select count(*)  [Population],DG.GroupAge from (select ms.dob,case when datediff(YEAR,ms.dob,getdate())< 10 then 'G10'
when datediff(YEAR,ms.dob,getdate()) between 10 and 19 then 'G 10-20'
when datediff(YEAR,ms.dob,getdate())between 20 and 29 then 'G 20-30'
when datediff(YEAR,ms.dob,getdate())between 30 and 39 then 'G 30-40'
when datediff(YEAR,ms.dob,getdate())between 40 and 49 then 'G 40-50'
when datediff(YEAR,ms.dob,getdate())between 40 and 49 then 'G 50-60'
else 'G60+' end as GroupAge 
from people ms 
where ms.id in (select max(id) 
					 from dbo.table
					 group by customer_id) 
and ms.location in ('CO','TX')) as DG
Group by DG.GroupAge
Order by GroupAge

....
               ELSE 'G60+' END AS GroupAge 
               FROM people ms 
               INNER JOIN 
               (
               		SELECT MAX(id) AS MaxId
				   FROM   dbo.table
				   GROUP BY
						  customer_id
               ) s ON s.MaxId = ms.id
           WHERE ms.location IN ('CO', 'TX')

quote:

---

Originally posted by xhostx

Oh, I really thought of this because it may sometimes round up/down!
Is this is right?

Please any alternative?

Thank you

--------------------------
Joins are what RDBMS's do for a living

---

...
CASE WHEN MONTH(GETDATE()) < MONTH(ms.dob) 
       OR (MONTH(GETDATE())= MONTH(ms.dob)
       AND DAY(GETDATE()) < DAY(dob))
     THEN datediff(YEAR,ms.dob,getdate())-1
     ELSE datediff(YEAR,ms.dob,getdate())
END AS Age

...

SELECT
(
   CAST(CONVERT(CHAR(8), GETDATE(), 112) AS INT)
   - 
   CAST(CONVERT(CHAR(8), ms.dob, 112) AS INT)
)/10000