I have the following view (SQL Server 2012 if it matters):
SELECT
EntityId
,EntityType
,StateId
FROM
SomeTable
INNER JOIN SomeOtherTable
When I generate an entity for this view (EF 6 - database first) it looks like this in the EDMX file:
<EntityType Name="VW_MyView">
<Key>
<PropertyRef Name="EntityId" />
<PropertyRef Name="EntityType" />
</Key>
<Property Name="EntityId" Type="Int32" Nullable="false" />
<Property Name="EntityType" Type="String" Nullable="false" MaxLength="2" FixedLength="false" Unicode="false" />
<Property Name="StateId" Type="Int32" />
</EntityType>
As you can see, the model generator created an entity key on the first two columns. The problem is, the first two columns do not guarantee uniqueness.
So for example I could have data like this in the view:
EntityId EntityType StateId
-------- ---------- -------
1234 CR 1
1234 CR 2
1234 CR 3
When I query the data using linq such as:
using (ContextA context = new ContextA())
{
var zList = context.VW_MyView.Where(f => f.EntityId == 1234
&& f.EntityType == "CR").ToList();
}
I get a list of three items, but like this (notice stateid duplicated):
EntityId EntityType StateId
-------- ---------- -------
1234 CR 1 <-- dupe
1234 CR 1 <-- dupe
1234 CR 1 <-- dupe
I migrated this exact same code from EF 4 (object context templates) to EF 6 (dbcontext templates), and before the migration it did not perform like this.
I know I can manually add an EntityKey to the StateId column, and it will work properly, but I have over 100 views in my model and I don't want to go through each one to check.
Why has this behavior changed, and is there a setting I can enable (globally) to correct this?
EDIT:
So based on the answers, I have been able to gather three ways to prevent this issue.
Add all primary key values from each consisting table into the view
Use nullif() tricks in the view to force columns to be non-nullable, and those be added by EF to the key
Manually add the Entity Key in the model myself
But this doesn't explain really why this happens, and how it could possibly be desired behavior? The EF linq query is simply returning entirely incorrect data, without any exceptions or warnings. I can't imagine this is correct.
I have the same "issue" in EF4 (with an .edmx file using the ObjectContext database-first approach) - not sure why it worked for you.
For Entity Framework, if it doesn't have a specified primary key (like on a table), it will fall back to using all non-nullable columns of that object (here: your view) as its compound PK.
These non-nullable columns are now the key for the table/view, and thus, only one value of that key can exist.
In order to resolve this, you need to either include more columns in your view to make the auto-detected key really unique (by including e.g. the primary key of all underlying base tables), or you need to manually set the key properly to something that works for you.
Another solution I found is by setting entity's MergeOption to NoTracking.
using (ContextA context = new ContextA())
{
context.VW_MyView.MergeOption = System.Data.Objects.MergeOption.NoTracking;
//Rest code goes here...
}
Solution found in this thread
Related
Well tested running system have already defined entity called 'User'.
Now I need to add a new property to User entity (ex: Age)
To do this in the safe way, I do not like to do any changes with the existing data base table, because that is very risky in my case. I need a way to rebuild the User entity with the minimum code changes.
So my proposal is:
Create a new table (user_age), with two columns (user_id, age)
Modify the user entity to add property 'age' and its getter-setters
So my entity (User) properties, will be saved to two different tables (user and user_age)
Loading the user is also similarly.
Is this possible to do with hibernate....??
If not, Any other safer way to do this with Hibernate...?
what are the available ORMs that provide this kind of feature (nhibernate, entityframwork,etc... or any other ORM)...?
Yes, there are various approaches:
[1] See JPA Secondary Tables. This allows you to map an Entity to two or more tables.
Section 2.2.7: http://docs.jboss.org/hibernate/annotations/3.5/reference/en/html_single/#d0e2235
[2] Create another Entity, say UserInfo, mapped to this new table. Create a one-to-one mapping from User to UserInfo.
Yes. You can do that.
