02-10-2014 03:28 AM
Could you please tell me how to count a variable only if certain criteria is met in the proc sql syntax. For example, I want to count both people and Hispanic people. So I coded the following way, but the calculated p_count and chm_count is the same. Also, if I want to count rows in which certain variable called sex is F, could I go for count(sex = "F")?
My old but stupid way is to use proc sql two time with the where statement and then join these two tables. That is not cool.
create table ipums2 as
select distinct year, statefip, county, count(pid) as p_count, count(race=4) as hs_count
group by year, statefip, county
order by year, statefip, county;
02-10-2014 05:11 AM
I don't know what your final result should be (in what form). In many cases, just adding your column to the group by clause is sufficient. If you want specific grouping, you can hold these in control tables (and apply them using SAS formats or SQL join). Adding specific criteria as a basic technique can lead to hard coded programs which lead higher maintenance costs.
Having that said, you could combine case with aggregation functions:
sum(case race when 4 then 1 else 0 end) as hs_count
Need further help from the community? Please ask a new question.