SQL –在一个表中查找所有不存在链接记录的记录

Question

I have a slight twist on the other “in one table but not another” threads I've seen.

I have 3 tables:

ProductFeed
    ProductFeedID INT
    ProductName NVARCHAR(30)

Product
    ProductID INT
    ProductFeedID INT
    StorefrontID INT

Storefront
    StorefrontID INT
    StorefrontName NVARCHAR(30)

• The ProductFeed table has one unique product per record.
• The Product table can have multiple of the same ProductFeedID, but can only have 0 or 1 unique ProductFeedID record per storefront.
• The Storefront table has 1 record per storefront.

What I wish to do is a query showing all ProductFeed records, with the StorefrontID and StorefrontName, which DO NOT contain a Product record for that storefront. So something like this:

ProductFeedID   ProductName    StorefrontID     StorefrontName
   123             iPod             1              MyStore1
   123             iPod             4              MyStore4
   234             TShirt           2              MyStore2
   234             TShirt           4              MyStore4
   345             Coffee Mug       5              MyStore5
   etc.

I'm using SQL Server 2012. Can someone help?

Answer 1

If I understand correctly, you can approach this problem is by generating all possible combinations of product feeds and store fronts, then use left join . The combinations that fail to match are the ones you want:

select pf.*, sf.*
from productfeed pf cross join
     storefront sf left join
     product p
     on p.ProductFeedID = pf.ProductFeedID and
        p.StoreFrontId = pf.StoreFrontId
where p.ProductId is null

Answer 2

-- I think this is more understandable.

    select pf.*, sf.*
    from productfeed pf  
    cross join storefront sf 
    where not exists (select 1 
                      from product p
                      where p.ProductFeedID = pf.ProductFeedID
                      and p.StoreFrontID = sf.StoreFrontID)