I've used for a similar problem a joined-subclass.
Base:
<class name="User" table="Users">
<id name="Code" type="System.Guid">
<column name="Code" />
<generator class="guid.comb" />
</id>
...
</class>
Subclass:
<joined-subclass name="UserExt" extends=User" table="UsersExt">
<key column="Code" />
<property name="Age">
<column name="Age" not-null="true" />
</property>
</joined-subclass>
A good reference here.
NHibernate's join mapping is for exactly this case.
See Ayende's blog and the documentation for more information. From the documentation:
Using the <join> element, it is possible to map properties of one class to several tables, when there's a 1-to-1 relationship between the tables.
From my searches, it looks like it is also possible to do this with Entity Framework: Simon J Ince - Mapping two Tables to one Entity in the Entity Framework . I think this article is about Entity Framework v1, and things could have changed by now, but it appears that there is an important limitation in Entity Framework's version of this mapping:
... it requires a record in each table to exist as the generated SQL uses an INNER JOIN. It makes sense if you're using a new model, but I guess this is more tricky if you're mapping to an existing schema and data.
With NHibernate, you can set the optional attribute on the join mapping to tell it to use outer joins instead of inner joins.
optional (optional - defaults to false): If enabled, NHibernate will insert a row only if the properties defined by this join are non-null and will always use an outer join to retrieve the properties.
I have in an NHibernate 2 application an Product entity which has an many-to-many relationship to an Location. The Location entity does not have any navigation properties or mappings back to the product. It is mapped like this:
<bag name="Locations" table="ProductLocation" cascade="none">
<key column="ProductId" />
<many-to-many column="LocationId" class="Location"/>
</bag>
The product also has an composite-element, a Component with a concentration mapped via the ProductComponent class. The class has no navigation property or mapping back to the product.
<bag name="ProductComponents" table="ProductComponent" access="nosetter.camelcase">
<key column="ProductId" />
<composite-element class="ProductComponent">
<property name="Concentration"/>
<many-to-one name="Component" column="ComponentId" access="nosetter.camelcase"/>
</composite-element>
</bag>
This all works fine when just inserting one product at a time. It however fails when batch inserting multiple products.
While the products itself get inserted fine, each product does get an own unique Id, the elements in the many-to-many (Locations) and composite-element (ProductComponent) doesn't get inserted well. This is because NHibernate multiple times executes the insert to the ProductLocation table with the same ProductId.
This causes an duplicate record in the link table. How can this be prevented?
You'll have to define one site of the relationship to be the owner so that only one side does the insert. This can be achieved with Inverse set to true on the other side.
Find a more detailed explanation here
I have a View which looks similar to this:
SELECT Id, Name
FROM Users
UNION ALL
SELECT NULL as [Id], NULL as [Name]
When I try to map to this view in Entity Framework, it just fails. I don't get an error, but the view does not exist in my data store. Why is this? Is there a way around it?
I know this is an older question already marked as answered, but I wanted to post an alternative to editing the edmx. I learned this one the hard way, after tons of Google searches and pulling my hair out for hours, trying different options...
For a view, EF attempts to to infer a primary key by identifying columns that are non-nullable
and non-binary (see Create an Entity Key When No Key Is Inferred).
With a view used to flatten related data for lookup purposes, this can result in many columns (or the wrong ones) being inferred as keys.
For a view with a UNION, it can cause the opposite problem, because there may be no true identity column that can be safely included as a key.
To force EF to identify columns as a key, you can use ISNULL() to ensure the value is non-nullable: ISNULL(column_name, replacement_value)
To force EF to not mark a column as a key, you can use NULLIF() to make it nullable: NULLIF(column_name, value_not_equal_to_parameter_1)
If you need to ensure a value is returned, but don't want to have it marked as a key, I believe COALESCE(column_name, replacement_value) will do the job of ISNULL without EF marking the column as a key
If there is truly no viable column available as a primary key (as in a UNION view), you can fake a non-nullable PK through the use of ISNULL() combined with ROW_NUMBER() in your SELECT statement: SELECT ISNULL(ROW_NUMBER() OVER (ORDER BY sort_criteria), 0) as 'ALIASNAME'
As an alternative, the edmx can absolutely be edited directly as Marcos Prieto suggested, but you run the risk of having those changes overwritten the next time you run "Update Model from Database".
