SQL SERVER 2008. How to select one row only based on dates - sql-server-2008

Question

Select MIN([appdate]), CustID, CustAtt, Surname, Firstname, itemType
From table
Where appDate > (getdate())
Group by CustID, CustAtt
CustAtt CustID Surname Firstname appDate itemType
53247 20675 A AX
49535 12/FX08 B BX 14/08/2017 solid
70433 400039 C CX
67119 413555 D DX
51406 27/EY07 E EX 14/07/2017 Liquid
51406 27/EY08 E EX 13/09/2017 Gas
51406 27/EY09 E EX 11/12/2017 Solid A
51406 27/EY10 E EX 06/06/2018 Liquid A
82820 410053 F FX
52395 29/FA72 G GX 25/09/2017 Gas A
89488 414282 H HX
55855 412799 I IX 30/08/2017 Solid
55855 412799 I IX 21/08/2017 Liquid
53248 16/EK15 J JX 06/07/2017 Gas
53248 16/EK15 J JX 17/07/2017 Solid B
89835 911528 K KX 08/05/2018 Solid B
The above is a snippet of my output from the above script, but isn't giving me the desired output. I want to only select one row for each CustID & CustAtt from my dataset based on the appdate column.
The desired output should include the next earliest date from current date, in the case the dates are the same any given row is acceptable:
The output table should yield:
CustAtt CustID Surname Firstname appDate itemType
53247 20675 A AX
49535 12/FX08 B BX 14/08/2017 solid
70433 400039 C CX
67119 413555 D DX
51406 27/EY07 E EX 14/07/2017 Liquid
82820 410053 F FX
52395 29/FA72 G GX 25/09/2017 Gas A
89488 414282 H HX
55855 412799 I IX 21/08/2017 Liquid
53248 16/EK15 J JX 06/07/2017 Gas
89835 911528 K KX 08/05/2018 Solid B

Answer 1

Example: http://sqlfiddle.com/#!6/c35e2/1
create table t(
id int,
x int,
y int,
z int,
a int,
b int,
c datetime
);
-- Build t table
INSERT INTO t values
(1, 1, 2, 3, 4, 5, getdate()),
(2, 1, 2, 3, 4, 5, dateadd(hh,+3,getdate())),
(3, 1, 2, 3, 4, 5, dateadd(hh,+1,getdate())),
(4, 1, 2, 3, 4, 5, dateadd(hh,+2,getdate())),
(5, 1, 2, 3, 4, 5, dateadd(hh,-2,getdate())),
(6, 1, 2, 3, 4, 5, dateadd(hh,-1,getdate())),
(7, 1, 2, 3, 4, 5, dateadd(hh,+1,getdate())),
(8, 1, 2, 3, 4, 5, dateadd(hh,+1,getdate())),
(9, 1, 2, 3, 4, 5, dateadd(hh,+1,getdate()));
select top 1 t.id, s.*, t.c
from t
join (
select x,y,z,a,b
from t
group by x,y,z,a,b
having count(*) > 1
) s
on s.x = t.x
and s.y = t.y
and s.z = t.z
and s.a = t.a
and s.b = t.b
and (getdate() < t.c)
order by t.c

Answer 2

You have to use HAVING keyword. It can filter "after group".
SELECT
MIN([appdate]),
id,
type
FROM
[table]
WHERE
appdate > GETDATE()
GROUP BY
id, type
HAVING
COUNT(*) > 1 -- select only records, that have duplicates, based on our "group by"
ORDER BY
id
added 12.07.17:
(for revised question)
Count row number (partion by CustID & CustAtt, order by appDate) and then select first one for each one.
;WITH cte AS (
SELECT
*,
Row_number() OVER(
PARTITION BY CustID, CustAtt ORDER BY appDate
) row_number
FROM
table
)
SELECT
*
FROM
cte
WHERE
row_number = 1