one-to-many efficient query

gistfile1.txt

CREATE TABLE [Foo] (
	[FooId] [int] NOT NULL,
	[SomeDateTime] [datetime2](7) NOT NULL,
	-- ... OTHER COLUMNS
	[Version] rowversion NOT NULL,
    CONSTRAINT [PK_Foo] PRIMARY KEY CLUSTERED ([FooId])
)

CREATE TABLE [FooBar] (
	[FooBarId] [int] NOT NULL,
	[FooId] [int] NOT NULL,
	[BarId] [int] NOT NULL
    CONSTRAINT [PK_FooBar] PRIMARY KEY CLUSTERED ([FooBarId])
)

CREATE NONCLUSTERED INDEX [IX_SomeDateTime] ON [Foo] ([SomeDateTime] DESC) INCLUDE ([version])
CREATE UNIQUE NONCLUSTERED INDEX [UX_FooId_BarId] ON [FooBar] ([FooId],[BarId])


;WITH selectFoos AS
(
	SELECT
		Foo.FooId,
		Foo.[Version],
		ROW_NUMBER() OVER(ORDER BY Foo.[DateTime] DESC) AS RowNumber
	FROM Foo
		LEFT OUTER JOIN FooBar ON FooBar.FooId = Foo.FooId
	WHERE Foo.[DateTime] >= '2012-01-02' AND Foo.[DateTime] <= '2012-01-03'
		AND (FooBar.BarId IS NULL OR FooBar.BarId = 1 OR FooBar.BarId = 2)
		and Foo.FooId = FooBar.FooId
	GROUP BY Foo.FooId, Foo.[Datetime], Foo.[Version]
)

SELECT
	FooId
	,[Version]
FROM selectFoos
WHERE RowNumber BETWEEN 50 AND 100
OPTION (RECOMPILE)