Hope this helps anyone who encounters this in the future!
It is because Visual Studio cannot infers the Primary Key of your View.
You can see the error message within edmx file by open it with XML editor and see the SSDL section.
Here is error message that results from my Model(which I created some View like yours within my Database just to emulate) :
Errors Found During Generation:
warning 6013: The table/view 'PhoneBook.dbo.ContactCustomer' does not have
a primary key defined and no valid primary key could be inferred.
This table/view has been excluded. To use the entity,
you will need to review your schema,
add the correct keys, and uncomment it.
It is not true that Union is not supported in EF 4.
But I think the problem is that Visual Studio saw your view as the odd View.
You can doing some experiment by create another View and compares them (using update model from database menu within model designer).
And you can modify the Model by hand (manual typing the edmx file) to define the Primary Key to resolve this.
I had a view that worked perfectly. I modified it (changed the view on a union of 2 tables), updated the model from database and this problem appeared.
I fixed it in 3 steps:
Open the .edmx file with a XML editor.
Uncomment the view's EntityType XML (edmx:StorageModels > Schema) code and add the Key:
<EntityType Name="your_view">
<Key>
<PropertyRef Name="your_id" />
</Key>
<Property Name="your_id" Type="int" Nullable="false" />
<Property Name="other_field" Type="varchar" MaxLength="45" />
</EntityType>
Be sure that EF didn't erased the view in edmx:StorageModels > Schema > EntityContainer (if you have code repository, copy the code from there):
<EntitySet Name="your_view" EntityType="Your_Model.Store.your_view" store:Type="Views" store:Schema="your_schema" store:Name="your_view">
<DefiningQuery>SELECT
`your_view`.`your_id`,
`your_view`.`other_field`,
FROM `your_view` AS `your_view`
</DefiningQuery>
</EntitySet>
I know this is an old question but i faced this issue recently & after trying to do the above mentioned methods i simply created another view that selects from the Union view & i was able to map the new view by updating my entity model.
This seems like it should be quite obvious but something about the entity framework is confusing me and I cannot get this to work.
Quite simply, I have three tables where the Id values are identity columns:
Users (userId, username)
Categories (categoryId, categoryName)
JoinTable (UserId, CategoryId) composite.
In the entities designer (this is .net 4.0), when I import these tables, as expected the join table does not appear but Users and Categories show a relationship. The following code:
var _context = new MyContext();
var myUser = new User();
myUser.UserName = "joe";
var myCategory = new Category();
myCategory.CategoryName = "friends";
_context.Users.AddObject(myUser);
myUser.Categories.Add(myCategory);
var saved = _context.SaveChanges();
Returns an error of (though nothing was added to the database):
An item with the same key has already been added.
If I add the following before saving:
_context.Categories.AddObject(myCategory);
myCategory.Users.Add(myUser);
I get the same error and nothing saved to the db. If I save the myUser and myCategory object before trying to associate them, they both save, but the second save throws an error, with nothing added to the join table:
Cannot insert the value NULL into column 'UserId', table '...dbo.JoinTable'; column does not allow nulls. INSERT fails. The statement has been terminated.
I'm clearly failing to understand how many to many relationships are inserted. What am I missing?
You do need to call SaveChanges() after adding User and Category entities to the database, and then set your association between them.
However, the real problem here is the second exception you listed. If you look at SqlProfiler or the ADO.NET profiler within the debugger, you will see that during the second SaveChanges call it looks something like this:
insert [dbo].[JoinTable]([UserId]) values (#0) select [CategoryId] from
[dbo].[JoinTable] where ##ROWCOUNT > 0 and [UserId] = #0 and [CategoryId] = scope_identity()
Obviously this won't work if you programmed your JoinTable correctly (composite PK on both columns).
