The latest EF Code First NuGet package comes with a custom implementation of IDatabaseInitializer called DontDropDbJustCreateTablesIfModelChanged. As the name implies, when a model change is detected, it will not drop and recreate the whole database, just drop and recreate the tables.
Say I have this model class:
public class User
{
public string Username { get; set; }
// This property is new; the model has changed!
public string OpenID { get; set; }
}
How would one go about implementing an IDatabaseInitializer that doesn't drop any tables either. In this case, it would just add an OpenID column to the User table?
I think it is a matter of SQL. So for SQL Server you can write something like:
public class MyInitializer : IDatabaseInitializer<MyContext>
{
public void InitializeDatabase(MyContext context)
{
context.Database.SqlCommand(
#"
IF NOT EXISTS (SELECT 1 FROM sys.columns AS col
INNER JOIN sys.tables AS tab ON tab.object_Id = col.object_Id
WHERE tab.Name = 'User' AND col.Name = 'OpenId')
BEGIN
ALTER TABLE dbo.User ADD OpenId INT;
END");
}
}
But in the same way you can execute such script without adding it to your application which I think is much better approach.
With the current version of Code First, you cannot simply amend your schema and preserve any data that you might have in your tables. If maintaining data, such as reference data / lookup tables is important with this release you can create your own Initializer and override the Seed method to populate your tables
public class MyDbInitializer : DropCreateDatabaseIfModelChanges<MyContext>
{
protected override void Seed(MyContext context)
{
var countries = new List<Country>
{
new Country {Id=1, Name="United Kingdom"},
new Country{Id=2, Name="Ireland"}
};
countries.ForEach(c => context.Countries.Add(c));
}
}
And then use this in your Application_Start:
Database.SetInitializer<MyContext>(new MyDbInitializer());
I believe that this is being addressed currently by the EF Team, but wasn't ready for release at the time the Code First drop came out. You can see a preview here: Code First Migrations
Related
I use Entity Framework 6.2 Code First (.net framework 4.6.1) and I map few entities to view via Table Attribute. It works for select operations and I handle Insert/Update/Delete with writing trigger to view at sql server side. It works as expected, however when I add a new migration, Entity Framework generate RenameTable scripts for used Table Attribute (actuallyis expected behavior for EF). But I want to intercept migration generation and change these entities tableName to original name.
my code like;
[MapToView("Users","UsersView")]
public class User
{
...
}
I wrote MapToView Attribute, this attribute inherited by TableAttribute and pass to second parameter to TableAttribute. I create this Attribute because if I intercept migration generation, return original table name with this attribute parameters.
In this case when I run "add-migration migrationName" it creates migration scripts like this;
RenameTable(name: "dbo.Users", newName: "UsersView");
but i want to create empty migration when I run "add-migration migrationName" script.
anyone can help me?
I solve the problem.
First: Problem is; When I Map Entity to View EF Code-first generate migration with ViewName. This is problem because I want to use View Instead of Table. So I solve problem with this instructions;
1- I Create BaseEntityConfiguration that Inherited from EntityTypeConfiguration and all entity configuration classes are inherited by.
for example:
public class UserConfig: BaseEntityConfiguration<User> //Generic Type is Entity
{
public UserConfig()
{
}
}
2- I Create MapToViewAttribute that inherited by TableAttribute
public class MapToViewAttribute : TableAttribute
{
public string TableName { get; }
public string ViewName { get; }
public MapToViewAttribute(string tableName, string viewName) : base(viewName)
{
TableName = tableName;
ViewName = viewName;
}
}
3- I Use MapToViewAttribute for example User Entity to use View.
[MapToView("User","UserView")]
public class User
{
...
}
And in BaseEntityConfiguration's Constructor I Get Generic Type and custom attributes. If any entity has MapToView Attribute, I pass to TableName parameter to ToTable Method. So at runtime EF uses View for these entities but doesn't create migration with RenameTable for these entities.
protected BaseEntityConfiguration()
{
var baseType = typeof(TEntityType);
var attributes = baseType.GetCustomAttributes(true);
foreach (var attr in attributes)
{
if (attr.GetType() == typeof(MapToViewAttribute))
{
var tableName = ((MapToViewAttribute)attr).TableName;
ToTable(tableName);
}
}
}
Last EF don't use your configuration files, so you must tell the EF to use this in DbContext class's InternalModelCreate method.
