Question

I have this expression:

var result = from pav in ProductAttributes
           join id in valueIds
           on pav.AttributeValueID equals id
           select pav.ProductImageID;

which works up to a point. The issue is that the collection ProductAttributes contains the same product many times, for each attribute. It’s structure is:

ID – unique

ProductID

ProductAttributeValueID

ProductImageID

So a Product may appear many times in the collection. I want the result to actually filter OUT all the products that DON’T have any matches at all in valueIds (which is a list of ProductAttributeValueIDs).

So I want to ONLY return products that have ALL of the COMBINED valueIds, not just ANY of them, which is what the above linq expression is doing.

PS I can post SQL code that shows what I mean in SQL if that helps!

@devgeezer posted an answer which was close enough but it only worked for one value.

I ended up with the code below, which works. I group on the ProductID, then use that in a 2nd query to filter the original collection:

        var result = 
            from pav in ProductAttributeValues
            join id in valueIds
            on pav.AttributeValueID equals id
            group pav by pav.ProductID into gj
            where gj.Count() == valueIds.Count()
            select gj.Key;

        var imageIds = from pav in ProductAttributeValues
                       join id in result
                       on pav.ProductID equals id 
                       select pav.ProductImageID;

Answer 1

You might try a group and filter approach something like the following:

var result = 
    from pav in ProductAttributes
    join id in valueIds
    on pav.AttributeValueID equals id
    group pav by pav.ProductImageID into gj
    where gj.Count() == valueIds.Count()
    select gj.Key;