If I look at the EntityModel store through Model Browser, it shows that the CategoryId column inside JoinTable does indeed have StoreGeneratedPattern set to Identity while UserId is set to None. Why EF did this during the generation phase when a composite PK was present is beyond me. I'll be posting a bug about this to MS, however in the mean time you can manually edit the edmx/ssdl file after generation to remove the Identity specifier. Find the StoreGeneratedPattern="Identity" string under the Property tag of the EntityType tag for your JoinTable and remove it:
Change:
<Property Name="CategoryId" Type="int" Nullable="false" StoreGeneratedPattern="Identity" />
To:
<Property Name="CategoryId" Type="int" Nullable="false" />
Then when you run your code you will get a much better insert query (and no more exception!):
insert [dbo].[JoinTable]([UserId], [CategoryId]) values (#0, #1)
The way I have done this is to first generate a valid Category entity with the entity key.
Category myCategory = _context.Categories.First(i => i.CategoryID == categoryIDToUse);
Or you can try to create the entity as a stub to save the hit to the DB:
Category myCategory = new Category{CategoryID = categoryIDToUse };
Then add that entity to the entity set(CategorySet) on the ObjectContext using the AttachTo method(you may want to check if it is already attached). Then you can add the Category to your User entity using the Add method. Something like this:
myUser.Categories.Add(myCategory);
Call SaveChanges(). That has worked for me.
When new parent added to Context, state of all object in object tree changes to Added and hence EF tries to save all such objects as new record.
Refer this link:
http://nileshhirapra.blogspot.in/2012/03/entity-framework-insert-operation-with.html
Using the Entity Framework, I've modeled a fairly simple database schema with an ever-so-slightly more complex class hierarchy. In two places, I'm using single table inheritance with a single NVARCHAR(20) NOT NULL discriminator column. In one of those two places, it works great, no issues. But in the other place, with an almost identical pattern, I get the following error:
Error 3023: Problem in Mapping Fragments starting at lines 371, 375, 379, 382: Column MediaStream.MediaStreamTypeID has no default value and is not nullable. A column value is required to store entity data.
An Entity with Key (PK) will not round-trip when:
((PK does NOT play Role 'MediaStream' in AssociationSet 'FK_MediaStream_SessionID' OR PK is NOT in 'MediaStream' EntitySet OR Entity is type [SlideLinc.Model].MediaStream) AND (PK plays Role 'MediaStream' in AssociationSet 'FK_MediaStream_SessionID' OR PK is NOT in 'MediaStream' EntitySet OR Entity is type [SlideLinc.Model].MediaStream) AND (PK plays Role 'MediaStream' in AssociationSet 'FK_MediaStream_SessionID' OR PK is in 'MediaStream' EntitySet))
Here's the table definition (not including various indexes, foreign keys, etc.):
CREATE TABLE [dbo].MediaStream(
[MediaStreamID] UNIQUEIDENTIFIER NOT NULL,
[SessionID] UNIQUEIDENTIFIER NOT NULL,
[RtmpUri] nvarchar(250) NOT NULL,
[MediaStreamTypeID] nvarchar(20) NOT NULL,
CONSTRAINT PK_MediaStream PRIMARY KEY CLUSTERED
(
[MediaStreamID] ASC
)
I'm using the MediaStreamtypeID column as the discriminator: if it's set to "video", a VideoMediaStream class should be created, and if it's set to "audio", an AudioMediaStream class should be created.
