I implemented inheritance with a discriminator field so all my records are in the same table. My basetype is Person (also the name of the table) and Driver and Passenger inherit from it. I receive instances of the correct type (Driver and Passenger) when I perform a query on the object context to Person. example:
var q = from d in ctx.Person
select d;
But I also create a function that calls a stored procedure and mapped the output of the function to the type Person. But now I get a list of Person and not Drivers or Passengers when I execute this method.
Anybody an idea how to solve this or is this a bug in EF4?
AFAIK, you can't use discriminator mapping (e.g TPH) when dealing with stored procedure mappings.
The stored procedure must be mapped to a complex type or custom entity (e.g POCO), the mapping cannot be conditional.
What you could do is map it to a regular POCO, but then project that result set into the relevant derived type (manual discrimination).
E.g:
public ICollection<Person> GetPeople()
{
var results = ExecuteFunction<Person>(); // result is ObjectResult<Person>
ICollection<Person> people = new List<Person>();
foreach (var result in results)
{
if (result.FieldWhichIsYourDiscriminator == discriminatorForDriver)
{
people.Add((Driver)result);
}
// other discriminators
}
}
If your always expecting a collection of one type (e.g only Drivers), then you wouldn't need the foreach loop, you could just add the range. The above is in case you are expecting a mixed bag of different people types.
Would be interested to see other answers, and if there is a better way though - but the above should work.
Related
Suppose I have a class Parent, with two subclasses, AChild and BChild. I have these mapped to a single table using Entity Framework 5.0.0 on .NET 4.5, using TPH.
public abstract class Parent {
public string Type { get; set; } // Column with this name in DB is discriminator.
public string Status { get; set; }
}
public class AChild : Parent {
// Other stuff.
}
public class BChild : Parent {
// Other stuff.
}
The code to configure the mapping:
class ParentConfiguration : EntityTypeConfiguration<Parent> {
Map((EntityMappingConfiguration<AChild> mapping) => mapping
.Requires("Type")
.HasValue("A"));
Map((EntityMappingConfiguration<BChild> mapping) => mapping
.Requires("Type")
.HasValue("B"));
}
I have a need to run a query that returns both AChild and BChild objects. However, it needs to filter ONLY the AChild rows by a second column, which in this example I will call Status.
Ideally I would want to do the following:
public IList<Parent> RunQuery() {
IQueryable<Parent> query =
this.context.Set<Parent>()
.Where((Parent parent) => !parent.Type.Equals("A") || parent.Status.Equals("Foo"))
.OrderBy((Parent parent) => parent.Number);
return query.ToList();
}
This doesn't work. It insisted on looking for a "Type1" column instead of just letting both the discriminator and a "Type" property be mapped to the same "Type" column.
I know of the "OfType" extension method that can be used to completely filter down to one type, but that's too broad a brush in this case.
I could possibly run multiple queries and combine the results, but the actual system I'm building is doing paging, so if I need to pull back 10 rows, it gets messy (and inefficient) to query since I'll either end up pulling back too many rows, or not pull back enough and have to run extra queries.
Does anyone have any other thoughts?
There are few problems. First of all you cannot have discriminator mapped as a property. That is the reason why it is looking for Type1 column - your Type property results in second column because the first one is already mapped to .NET types of your classes. The only way to filter derived types is through OfType.
The query you want to build will be probably quite complex because you need to query for all Bs and concatenate them with result of query for filtered As. It will most probably not allow you to concatenate instances of Bs with As so you will have to convert them back to parent type.
Can anyone explain the behavior I am seeing in the minimal code example below? It seems that for a given field or property, the same two instances of the Entry class are being reused in each iteration of the LINQ to SQL query, even though I use the new operator. The same problem does not show up for LINQ to objects queries. I created a C# console application project using .NET Framework 4 and connecting to a SQL Server 2005 Enterprise database.
public class Set
{
public Entry Field;
public Entry Property { get; set; }
}
public class Entry
{
public int ID;
public string Name { get; set; }
}
class Program
{
static void Main(string[] args)
{
var dc = new DataClasses1DataContext(); // just create a simple dbml with some table from some database
var resultQuery = (
from x in dc.SomeTable
select new Set()
{
Field = new Entry(),
Property = new Entry()
}
);
var sets = resultQuery.ToArray();
Test(sets);
var source = Enumerable.Range(0, 10);
var sourceQuery = (
from x in source
select new Set()
{
Field = new Entry(),
Property = new Entry()
}
);
var sets2 = sourceQuery.ToArray();
Test(sets2);
}
static void Test(Set[] sets)
{
var f = sets[0].Field;
Console.WriteLine(sets.All(x => object.Equals(x.Field, f)));
var p = sets[0].Property;
Console.WriteLine(sets.All(x => object.Equals(x.Property, p)));
Console.Writeline(sets.Length);
Console.WriteLine(object.Equals(f, p));
Console.WriteLine();
}
}
The output of Test() for the LINQ to SQL query is
True
True
1362
False
which indicates that for all of the Set objects produced, all the Field members point to the same single Entry instance and all the Property members point to the same single Entry instance. I.e., the same instance is reused for a respective member in every iteration of the query.
The output of Test() for the LINQ to objects query is
False
False
10
False
which indicates that a new distinct instance is created in each iteration of the query.
Is the LINQ to SQL behavior expected or a bug? Does anyone know if this happens with the Entity Framework?
I don't know if this is a bug or if and why this is expected in LINQ to SQL. I can only answer your last question if that also happens with Entity Framework.
The answer is: No.
