Is this what you are looking for? I didn't find any cube related information in your question, so I did this in straight SQL..

DECLARE @Input TABLE

(

pn VARCHAR(10),

prcode VARCHAR(10),

dos DATE,

Expected DECIMAL(10,2),

Payment DECIMAL(10,2)

)

INSERT INTO @Input VALUES('MS0000003', 'PT001', '1/2/2013', 29.72, 0), ('MS0000003', 'PT001', '1/21/2013', 57.1, 0)

,('MS0000003', 'PT002', '6/7/2013', 26.69, 0),

('MS0000003', 'PT002', '6/7/2013', 89.16, 0),

('MS0000003', 'PT002', '6/6/2013', 57.1, 0),

('MS0000003', 'PT002', '6/7/2013', 12.28, 0),

('MS0000003', 'PT002', '6/7/2013', 26.69, 26.69),

('MS0000003', 'PT002', '6/7/2013', 89.16, 77.16),

('MS0000003', 'PT002', '6/17/2013', 57.1, 57.1)

;WITH CTE AS

(

SELECT pn, prcode, MAX(dos) AS Max_dos, MIN(dos) AS min_dos, SUM(expected) as expected, SUM(payment) AS payment

FROM @Input

GROUP BY pn, prcode

)

SELECT a.pn, a.prcode, Max_dos, DATEDIFF(d, min_dos, max_dos) AS differnce, a.Expected, a.Payment

FROM CTE a