My implementation like this;
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
var typesToRegister = Assembly.GetExecutingAssembly()
.GetTypes().Where(IsConfigurationType);
foreach (var type in typesToRegister)
{
dynamic configurationInstance = type.BaseType != null
&& type.BaseType.IsGenericType
&& type.BaseType.GetGenericTypeDefinition() == typeof(BaseEntityConfiguration<>)
? Activator.CreateInstance(type, culture)
: Activator.CreateInstance(type);
modelBuilder.Configurations.Add(configurationInstance);
}
modelBuilder.Types().Configure(t => t.ToTable(t.ClrType.Name));
BaseDbContext.InternalModelCreate(modelBuilder);
}
But if you use this approach you must create Insert, Update and Delete Triggers/Rule (if you use SQLServer trigger is an option but if you use postgresql rule is better option) because EF uses this view for insert, update and delete operations.
I have a model which has an auto-incrementing ID field by default as is normal. However, I wish to seed the database with initial data and because there are foreign keys I wish to explicitly set the IDs of the seeded data.
My model
public class EntAttribute
{
public int ID { get; set; }
public string Title { get; set; }
}
My seeding code:
public class Seeder
{
private class AllAttributes
{
public List<EntAttribute> Attributes { get; set; }
}
public bool SeedData()
{
AllAttributes seedAttributes;
string strSource;
JsonSerializer JsonSer = new JsonSerializer();
strSource = System.IO.File.ReadAllText(#"Data/SeedData/Attributes.json");
seedAttributes = JsonConvert.DeserializeObject<AllAttributes>(strSource);
_context.AddRange(seedAttributes.Attributes);
_context.SaveChanges();
return true;
}
}
Please note, I'm very new to both EFCore and C#. The above is what I've managed to cobble together and it seems to work right up until I save the changes. At this point I get:
SqlException: Cannot insert explicit value for identity column in table 'Attribute' when IDENTITY_INSERT is set to OFF.
Now I'm smart enough to know that this is because I can't explicitly set the ID field in the EntAttribute table because it wants to assign its own via auto-increment. But I'm not smart enough to know what to do about it.
Any help appreciated.
EDIT: Adding the solution based on the accepted answer below because the actual code might help others...
So I added to my Context class the following:
protected override void OnModelCreating(ModelBuilder modelBuilder)
{
modelBuilder.HasSequence<int>("EntAttributeNumbering")
.StartsAt(10);
modelBuilder.Entity<EntAttribute>()
.Property(i => i.ID)
.HasDefaultValueSql("NEXT VALUE FOR EntAttributeNumbering");
}
This first ensures the a sequence is created (the name is arbitrary) and then secondly, sets it to be used for the relevant table instead of auto-increment. Once this was done I was able to my seed data. There are fewer than 10 records so I only needed to set the start value for the sequence to 10. More would normally make sense but I know there will never be more.
I also had to blitz my migrations because they'd somehow got in a mess but that's probably unrelated.
With EF Core you can create and use a Sequence object to assign the IDs, and you can reserve a range of IDs for manual assignment by picking where the sequence starts. With a Sequence you can assign the IDs yourself, or let the database do it for you.
FYI for people using EF Core 3, if using int for your key you can set the start sequence value incase you have seeded data. I found this a much cleaner to solve this problem in my use case which just had a single seeded record.
e.g
modelBuilder.Entity<TableA>()
.Property(p => p.TableAId)
.HasIdentityOptions(startValue: 2);
modelBuilder.Entity<TableA>()
.HasData(
new TableA
{
TableAId = 1,
Data = "something"
});
https://github.com/npgsql/efcore.pg/issues/367#issuecomment-602111259
Is it possible to call a TVF in EF6 Code First?
I started a new project using EF6 Database first and EF was able to import a TVF into the model and call it just fine.
But updating the model became very time consuming and problematic with the large read-only db with no RI that I'm stuck dealing with.
