Filter the "Includes" table on Entity Framework query

This is for Entity Framework for .NET 3.5:

I have the need to query a table and include a collection of the "many" table of a one-to-many relationship. I'm trying to filter that collection as part of the query - I'm pretty new to Entity Framework, and I'm having trouble figuring it out.

Simplified example: Author has Books, and Book has an IsFiction column. I want a filtered list of authors, along with all fiction books.

Without the filter, it's easy:

var q = from a in db.Authors.Include("Books")
        where a.BirthYear > 1900
        select a;

I can filter after the fact, something like:

var fictionBooks = a.Books.Where(b => b.IsFiction);

But the problem is that the original query already ran, and included those results, which is unnecessary database processing.

I can query separately, like:

var q = from a in db.Authors where a.BirthYear > 1900 select a;
foreach (var a in q)
{
    var books = from b in db.Books 
                where ((b.Author.Id == a.Id) && (b.IsFiction))
                select b;
}

But of course that's one call for every author, which I want to avoid as well.

I can go backwards, like:

var allBooks = from b in db.Books.Include("Author")
    开发者_StackOverflow中文版           where b.IsFiction
               select b;

But then I'm back to the original problem, except now on the author side instead of the book side.

There must be a solution that encompasses everything - I can do it in SQL pretty easily:

select * from author a
left join book b on a.id = b.author_id and b.is_fiction = 1
where a.birth_year > 1900

Any suggestions?

The forward way:

var q = from a in db.Authors.Include("Books")
        where a.BirthYear > 1900
        select new {
            Author = a,
            FictionBooks = a.Books.Where(b => b.IsFiction)
        };

Another way, derived from the SQL at the bottom of your question:

var q = from a in db.Authors
        from b in db.Books.Include("Author")
        where a.BirthYear > 1900 && b.IsFiction && a.Id == b.Author.Id
        select new { Author = a, Book = b };

The main difference between these two is that the first one will basically give you a collection of authors plus the list of fiction books for each author (which may be empty); while the second will give you a collection of author/books pairs (so it doesn’t return any authors with no fiction books).