choosing row (from group) with max value in a SQL Server Database - sql-server

Question

I have a large database and am putting together a report of the data. I have aggregated and summed the data from many tables to get two tables that look like the following.
id | code | value id | code | value
13 | AA | 0.5 13 | AC | 2.0
13 | AB | 1.0 14 | AB | 1.5
14 | AA | 2.0 13 | AA | 0.5
15 | AB | 0.5 15 | AB | 3.0
15 | AD | 1.5 15 | AA | 1.0
I need to get a list of id's, with the code (sumed from both tables) with the largest value.
13 | AC
14 | AA
15 | AB
There are 4-6 thousand records and it is not possible to change the original tables. I'm not too worried about performance as I only need to run it a few times a year.
edit:
Let me see if I can explain a bit more clearly, imagine the id is the customer id, the code is who they ordered from and the value is how much they spent there.
I need a list of the all the customer id's and the store that customer spent the most money at (and if they spent the same at two different stores, put a value such as 'ZZ' in for the store name).

Answer 1

try this:
DECLARE #Table1 table (id int, code char(2), value decimal(5,1))
INSERT #Table1 VALUES (13 , 'AA' , 0.5)
INSERT #Table1 VALUES (13 , 'AB' , 1.0)
INSERT #Table1 VALUES (14 , 'AA' , 2.0)
INSERT #Table1 VALUES (15 , 'AB' , 0.5)
INSERT #Table1 VALUES (15 , 'AD' , 1.5)
DECLARE #Table2 table (id int, code char(2), value decimal(5,1))
INSERT #Table2 VALUES (13 , 'AC' , 2.0)
INSERT #Table2 VALUES (14 , 'AB' , 1.5)
INSERT #Table2 VALUES (13 , 'AA' , 0.5)
INSERT #Table2 VALUES (15 , 'AB' , 3.0)
INSERT #Table2 VALUES (15 , 'AA' , 1.0)
SELECT
dt.id, MAX(dt.code) AS code, sum(dt.value) as value
from (select id, code, value
from #Table1
UNION ALL
select
id, code, value
from #Table2
) dt
group by dt.id
order by id
OUTPUT:
id code value
----------- ---- ---------------------------------------
13 AC 4.0
14 AB 3.5
15 AD 6.0
(3 row(s) affected)
I'm not sure what you are after? this is the MAX code per id summing the value. if this is not what you are after please specify i nthe question more clearly
EDIT after OP's edit, using same tables as code from above:
;WITH AllTAbles AS
(select
id, code, value
from #Table1
UNION ALL
select
id, code, value
from #Table2
)
, MaxValues AS
(SELECT
dt.id, MAX(dt.value) as MaxValue, SUM(dt.value) AS SumValue
from AllTAbles dt
group by dt.id
)
, StoreCount AS
(SELECT
a.id,a.Code, COUNT(*) AS StoreCount
FROM AllTAbles a
INNER JOIN MaxValues m ON a.id=m.id AND a.value=m.MaxValue
GROUP BY a.id,a.Code
)
SELECT
s.id
,CASE
WHEN s.StoreCount=1 THEN s.Code
ELSE 'ZZ'
END AS code
,m.SumValue
FROM StoreCount s
INNER JOIN MaxValues m ON s.id=m.id
ORDER BY s.id
OUTPUT:
id code SumValue
----------- ---- ----------
13 AC 4.0
14 AA 3.5
15 AB 6.0
(3 row(s) affected)
OP doesn't say the version of SQL Server, so here is a pre SQL Server 2005 version that does not use CTEs, has same output as the CTE version above:
SELECT
s.id
,CASE
WHEN s.StoreCount=1 THEN s.Code
ELSE 'ZZ'
END AS code
,s.SumValue
FROM (SELECT
a.id,a.Code, COUNT(*) AS StoreCount, m.SumValue
FROM (select
id, code, value
from #Table1
UNION ALL
select
id, code, value
from #Table2
) a
INNER JOIN (SELECT
dt.id, MAX(dt.value) as MaxValue, SUM(dt.value) AS SumValue
from (select
id, code, value
from #Table1
UNION ALL
select
id, code, value
from #Table2
) dt
group by dt.id
) m ON a.id=m.id AND a.value=m.MaxValue
GROUP BY a.id,a.Code,m.SumValue
) s
ORDER BY s.id

Answer 2

select id, code, sum(value) as value
from
(
select id, code, value
from yyy
UNION
select id, code, value
from zzz
) aaa
group by id, code
order by sum(value)
or this removing id from the grouping:
select code, sum(value) as value
from
(
select id, code, value
from yyy
UNION
select id, code, value
from zzz
) aaa
group by code
order by sum(value)