Performance issue regarding EF 4.1, code first and TPH - entity-framework

Question

I'm trying to create a datamodel using EF 4.1 and code first.
My Poco classes looks like this:
public abstract class AttributeBase
{
public int Id { get; set; }
public string Name { get; set; }
public string AttributeDescription { get; set; }
public int Length { get; set; }
public bool AllowNull { get; set; }
public bool DisplayInWeb { get; set; }
}
public class StringListAttribute : AttributeBase
{
public ICollection<StringValue> Values { get; set; }
public StringListAttribute()
{
Values = new List<StringValue>();
}
}
public class StringValue
{
public int StringValueId { get; set; }
public string Value { get; set; }
}
The basic idea is that StringListAttribute inherits from the AttributeBase class (other classes also inherits from the AttributeBase class, but I didn't include them in this post)
The StringListAttribute class have a collection of StringValues. (a zero-to-many relationship). What I try to achieve in my MyContextExtention class is that whenever I want a list of AttributeBase objects from the database, I want to get all the StringListAttribute objects and to populate their StringValue collection in one go. I don't want any lazy loading. I want to retrieve as much data I can in as few queries as possible.
My Context class inherits from DbContext and I created an extension method on my BraArkivContext class
public class MyContext : DbContext
{
public DbSet<AttributeBase> Attributes { get; set; }
public MyContext()
{
Configuration.LazyLoadingEnabled = false;
}
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
modelBuilder.Entity<StringListAttribute>()
.HasMany(t=>t.Values).WithOptional().WillCascadeOnDelete();
base.OnModelCreating(modelBuilder);
}
}
public static class MyContextExtension
{
private const string Values = "Values";
public static List<AttributeBase> AttributesWithData1(this MyContext context)
{
var attributeBases = new List<AttributeBase>();
var stringListAttributes = context.Attributes.Include(Values).OfType<StringListAttribute>().ToList();
attributeBases.AddRange(stringListAttributes);
return attributeBases;
}
}
My test project (I'm using nUnit) )is nothing much to write home about and contains these methods:
public class TestHelper
{
public static StringListAttribute CreateStringListAttribute(string name, int listSize)
{
var listAttribute = new StringListAttribute();
listAttribute.Name = name;
for (int i = 0; i < listSize; i++)
{
var attr1 = new StringValue();
attr1.Value = String.Format("StringValue_{0}", i);
listAttribute.Values.Add(attr1);
}
return listAttribute;
}
}
[Test]
public void AttributeBase_GetAllAttributesWithData1()
{
//First add some data to the database
var numberOfObjects = 100;
var name = "StringListAttribute_ReadMultipleStringsListsFromDb_" + Guid.NewGuid().ToString();
var listAttributes = TestHelper.CreateStringLists(name, numberOfObjects);
using (var context = new MyContext())
{
foreach (var stringListAttribute in listAttributes)
{
context.Attributes.Add(stringListAttribute);
}
context.SaveChanges();
}
Stopwatch sw = new Stopwatch()
var attributes = new List<AttributeBase>();
using (var context = new MyContext())
{
sw.Start();
attributes = context.AttributesWithData1().ToList();
sw.Stop();
}
Debug.WriteLine(String.Format("Total time for {0} attributes is: {1}.", attributes.Count, sw.Elapsed));
Assert.IsNotNull(attributes);
}
When running the test method "AttributeBase_GetAllAttributesWithData1" The sql statement that is executed looks like this:
SELECT
[Project1].[Id] AS [Id],
[Project1].[C1] AS [C1],
[Project1].[Name] AS [Name],
[Project1].[AttributeDescription] AS [AttributeDescription],
[Project1].[Length] AS [Length],
[Project1].[AllowNull] AS [AllowNull],
[Project1].[DisplayInWeb] AS [DisplayInWeb],
[Project1].[Document_Id] AS [Document_Id],
[Project1].[C2] AS [C2],
[Project1].[StringValueId] AS [StringValueId],
[Project1].[Value] AS [Value],
[Project1].[StringListAttribute_Id] AS [StringListAttribute_Id]
FROM ( SELECT
[Extent1].[Id] AS [Id],
[Extent1].[Name] AS [Name],
[Extent1].[AttributeDescription] AS [AttributeDescription],
[Extent1].[Length] AS [Length],
[Extent1].[AllowNull] AS [AllowNull],
[Extent1].[DisplayInWeb] AS [DisplayInWeb],
[Extent1].[Document_Id] AS [Document_Id],
'0X0X' AS [C1],
[Extent2].[StringValueId] AS [StringValueId],
[Extent2].[Value] AS [Value],
[Extent2].[StringListAttribute_Id] AS [StringListAttribute_Id],
CASE WHEN ([Extent2].[StringValueId] IS NULL) THEN CAST(NULL AS int) ELSE 1 END AS [C2]
FROM [dbo].[AttributeBases] AS [Extent1]
LEFT OUTER JOIN [dbo].[StringValues] AS [Extent2] ON [Extent1].[Id] = [Extent2].[StringListAttribute_Id]
WHERE [Extent1].[Discriminator] = 'StringListAttribute'
) AS [Project1]
ORDER BY [Project1].[Id] ASC, [Project1].[C2] ASC
When I run this sql statement in SQL Server Management Studio it's quite fast, but the Entity Framework is taking a very long time to process the result and return the poco classes populated with data.
Is there any other way to model the AttributeBase, StringListAttribute and StringValue classes and the relation between them to decrease the processing time in EF?