With EF you have to use object initializer syntax though when you instantiate the Entry objects. Using the default constructor leads to an exception:
var resultQuery = (
from x in dc.SomeTable
select new Set
{
Field = new Entry { Name = "X" },
Property = new Entry { Name = "X" }
}
);
It doesn't matter how you initialize. Using the code above (and with 4 rows in a small test table) I get this output with your test program:
False
False
4
False
False
False
10
False
It looks that there is a big difference between LINQ to SQL and Entity Framework regarding object materialization during projections.
(I've tested with EF 4.1/DbContext.)
Edit
If I take the modified query in my code snippet above also for your LINQ to SQL query and watch the generated SQL I get the following:
SELECT NULL AS [EMPTY]
FROM [dbo].[SomeTable] AS [t0]
Whereas the same with LINQ to Entites creates this query:
SELECT
1 AS [C1],
N'X' AS [C2],
N'X' AS [C3]
FROM [dbo].[SomeTable] AS [Extent1]
My interpretation is that LINQ to SQL parses the projection code and queries only columns for property values which depend on the "row variable" x. All other properties are filled on the client when the objects get materialized. If an object does not depend on a column value at all, LINQ to SQL creates a single constant object and reuses it in the whole result collection.
In constrast to this Entity Framework also sends constant values (independent of x) to the database server. The values get send back to the client and EF treats those values as if they were column values and updates the properties of the objects in the projection.
This leads also to the big difference that something like this...
Random random = new Random();
var resultQuery = (
from x in dc.SomeTable
select new Set
{
Field = new Entry { ID = random.Next() },
Property = new Entry { Name = "X" }
}
);
...works in LINQ to SQL because apparently the random function value (which is independent of x) is evaluated on the client and then assigned to the property. But EF wants to translate the right side of the property assignment into SQL and send it as SQL fragment to the database server - which fails and leads to the infamous "...cannot translate into store expression..." exception.
Edit 2
BTW: The last code snippet above still creates only a single Field instance in the whole collection: random.Next() is only evaluated once (and also the constructor of Entry is only called once for the Field object). This now is indeed confusing because writing such code one would expect that you want to have a random value for each row returned from the database. It's not the case.
I have 2 Entities, each of which is managed by EF Code First, and each happily sitting in its own table. Entity_A has a property, called "SumTotal", which should be the sum of a specific column in Entity_B that matches a certain criteria.
SumTotal should not be persisted in the DB, but rather calculated each time an instance of Entity_A is retrieved.
I have looked at ComputedColumns, but it appears that the computedcolumn can only be defined relative to columns in the same table.
I also have a feeling that I need to set SumTotal to NotMapped (or something similar with AutoGenerated), but dont know how to get the actual value into SumTotal.
Hope this question makes sense, thanks in advance
You can project the results to an anonymous object and transform that it to your entity
var projection = db.EntityAs.Where(/* */)
.Select(a => new {A = a, Sum = a.Bs.Sum(b => b.Total)})
foreach(p in projection)
{
p.A.SumTotal = p.Sum;
}
var As = projection.Select(p => p.A);
I have a statement:
var items = from e in db.Elements
join a in db.LookUp
on e.ID equals a.ElementID
where e.Something == something
select new Element
{
ID = e.ID,
LookUpID = a.ID
// some other data get populated here as well
};
As you can see, all I need is a collection of Element objects with data from both tables - Elements and LookUp. This works fine. But then I need to know the number of elements selected:
int count = items.Count();
... this call throws System.NotSupportedException:
"The entity or complex type 'Database.Element' cannot be constructed in a LINQ to Entities query."
How am I supposed to select values from multiple tables into one object in Entity Framework? Thanks for any help!
You are not allowed to create an Entity class in your projection, you have to either project to a new class or an anonymous type
select new
{
ID = e.ID,
LookUpID = a.ID
// some other data get populated here as well
};
Your code doesn't work at all. The part you think worked has never been executed. The first time you executed it was when you called Count.
As exception says you cannot construct mapped entity in projection. Projection can be made only to anonymous or non mapped types. Also it is not clear why you even need this. If your class is correctly mapped you should simply call:
var items = from e in db.Elements
where e.Something == something
select e;
If LookupID is mapped property of your Element class it will be filled. If it is not mapped property you will not be able to load it with single query to Element.
I've set up a many-to-many association between two tables based on a third table that just holds a pair of key values. Now I'd like to do a query that groups the right tables key values by the lefts without needing other data.
LeftTable { LeftID, LeftField1, LeftField2 }
JoinTable { LeftID, RightID}
RightTable { RightID, RightField1, RightField2 }
Is there any way to essentially just query the JoinTable and get all the 'RightIDs' grouped by the 'LeftIDs' without the SQL trying to fetch the fields from either side?
The JoinTable is not an entity in its own right in the model, but is mapped to the association.
I've experimented a bit with both using ObjectQuery and EntityCommand (ESQL) and both seem to still load in the other fields by joining to RightTable which I don't need.
My ESQL looks something like:
SELECT lt.LeftID, (SELECT rt.RightID
FROM NAVIGATE(lt, MyModel.LeftToRightAssoc, RightTable) as rt)
FROM MyEntities.LeftTable as lt;
but the generated SQL is still fetching in RightField1 and RightField2.
Surely there must be a simpler way to do this?
Assuming that your class Left has a navigation property Rights (a collection of Right entities) you could try this:
var list = context.Lefts.Select(l => new
{
LeftId = l.LeftId,
RightIds = l.Rights.Select(r => r.RightId)
});
foreach (var item in list)
{
Console.WriteLine("LeftId = {0}", item.LeftId);
foreach (var rightId in item.RightIds)
{
Console.WriteLine("RightId = {0}", rightId);
}
}
You would get a collection of anonymous type objects where each element has the LeftId and a collection of corresponding RightIds. This query should not touch the other fields like RightField1, etc. Instead of an anonymous type you could also create your own custom type and then project into this type in the query above.