The relevant portions of the EDMX file look like this:
<EntitySetMapping Name="MediaStream">
<EntityTypeMapping TypeName="IsTypeOf(SlideLinc.Model.MediaStream)">
<MappingFragment StoreEntitySet="MediaStream">
<ScalarProperty Name="RtmpUri" ColumnName="RtmpUri" />
<ScalarProperty Name="MediaStreamID" ColumnName="MediaStreamID" /></MappingFragment></EntityTypeMapping>
<EntityTypeMapping TypeName="IsTypeOf(SlideLinc.Model.VideoMediaStream)">
<MappingFragment StoreEntitySet="MediaStream" >
<ScalarProperty Name="MediaStreamID" ColumnName="MediaStreamID" />
<Condition ColumnName="MediaStreamTypeID" Value="video" /></MappingFragment></EntityTypeMapping>
<EntityTypeMapping TypeName="IsTypeOf(SlideLinc.Model.AudioMediaStream)">
<MappingFragment StoreEntitySet="MediaStream" >
<ScalarProperty Name="MediaStreamID" ColumnName="MediaStreamID" />
<Condition ColumnName="MediaStreamTypeID" Value="audio" /></MappingFragment></EntityTypeMapping></EntitySetMapping>
<AssociationSetMapping Name="FK_MediaStream_SessionID" TypeName="SlideLinc.Model.FK_MediaStream_SessionID" StoreEntitySet="MediaStream">
<EndProperty Name="MediaStream">
<ScalarProperty Name="MediaStreamID" ColumnName="MediaStreamID" /></EndProperty>
<EndProperty Name="Session">
<ScalarProperty Name="SessionID" ColumnName="SessionID" /></EndProperty></AssociationSetMapping>
So there are multiple things about this error that I don't get:
(1) Why does exactly this same approach work for my other class hierarchy, but not this one? I thought it might be the Entity Designer getting confused, so I deleted this portion of my hierarchy (in the XML), and recreated it, but I'm still getting it. I could try recreating the whole damn thing, but hell, that's a lot of work, and if I'm gonna have to be doing this very often, that doesn't leave a great taste in my mouth about the entity framework.
(2) What is it complaining about in the first place? I don't get how MediaStreamTypeID (which isn't a member of the primary key) has anything to do with the primary key at all, or why the fact that it can't be null is a problem (especially given that this same setup works elsewhere in my model!).
Any thoughts or suggestions?
I had a similar problem, and was able to solve it by setting the "Abstract" property for the base class to "True" and by removing the discriminator column from the base class in the model (either in the *.edmx file or in the designer view within Visual Studio).
I got the exactly same error but maybe with different reason from yours.
In the code below, I mis-copied some codes so 2 of inherited classes (LeveledItem and Team) are of the same "Type".
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
Database.SetInitializer(new ScrumDbContextInitializer());
base.OnModelCreating(modelBuilder);
modelBuilder.Conventions.Remove<PluralizingTableNameConvention>();
modelBuilder.Entity<LeveledItem>()
.Map<LeveledItem>(m => m.Requires("Type").HasValue(typeof(LeveledItem).Name))
.Map<Team>(m => m.Requires("Type").HasValue(typeof(LeveledItem).Name))
.Map<Story>(m => m.Requires("Type").HasValue(typeof(Story).Name))
.Map<Task>(m => m.Requires("Type").HasValue(typeof(Task).Name)) .Map<Sprint>(m => m.Requires("Type").HasValue(typeof(Sprint).Name));
After the second one was changed to "typof(Team).Name", the error was fixed.
I had a similar problem to this, just out of interest, does deleting the .EDMX file and recreating it from scratch solve your problem?
What caused the problem:
Created a set of tables with some 0..* mappings
Generated the Entity Framework class, customised a whole bunch of Navigation properties
Went back to the DB and changed the cardinality of some of the 0..* to 1..* relationships - ie: set some of the FKs to !nullable
Updated the Entity Framework
Recompiled and BOOM I got a compilation error similar to yours
There seems to be an issue where the "Update Model from Database" command doesn't update the changed relationships correctly.
The solution was to open the EDMX file, look for the <Association Name="FK_XXX_XXX"> elements that were generated the first time and change the Multiplicity attribute on the relevant End point from Multiplicity="0..1" to Multiplicity="1"
I encountered this error when I inadvertently mapped 2 classes to the same table via the [Table] attribute (same effect via modelBuilder ToTable())