Exclude Id from Entity<TEntity> when using stored procedure

SORRY for "Exclude Id from Entity<TEntity>" Thread but there has been a problem.

If I want to use stored procedure I must to create a derived class (such as CustomClass) from Entity<TEntity> and automatically UnitOfWork.Find<CustomClass>("procedurename", procQueryInstance) will return entities with Id.
----------------
Stored procedure looks like this:

CREATE  PROCEDURE [dbo].[procedurename] @XYZvalue varchar(20) = NULL AS
SELECT ID, .., .., ..,
FROM tableName
WHERE XYZ = @XYZValue
--------------------

In some cases I need automatically entities without ID and that means I don`t need to use Entity<TEntity>, but I tried something like this:

I added a custom class derived from Entity:

1)

using............
namespace Mindscape.LightSpeed
{
    [Serializable]
    public abstract class EntityBase : Entity
    {      
        /// <summary>
        /// Initializes a new instance of the <see cref="EntityBase"/> class.
        /// </summary>
        protected EntityBase()
        {}      
        protected EntityBase(bool autoInitialize)
            : base(false)
        {}
        internal override void SetIdInternal(object value)
        {
            throw new NotImplementedException();
        }
        internal override object IdInternal
        {
            get { return 0; }
        }

        //modified as virtual method in abstract class Entity because I want to change format
        internal override string IdentifierString
        {
            get
            {
                return StringUtils.FormatInvariant(Resources.EntityToStringFormat, GetType().Name);
            }
        }      
    }
}

AND :
2)
 
protected override void VisitTypeModel(TypeModel typeModel, QueryExpression queryExpression, Order order, string aggregateName, bool includeDeleted, bool identifiersOnly)
{
 .........

 while(......)
 {
   //I modified "var id = TypeModel.IdFieldModel.GetValueFromReader          (_queryStrategy.DataReader); "
  //with:
    var id = new object();
    if (typeModel.IdFieldModel != null)       
       id = typeModel.IdFieldModel.GetValueFromReader (_queryStrategy.DataReader);

  ........................
 }
 .........
}


Results:
- EntityBase (my custom class) and Entity<TEntity> work together very good:
when I need entities without Id I`ll use such as CustomClass: EntityBase otherwise CustomClass: Entity<TEntity>

What you think about that? Is it a bad idea until the new version is released (maybe in next version (3.0) of LightSpeed Id is optional).

Have a nice day.
Thanks in advance.

We can't guarantee how this will work with LightSpeed.  Your EntityBase class is effectively opting out of the normal identity conventions.  This means that there will be no way to handle associations, to identify existing entities, to save updates, etc.  You may also have a problem because it looks like everything is getting Id 0; that will cause problems for the unit of work's identity map and if you perform a second query in the same UOW instance, then you may find yourself getting the "cached item with Id 0" back for each result.

Effectively what you have here is a way of loading isolated value types, rather than domain entities.  If this is what you want, a better idea may be to use projections.  You can't do a database-level projection over a stored procedure, but you can do it client side:

var addresses = unitOfWork.Find(new ProcedureQuery(...))
                                        .Select(c => new Address(c.HouseNumber, c.Street, c.Town));

Note that the stored procedure still has to return an ID column because this materialises the entities and then projects them into the value type.

If you have stored procedures that *don't* return an ID column, that's not something we support.  We could look at extending the IUnitOfWork.Project() method to support stored procedures, and the LINQ provider to enable stored procedures to be projected into a named or anonymous type.  Perhaps you could say more about what problem you're trying to solve, and we'll see if we can offer some advice on an alternative approach.

 I know S.P. return an ID column, buuuut I have S.P. that don`t return an ID column.

There is my scenarios and maybe you will understand why I created custom class EntityBase derived from Entity.

1)  When SP return an ID Column:

CREATE  PROCEDURE [dbo].[StoredProcName] @Xvalue varchar(10) = NULL,@Yvalue varchar(10) = NULL AS  SELECT DISTINCT [ID],
      [returnX1] as [X1]
      ,[Description] as [X2]
      ,[returnX3] as [X3]
      ,0 as [X4]
  FROM [dbo].[table_name]
  WHERE ((@Xvalue IS NULL) OR (returnX1 LIKE '%' + @Xvalue + '%'))
        AND ((@Yvalue IS NULL) OR (Description LIKE '%' + @Yvalue + '%'))     
  Order by Description

results: more than 1000 records . Wrong!!

2)

CREATE  PROCEDURE [dbo].[StoredProcName] @Xvalue varchar(10) = NULL,@Yvalue varchar(10) = NULL AS  SELECT DISTINCT [0] as [ID],
      [returnX1] as [X1]
      ,[Description] as [X2]
      ,[returnX3] as [X3]
      ,0 as [X4]
  FROM [dbo].[table_name]
  WHERE ((@Xvalue IS NULL) OR (returnX1 LIKE '%' + @Xvalue + '%'))
        AND ((@Yvalue IS NULL) OR (Description LIKE '%' + @Yvalue + '%'))     
  Order by Description

 results: less than 500 records . that`s GOOD  !!!

3) but I would like that SP not return an ID Column and I created this EntityBase custom class

my SP look like this:

CREATE  PROCEDURE [dbo].[StoredProcName] @Xvalue varchar(10) = NULL,@Yvalue varchar(10) = NULL AS  SELECT DISTINCT [returnX1] as [X1]
      ,[Description] as [X2]
      ,[returnX3] as [X3]
      ,0 as [X4]
  FROM [dbo].[table_name]
  WHERE ((@Xvalue IS NULL) OR (returnX1 LIKE '%' + @Xvalue + '%'))
        AND ((@Yvalue IS NULL) OR (Description LIKE '%' + @Yvalue + '%'))     
  Order by Description

results: same as 2)

 well...  I expect the new version is released LS 3.0 ;)