topic Syntax to Print only Categorical variables only in a data set? in SAS Programming

Syntax to Print only Categorical variables only in a data set?

Guru_MG — Mon, 17 Aug 2020 12:10:33 GMT

Hi There,

what is the syntax to print only categorical variables in a dataset?

For eg: Syntax to print only numeric variables in a dataset is

proc print data=sashelp.cars;
var _numeric_;
run;

Re: Syntax to Print only Categorical variables only in a data set?

PaigeMiller — Mon, 17 Aug 2020 12:16:59 GMT

Use

var _character_;

Re: Syntax to Print only Categorical variables only in a data set?

Tom — Mon, 17 Aug 2020 12:40:42 GMT

SAS has two variable types. Floating point numbers and fixed length character strings. Whether a variable is categorical or not is something that SAS does not know. You would have to tell it.

The _NUMERIC_ variable list includes all of the numeric variables (whether they are "categorical" or not). The _CHARACTER_ variable list includes all of the character variables (whether they have defined categories or are free text).

Re: Syntax to Print only Categorical variables only in a data set?

Guru_MG — Mon, 17 Aug 2020 12:44:05 GMT

Thank you very much Pagiemiller 🙂

Re: Syntax to Print only Categorical variables only in a data set?

ballardw — Mon, 17 Aug 2020 14:46:33 GMT

Since SAS doesn't know which variables are 'categorical' you might look into the Proc freq NLEVELS option

ods select nlevels;
proc freq data=sashelp.class nlevels;
run;

As an example. The ODS SELECT option will have Proc Freq only display the NLEVELS table and not every single value of every single variable in the data set.

This shows the number of different values a variable has. In this case AGE, which is numeric, shows 6 values and might be considered "categorical". Name, which has 19 values in 19 records seems like less a "category" than "free text". Sex, character with 2 levels, likely would be a category.

The common meanings of the variable name in the SASHELP.CLASS data set may be sufficient but this technique may be helpful exploring a new data set.