对于sql server

WITH CTE AS(

    SELECT 
        *
    FROM (
        VALUES  
            ('2018-08-23', 'A', 10),
            ('2018-08-23', 'B', 10),
            ('2018-08-24', 'A', 14),
            ('2018-08-31', 'A', 8),
            ('2018-08-31', 'B', 10)
        ) as list (Date_, Item, Amount)
)

select 
    WKnum, MIN(Date_), MAX(Date_), Item, AVG(Amount)    
from(
    select 
        *,
        DATEPART(WK, Date_) WKnum
    from CTE
    ) as A
group by WKnum, Item

我知道我参加聚会迟到了，但万一有人在找我，这也许行得通：

select date_part('week', transaction_date) week_of_the_year, item_id,
    avg(sold_qty) from your_table group by 1, 2 order by 1

尝试下面的查询：它将在sql server上工作

select datepart(week,DateOfPurchase),avg(item)
from
(
select DateOfPurchase,count(SoldItemID) item
from tablename group by DateOfPurchase) a
group by datepart(week,DateOfPurchase)