Question

Can somebody please explain the difference between the following 3 possibilities for using the ColumnAttribute:

A: attribute on field

[Column(Name="ParentId")]
private int m_parentid;
public int ParentId { get { return m_parentid; } set { m_parentid = value; } }

B: attribute on property

private int m_parentid;
[Column(Name="ParentId")]
public int ParentId { get { return m_parentid; } set { m_parentid = value; } }    

C: attribute on property with storage set

private int m_parentid;
[Column(Name="ParentId", Storage="m_parentid")]
public int ParentId { get { return m_parentid; } set { m_parentid = value; } }    

I can understand B would be different from A and C in the case of a non-trivial getter / setter (Linq would presumably use the getter/setter in case B but not in case A or C, is that correct?)

But I don’t understand how there can ever be any difference between A and C.

Answer 1

C is useful to allow the engine to understand expression trees, for example a predicate:

var items = ctx.SomeTable.Where(x => x.ParentId == 21);

With “a” I would expect this to fail, as it doesn’t really know about the property ParentId (only the field m_parentid and the database column). Specifying the “Storage” in “c” allows the db-value to be saved directly to the field rather than using the property, avoiding code that shouldn’t be necessary during materialisation from the db.