=sashelp.cars out=cars_sorted;
by origin;
run;
title; footnote;
options nocenter nodate nonumber;
PROC TABULATE DATA=cars_sorted;
by origin;
where Origin in ("Asia", "Europe", "USA");
where also make >= 'P'; * further subset for reduced size of output screen shot;
CLASS Make DriveTrain ;
TABLE (Make), (DriveTrain) / nocellmerge ;
run;
ods html close;
Alternatively, use a `TABLE` statement of form `<page dimension>,<row dimension>,<column dimension>` in lieu of `BY` group processing. Such a form does not need presorted data because the <page dimension> is constructed from `CLASS` variables.
Example:
PROC TABULATE DATA=sashelp.cars; /* original data, not sorted */
where Origin in ("Asia", "Europe", "USA");
where also make >= 'P'; * further subset for reduced size of output screen shot;
CLASS Origin Make DriveTrain ; /* Origin added to CLASS */
TABLE Origin, (Make), (DriveTrain) / nocellmerge ; /* Origin is page dimension */
run;
Output BY statement
Output PAGE dimensioned table
... View more