So I tried to convert to EF6 code first using the Power Tools Reverse Engineering tool to generate a context and model classes.
Unfortunately the Reverse Engineering tool didn't import the TVFs.
Next I tried to copy the DBFunctions from my old Database First DbContext to the new Code First DbContext, but that gave me an error that my TVF:
"cannot be resolved into a valid type or function".
Is it possible to create a code first Fluent mapping for TVFs?
If not, is there a work-around?
I guess I could use SPs instead of TVFs, but was hoping I could use mostly TVFs to deal with the problematic DB I'm stuck with.
Thanks for any work-around ideas
This is now possible. I created a custom model convention which allows using store functions in CodeFirst in EF6.1. The convention is available on NuGet http://www.nuget.org/packages/EntityFramework.CodeFirstStoreFunctions. Here is the link to the blogpost containing all the details: http://blog.3d-logic.com/2014/04/09/support-for-store-functions-tvfs-and-stored-procs-in-entity-framework-6-1/
[Tested]
using:
Install-Package EntityFramework.CodeFirstStoreFunctions
Declare a class for output result:
public class MyCustomObject
{
[Key]
public int Id { get; set; }
public int Rank { get; set; }
}
Create a method in your DbContext class
[DbFunction("MyContextType", "SearchSomething")]
public virtual IQueryable<MyCustomObject> SearchSomething(string keywords)
{
var keywordsParam = new ObjectParameter("keywords", typeof(string))
{
Value = keywords
};
return (this as IObjectContextAdapter).ObjectContext
.CreateQuery<MyCustomObject>(
"MyContextType.SearchSomething(#keywords)", keywordsParam);
}
Add
public DbSet<MyCustomObject> SearchResults { get; set; }
to your DbContext class
Add in the overriden OnModelCreating method:
modelBuilder.Conventions.Add(new FunctionsConvention<MyContextType>("dbo"));
And now you can call/join with
a table values function like this:
CREATE FUNCTION SearchSomething
(
#keywords nvarchar(4000)
)
RETURNS TABLE
AS
RETURN
(SELECT KEY_TBL.RANK AS Rank, Id
FROM MyTable
LEFT JOIN freetexttable(MyTable , ([MyColumn1],[MyColumn2]), #keywords) AS KEY_TBL
ON MyTable.Id = KEY_TBL.[KEY]
WHERE KEY_TBL.RANK > 0
)
GO
I was able to access TVF with the code below. This works in EF6. The model property names have to match the database column names.
List<MyModel> data =
db.Database.SqlQuery<MyModel>(
"select * from dbo.my_function(#p1, #p2, #p3)",
new SqlParameter("#p1", new System.DateTime(2015,1,1)),
new SqlParameter("#p2", new System.DateTime(2015, 8, 1)),
new SqlParameter("#p3", 12))
.ToList();
I actually started looking into it in EF6.1 and have something that is working on nightly builds. Check this and this out.
I have developed a library for this functionality. You can review my article on
UserTableFunctionCodeFirst.
You can use your function without writing SQL query.
Update
First of all you have to add reference to the above mentioned library and then you have to create parameter class for your function. This class can contain any number and type of parameter
public class TestFunctionParams
{
[CodeFunctionAttributes.FunctionOrder(1)]
[CodeFunctionAttributes.Name("id")]
[CodeFunctionAttributes.ParameterType(System.Data.SqlDbType.Int)]
public int Id { get; set; }
}
Now you have to add following property in your DbContext to call function and map to the property.
[CodeFunctionAttributes.Schema("dbo")] // This is optional as it is set as dbo as default if not provided.
[CodeFunctionAttributes.Name("ufn_MyFunction")] // Name of function in database.
[CodeFunctionAttributes.ReturnTypes(typeof(Customer))]
public TableValueFunction<TestFunctionParams> CustomerFunction { get; set; }
Then you can call your function as below.
using (var db = new DataContext())
{
var funcParams = new TestFunctionParams() { Id = 1 };
var entity = db.CustomerFunction.ExecuteFunction(funcParams).ToList<Customer>();
}
This will call your user defined function and map to the entity.
