Question

So I’ve been told that looping is not a desirable thing to do in SQL. The following question makes a histogram based on age ranges. As you can see, all of the groups are hardcoded. In this instance all of the values for which I’m looking are under 100. What to do if these ranged into the thousands? How could this be made more extensible?

select bins, count(*) as numbers from
(

select id, patientage,
case 
    when patientage between 20 and 29 then '20-29' 
    when patientage between 30 and 39 then '30-39'
    when patientage between 40 and 49 then '40-49'
    when patientage between 50 and 59 then '50-59'
    when patientage between 60 and 69 then '60-69'
    when patientage between 70 and 79 then '70-79'
    when patientage between 80 and 89 then '80-89'
    when patientage between 90 and 99 then '90-99'


end as bins
from patient
inner join tblhospitals on tblhospitals.hospitalnpi=patient.hospitalnpi
where (tblhospitals.hospitalname like '%university%')
) as t
group by bins
order by bins

Answer 1

Instead of your case statement, you could instead group by patientage / 10 For instance, if you rewrote your query as:

SELECT bins, COUNT(*) AS numbers
FROM (
    SELECT id, patientage, patientage / 10 AS bins
    FROM patient P
        INNER JOIN tblhospitals H ON H.hospitalnpi = P.hospitalnpi
    WHERE H.hospitalname LIKE '%university%'
) AS T
GROUP BY bins
ORDER BY bins

You will get the same data, except instead of ’50-59′ you will get 5, which is really just a display issue that can be handled outside of SQL. (Or if you really must, you could do some massaging of the data for display purposes like casting bins * 10 and bins * 10 + 9 to varchar etc)