Since most of those variables are basically categories by aggregate I have to assume count.
Do you need a data set for further manipulation or a report for people to read?
If you need a data set you need to provide some example of what the result would look like and rules.
I might start with something like:
Proc tabulate data=have;
class zipcode county_name state_name age_cat;
/* row based count*/
table zipcode* (state_name county_name age_cat),n;
/* crosstab */
table state_name county_name age_cat,
zipcode*n
;
/* nested cross tab */
table zipcode ,
state_name*county_name*n
;
/* ugly and quite possibly not very useful */
table zipcode*age_cat,
state_name * county_nam *n
;
run;
First warning with Proc Tabulate, any record missing one or more of the Class variables is dropped from the summary by default. If you have missing values and want that level reported you add a /missing; option to Class statement with the variable(s) that may have missings.
You can dump a LOT of variables into a table statement, and many statistics, and the result may take some time to appear in a results window.
There are many other statistics you can request. Every where there is an N you could request percentages with Pctn, RowPctn, Colpctn, Reppctn or PagePctn.
The Comma separates dimensions: page (if there are 3), row, column. The * requests a nest, that means each level of the second variable or statistic follows each level of the first
If you have variables where requesting a mean, max, min, sum etc. would make sense, Those go on a Var statement and then request the statistics.
There are many appearance options. This is a complex enough procedure SAS publishes a book just on Proc Tabulate.