I'm using the newest Entity Framework and ran into a problem with Many To Many Relationship when I want to create an extra column.
The issue is the same raised in this older post:
EF Code First Additional column in join table for ordering purposes
Is it still the problem today that one can not add an extra column without loosing the many to many relation ship (link from object A to B as A.B because the mapping becomes and entity it self) ?
What are the work a rounds ?
Look up the a of class A I need and then query for mapping table where(e=>e.A == a) to get my Bs? And when I need the extra colums i would do MappingTable.find(a,b) ?
Are there other modeling options, linq to sql that would make it easier ?
As far as I know things haven't changed with EF 5. You would need to do it as the link says to. I like to stick with EF as its easy to use, but that's just my opinion...
I had the same problem. What I did to work-around it was create another derivative DbContext specifically to handle joins. I.E.:
public class JoinContext : DbContext
{
internal JoinContext() : base("name=SampleConnectionString")
{
PreventErrorIfDatabaseSchemaChanges();
// Get the ObjectContext related to this DbContext
var objectContext = (this as IObjectContextAdapter).ObjectContext;
}
public DbSet<StudentImage> StudentImages { get; set; }
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
modelBuilder.Entity<StudentImage>()
.ToTable("StudentImages");
.HasKey(joinTable => new { joinTable.StudentId, joinTable.ImageId });
base.OnModelCreating(modelBuilder);
}
private static void PreventErrorIfDatabaseSchemaChanges()
{
Database.SetInitializer<JoinContext>(null);
}
}
I left the other application context with the Student/Image many-to-many join mapping as-is. Don't forget to specify a compounded key for the join table (refer to HasKey method above), else EF bombs on databse initialization.
After you have your special join context, use a repository to access this context and get or set the desired fields from mapped join table:
public class StudentRepository
{
public int GetImageSortOrder(int studentId, int imageId)
{
var joinContext = new JoinContext();
var joinTuple = joinContext.StudentImages.Find(studentId, imageId);
return joinTuple.SortOrder;
}
}
Hope this helps!
I just want to know if there's a way on how to create an Updatable model. Right now, I have to create procedures for insert, update, and delete for all of the tables in my model. This is very tedious so I was wondering if there is one way which I could do to resolve this?
I remember before in my previous work that we used to make models and access them (CRUD) without creating procedures. But i'm not really certain now on how it was made.
Thank you!
There are various ways in which you can automate the generation (on the fly or already generated at compile time) of the actual SQL calls to the database to insert, select, update and delete within the Entity Framework.
You can use the ORM tools (e.g. Linq to Entities) to minimise or eliminate the writing of raw SQL. This means you still have to use the correct attributes on your entities and the properties/methods therein and that's a manual process. (Some backgrounding on this MSDN page)
You can allow the framework to automatically generate your entities based on some existing database schema (only possible with SqlServer-type databases) which basically does 90% of the work for you. There may be some cases where you need to override, for example, the default insert SQL with something custom. This is achieved via the Generate Database Wizard (which I think is a part of Visual Studio 2008+).
You can use POCO classes with EF. If you're using 4.1 and above, you can use the DbContext class. To map your model to the table / columns, simply override OnModelCreating in your context class (which inherits from DbContext). Say you have a model called User, a table called Users, and the context class MyContext, the code could be smth like this:
public class User
{
public int UserId { get; set; }
public string UserName { get; set; }
}
public class MyContext : DbContext
{
public MyContext() :
base("MyContext")
{
}
public DbSet<User> Users { get; set; }
protected override void OnModelCreating(ModelBuilder modelBuilder)
{
base.OnModelCreating(modelBuilder);
modelBuilder.Entity<User>().
.ToTable("Users");
modelBuilder.Entity<User>().
.Property(d => d.UserId)
.HasColumnName("UserId")
modelBuilder.Entity<User>().
.Property(d => d.UserName)
.HasColumnName("UserName");
}
}
To use it, simply add the User instance to your DbSet, then call SaveChanges:
using(MyContext ctx = new MyContext())
{
var u = new User() { UserId = 1, UserName = "A" };
ctx.Users.Add(u);
ctx.SaveChanges();
}