SQL Sum date from multiple rows, Sql Server

Language: Layout:

DECLARE @sample TABLE(
    idnum nvarchar(4), 
    itemid nvarchar(10),
	itemnum nvarchar(10), 
    poolnum nvarchar(2), 
    startdate datetime, 
    enddate datetime	
  )

INSERT INTO @sample
select '1001','1030-B','P00001','PC','2016-11-26 14:13:00.000','2016-11-26 20:50:00.000'
union 
select '1001','1030-B','P00002','RP','2016-11-28 15:02:00.000','2016-11-29 10:00:00.000'
union
select '1001','1030-B','P00004','PL','2016-11-28 20:02:00.000','2016-11-28 21:06:00.000'
union 
select '1001','1030-B','P00005','PC','2016-12-01 19:59:53.000','2016-12-05 11:35:44.000'
union
select '1002','1030-B','P00006','PC','2017-01-20 13:46:11.000','2017-01-21 03:12:38.000'

;WITH aCTE
AS
(   select DISTINCT
         idnum
        ,FIRST_VALUE(startDate) OVER (PARTITION BY idNum 
                                      ORDER BY startdate ASC
                                      ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS first_StartDate
        ,FIRST_VALUE(enddate) OVER (PARTITION BY idNum 
                                      ORDER BY enddate DESC
                                      ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS first_EndDate
	from @sample
	where poolnum='PC'
)

SELECT
    idnum
    ,first_StartDate
    ,first_EndDate
    ,DATEDIFF(DD,first_StartDate,first_EndDate) AS TT
    ,CAST(first_EndDate - first_StartDate as decimal(10,2)) AS calculatedTT
FROM
    aCTE

View schema
Execution time: 0,02 sec, rows selected: 2, rows affected: 5, absolute service time: 0,19 sec